3631

1 IN THE UNITED STATES DISTRICT COURT

FOR THE DISTRICT OF COLUMBIA
2

3 UNITED STATES OF AMERICA,

et al., Civil Action
4 Plaintiffs, No. 1:20-cv-3010

5 vs. Washington, DC
October 2, 2023
6 GOOGLE, LLC, 1:36 p.m.

7 Defendant. Day 14
_____________________________/ Afternoon Session
8

10 TRANSCRIPT OF BENCH TRIAL

BEFORE THE HONORABLE AMIT P. MEHTA
11 UNITED STATES DISTRICT JUDGE

12
APPEARANCES:
13
For DOJ Plaintiffs: KENNETH DINTZER
14 U.S. Department of Justice
1100 L Street, NW
15 Washington, DC 20005

16 ADAM SEVERT
MEAGAN BELLSHAW
17 U.S. Department of Justice
450 Fifth Street, NW
18 Washington, DC 20001

19 DAVID DAHLQUIST
U.S Department of Justice
20 209 South LaSalle Street, Suite 600
Chicago, IL 60604
21

22 For Plaintiffs
State of Colorado &
23 State of Nebraska: WILLIAM CAVANAUGH, JR.
Patterson, Belknap, Webb & Tyler, LLP
24 1133 Avenue of the Americas #2200
Suite 2200
25 New York, NY 10036
 3632

1 APPEARANCES CONT:

2 For Plaintiff
State of Colorado: JONATHAN SALLET
3 Colorado Department of Law
CPS/Antitrust Section
4 1300 Broadway, 7th Floor
Denver, CO 80203
5

6
For Defendant Google: JOHN SCHMIDTLEIN
7 KENNETH SMURZYNSKI
Williams & Connolly, LLP
8 680 Maine Avenue, SW
Washington, DC 20024
9

10
For Non-Party MATTHEW LARRABEE
11 Petitioner Microsoft: Dechert, LLP
1095 Avenue of The Americas
12 27th Floor
New York, NY 10036
13

14

15

16

17

18

19

20

21

22

23 Court Reporter: JEFF HOOK

Official Court Reporter
24 U.S. District & Bankruptcy Courts
333 Constitution Avenue, NW
25 Washington, DC 20001
 3633

1 I N D E X

2 WITNESS PAGE

3 SATYA NADELLA

4 Continued Cross-Examination by Mr. Schmidtlein 3634

5 Redirect Examination by Ms. Severt 3661

7 SRIDHAR RAMASWAMY

8 Direct Examination by Mr. Dintzer 3666

9 Direct Examination by Mr. Cavanaugh 3715

10 Cross-Examination by Mr. Smurzynski 3730

11

12

13 E X H I B I T S

14 EXHIBIT PAGE

15 Exhibit DX680 ......Admitted into evidence...... 3637

16 Exhibit DX524 ......Admitted into evidence...... 3637

17 Exhibit DX673 ......Admitted into evidence...... 3648

18 Exhibit DX501 ......Admitted into evidence...... 3653

19 Exhibit UPX940 ......Admitted into evidence...... 3683

20 Exhibit UPX715 ......Admitted into evidence...... 3746

21

22

23

24

25
 3634

1 P R O C E E D I N G S

2 THE COURT: Mr. Schmidtlein, I'm ready when you are.

3 CONTINUED CROSS-EXAMINATION OF SATYA NADELLA

4 BY MR. SCHMIDTLEIN:

5 Q. Thank you, Your Honor.

6 Now, Mr. Nadella, before the lunch break, we were

7 discussing the new Bing and ChatGPT. I just want to mop up a

8 few questions on that.

9 A. Sure.

10 Q. Since the introduction of the new Bing, which

11 incorporates some of OpenAI's ChatGPT technology, Bing has

12 grown to have more than 100 (sic) daily active users, does that

13 sound right?

14 A. A hundred million, it was roughly around that same

15 number. So what we're tracking is more the Bing Chat usage of

16 that hundred million. But yes, we do have a hundred million

17 daily active.

18 Q. And you believe that since the introduction of the new

19 Bing, Bing is gaining new mobile users; is that right?

20 A. Mobile users on a relative basis, yes, but not enough.

21 And even on the desktop, nothing has fundamentally changed.

22 But we feel confident that we can continue to make the progress

23 that we've been making over the last multiple years.

24 Q. Daily installs of the Bing app quadrupled after the

25 introduction?
 3635

1 A. In the initial period, yeah.

2 Q. And at least initially, the Bing app's popularity was

3 certainly elevated; is that fair?

4 A. That's correct.

5 Q. Press accounts indicated that it was the third most

6 popular app in the Apple App Store for a time, right?

7 A. I think when it initially launched, yeah.

8 Q. And in the week following the announcement of the new

9 Bing, the Bing app received almost as many downloads as it did

10 in all of 2022, right?

11 A. If you say so.

12 Q. I mean, that comports basically with the reports you

13 were getting from your team?

14 A. Yeah. I mean, the App Store downloads are sort of

15 interesting, but not sort of something that you write home

16 about.

17 Q. Right. I mean, people download lots of apps,

18 sometimes they try them and they continue to use them and other

19 times they don't, right?

20 A. That's one. And then also, on a relative basis, the

21 install base that downloads is a small number each time. But

22 if you sort of were a number one app for like a period of a

23 year, then yeah, it's a completely different game.

24 Q. And users have found the ChatGPT app, too?

25 A. That's correct.
 3636

1 Q. ChatGPT has managed to accumulate 10,000,000 daily

2 users in just 40 days, right?

3 A. I don't have those facts, but yeah.

4 Q. I mean, as a -- well, I don't know if investor is the

5 right word, but as somebody who has invested a substantial

6 amount of money in ChatGPT and OpenAI, you monitor how they are

7 performing?

8 A. Well, we don't have rights to monitor any of their

9 commercial activity, and we don't exercise that.

10 Q. ChatGPT released an iOS app earlier this year, right?

11 A. Yep.

12 Q. And that app has received over half a million

13 downloads in just six days, right?

14 A. If you say so.

15 Q. Bing is the default search engine in the ChatGPT app,

16 correct?

17 A. That's correct.

18 Q. Now, I want to show you -- if you'll turn in your

19 binder to DX680. Are you there?

20 A. Yeah.

21 Q. Okay. Now, DX680 is an e-mail exchange between you --

22 it initially started off as an e-mail exchange between you and

23 Mr. Gates and Kevin Scott; is that correct?

24 A. That's correct.

25 Q. And can you remind the Court, who is Kevin Scott?

 3637

1 A. He's our CTO.

2 Q. Chief technology officer?

3 A. That's right.

4 Q. And am I correct that he was the CTO in 2019?

5 A. Yes, he was.

6 Q. Now, again -- well, Your Honor, I would move into

7 evidence DX680.

8 MR. SEVERT: No objection.

9 THE COURT: 680 will be admitted.

10 (Exhibit DX680 admitted into evidence)

11 MR. SCHMIDTLEIN: Your Honor, I'm told I failed to move in

12 DX524, so I'd move that into evidence as well.

13 MR. SEVERT: No objection on that one.

14 THE COURT: 524 will be admitted, counsel.

15 (Exhibit DX524 admitted into evidence)

16 BY MR. SCHMIDTLEIN:

17 Q. Now, Mr. Nadella, I am -- your counsel has designated

18 this document confidential, and so we're not going to put it on

19 the screen, and I am not going to ask you about sensitive

20 areas -- or at least I'm going to try not to ask you about

21 things that would require a sensitive response.

22 But this is an e-mail exchange about -- at a high level,

23 about Microsoft's investment in OpenAI; is that fair?

24 A. That's correct.

25 Q. And I want to focus you on Mr. Scott's e-mail on the

 3638

1 first page. Do you see that?

2 A. Yes.

3 Q. Okay. And he makes reference in this e-mail to OpenAI

4 and something called DeepMind and Google Brain?

5 A. Yes.

6 Q. What is DeepMind and Google Brain?

7 A. They're the two groups that were working on Google's

8 AI efforts.

9 Q. But both of those are Google --

10 A. That's correct.

11 Q. -- AI technology development groups, right, owned and

12 funded by Google, right?

13 A. That's correct.

14 Q. And is it fair to say that Mr. Scott is discussing the

15 competition that Microsoft faces with Google regarding AI

16 technologies?

17 A. That's correct.

18 Q. And if you will go down in the second paragraph of his

19 e-mail, and is it fair to say that Mr. Scott in here is

20 describing -- again, at a high level, his initial views of the

21 significance of the technological developments that Google's

22 DeepMind and Google Brain had been engaged in?

23 A. Sorry, can you repeat that?

24 Q. Sure. If you go to the question -- or the sentence

25 that says: "I has highly" -- and I think he meant -- I think

 3639

1 he was trying to say "I was highly." Do you see that? It's

2 about the fourth or fifth line from the bottom of the third

3 paragraph.

4 A. Third paragraph, okay.

5 Q. The one that starts: "The thing that's interesting."

6 A. Okay, got it. Yeah.

7 Q. And there he's expressing at least some of his initial

8 reactions to the AI technology efforts that Google had been

9 engaged in; is that fair?

10 A. Yeah, I mean, I think he was making a comment on this

11 generation of AI technology and how promising is it or not,

12 yeah.

13 Q. And his initial -- he had some initial reactions, but

14 he then subsequently came to believe that they were more

15 significant; is that fair?

16 A. That's correct.

17 Q. And then the next sentence that says that, and then we

18 go to the next one, he says: "When they took all of the

19 infrastructure," and then he goes on there. Again, Mr. Scott

20 is sort of explaining -- and in some respects justifying, why

21 it was important for Microsoft to make an investment in OpenAI?

22 A. That's correct.

23 Q. Because Mr. Scott knew and understood that there was

24 already very significant AI competition from Google?

25 A. That's correct.
 3640

1 Q. And that competition was going to affect, and already

2 had been affecting, search, correct?

3 A. You know, this particular investment was not made with

4 the narrow focus on just search. I think we have a broader

5 thing at stake, which is like mobile or cloud, where AI is

6 going to affect every software category. One such category is

7 going to be search.

8 Q. Right. And later on in this e-mail, Mr. Scott goes on

9 at some length to discuss the financial upside that supporting

10 OpenAI would have for Microsoft's cloud business, right?

11 A. Yeah, he does. And I think -- I've not read the

12 entire e-mail now, but the fundamental thing about AI, at least

13 as I look at it -- and that's one of the reasons why we want to

14 prioritize search, is the economics of search is one of the

15 most attractive areas to apply this AI innovation. So

16 therefore, yes, Google obviously has an inherent advantage, but

17 we want to compete.

18 Q. And the last sentence of that paragraph, when he

19 reads: "And as I dug into it to try to understand," can you

20 just read that to yourself.

21 (Witness reads document)

22 A. Yeah.

23 Q. So, again, this is Mr. Scott confirming that, in 2019,

24 it was critical that Microsoft make sure that it not fall

25 further behind in AI technology to Google?

 3641

1 A. That's correct.

2 Q. Now, earlier today, you had given some testimony about

3 Microsoft's agreement with Yahoo! in I believe it was 2009?

4 A. That's correct.

5 Q. And, at that time, that arrangement -- which it wasn't

6 a merger, but it was an arrangement whereby Yahoo! would send

7 all of its search queries to Microsoft, and Bing would return

8 search results to Yahoo!; is that fair?

9 A. Right, essentially we were white labeling Bing for

10 Yahoo!. So it is not as much about them transferring over to

11 Bing, but it was about providing Bing results or Bing index to

12 Yahoo! to run their own search experience.

13 Q. It would -- the end user consumer product would still

14 be labeled Yahoo! Search?

15 A. That is correct.

16 Q. But the search results that were going to be displayed

17 were at least significantly influenced by --

18 A. That's correct.

19 Q. -- Bing search results that were provided?

20 A. It's the same underlying technology, yeah.

21 Q. Because Yahoo -- in 2009, as part of that deal, Yahoo!

22 stopped crawling the web, right?

23 A. That's correct.

24 Q. Yahoo! stopped investing in search technologies,

25 correct?
 3642

1 A. That's correct.

2 Q. The idea was they were going to get out of sort of the

3 search infrastructure business, let Microsoft do that, and then

4 you would send them the search results, right?

5 A. That's correct.

6 Q. And you had a revenue share around the search

7 advertising that would come with those search results; is that

8 fair?

9 A. That's correct.

10 Q. I know it's a complicated deal, but at a high level,

11 is that right?

12 A. That's correct.

13 Q. Again, so 2009 up until today, Bing has enjoyed

14 getting all of the search queries that people enter into Yahoo!

15 search, right?

16 A. That's correct.

17 Q. And you've enjoyed all of that additional scale as --

18 A. That's correct.

19 Q. -- part of Bing's development efforts?

20 A. That's correct.

21 Q. You get to see every Yahoo! search query, right?

22 A. The dynamic data, yeah.

23 Q. You get to see every click?

24 A. That's correct.

25 Q. And notwithstanding having this arrangement from 2009

 3643

1 to the present, it hasn't really moved the needle in terms of

2 Bing's ability to get -- make significant inroads into search

3 market share?

4 A. I mean, all up on the desktop side, I would say we've

5 been able to hold on. As you rightly pointed out, Google is

6 very, very strong and has built a great product which has got

7 great penetration. So the fact that we have even a credible

8 desktop search competitor is thanks to some of the scale that

9 we were able to build up with Yahoo!.

10 Q. At the time -- do you remember -- and I know this is

11 unfair asking these memory tests going way back in time. Do

12 you remember roughly what the market share that Bing had and

13 Yahoo! had at the time of the deal in --

14 A. I don't recall, but I'm assuming they had more --

15 maybe 20 percent, and we were 10 percent perhaps.

16 Q. Okay. So as a result of that deal, Bing from a scale

17 perspective was in roughly 30 percent plus or minus in terms of

18 market share?

19 A. That's right.

20 Q. And today, Bing has teens market share on Windows,

21 right?

22 A. That's just our owned and operated. So if you add up

23 I think Yahoo!, it will be perhaps in the 20s.

24 Q. Right. Since the time of this arrangement with Yahoo!

25 in 2009, Bing has taken share from Yahoo!, right?

 3644

1 A. You can sort of say most of it was share shift within

2 Bing and Yahoo!, yeah.

3 Q. In other words, if you look at the combined share of

4 Bing and Yahoo! in '09, and the combined share of Bing and

5 Yahoo! today, today's number is lower?

6 A. That's right.

7 Q. And I believe you testified earlier today that

8 Microsoft's Bing is a profitable standalone product?

9 A. That's correct.

10 Q. And making Bing profitable, I think you testified, was

11 one of the first priorities that you made when you became CEO?

12 A. That's correct.

13 Q. And Bing has been a profitable multibillion-dollar

14 business for Microsoft for numerous years, right?

15 A. Yeah, I would say maybe 2015 or '16, I don't remember

16 the exact crossover. But, yes, for a substantial number of

17 years now.

18 Q. Microsoft has made billions of dollars in profits on

19 its search business since the 2015-16 time period?

20 A. Yeah, it's marginally profitable. I wouldn't say

21 billions of dollars.

22 Q. I wish my law firm could talk about marginal

23 profitability in terms of billions of dollars, but that's

24 another question.

25 A. You should get into the search business.

 3645

1 Q. I can't do that with a law school degree. I'm going

2 to talk about Bing's -- who Bing views as its competitors.

3 Obviously you've talked about Google. We've talked a little

4 about Yahoo! as another general search engine. Bing competes

5 with Yahoo! today?

6 A. Yeah, I mean, in the broader general search market,

7 they're different brands, yes, and do compete.

8 Q. But Bing doesn't only compete against general search

9 engines, does it?

10 A. Predominant -- I mean, our focus on why -- going back

11 to your question about why we continue to invest in Bing, why

12 are we interested in Bing. It's for general purpose search.

13 Q. Can you just give me a rough range of what percentage

14 of search queries that Bing receives Bing shows ads against?

15 A. Oh, that's --

16 Q. And I don't think that's -- I assume that's not a

17 highly confidential --

18 A. I don't know, it's a coverage metric. I don't have

19 the -- it's sort of a little beyond my current day-to-day sort

20 of feel for exactly what the coverage would be. But it's -- at

21 the most, the coverages maybe are 20, 30 percent maybe.

22 Q. It's a minority of the queries?

23 A. Yeah.

24 Q. And for that minority percentage of queries -- that

25 sometimes we refer to as commercial queries, you're familiar

 3646

1 with that term?

2 A. That's correct, yep.

3 Q. Bing competes heavily against both general search

4 engines and vertical search engines, for example, Amazon?

5 A. Correct. I mean, the commercial -- I mean, the search

6 economics is all about gaining share in those commercial query

7 domains. Of course, you can't gain share in just commercial

8 query domains without doing a good job with all the others.

9 And, of course, vertical search engines also compete with you

10 in two forms. One is to basically get to SEO ranking as well

11 as they advertise on the general search engine.

12 Q. And Microsoft has recognized for many years that for

13 some of these commercial queries, users can and will begin

14 their searches on the vertical search engines rather than at a

15 Bing or another type of general search engine, right?

16 A. The only one who's had, I would say, sizable success

17 in that is probably Amazon, but mostly others have not. But

18 the vertical search engines still fundamentally are dependent

19 on SEO traffic like we are. Like even a social network like

20 LinkedIn is predominantly SEO traffic.

21 Q. Bing and Microsoft have looked at and analyzed the

22 issue of how much search is starting on vertical sites versus

23 Bing, correct?

24 A. We have over the years. In fact, that was one of the

25 hopes was will search get a little more fragmented in the sense
 3647

1 will there be other entry points. And I would -- at least my

2 data tells me, and over time -- in fact, I would love to see

3 such a structure. But other than Amazon's success, I've not

4 seen much.

5 Q. Microsoft has recognized that people also have begun

6 going to social media sites to search, too, haven't they?

7 A. Again, we tried long back -- and I forget the exact

8 year. Even with Facebook, we did a big sort of search campaign

9 with them with Bing, and that didn't go anywhere. So, again,

10 at least up to now, we've not yet seen any major change in

11 search behavior. General search remains still the predominant

12 driver of the search market.

13 Q. If you'll take a look in your binder at DX673. This

14 is a 2010 e-mail from Justin Jed to you. Do you know who

15 Mr. Jed is?

16 A. I just can't recall at this very moment, but I'm

17 assuming he's in our business development team.

18 Q. And the e-mail was to you and an Erik Jorgensen?

19 A. Yep.

20 Q. Who is Mr. Jorgensen?

21 A. He ran our -- at that point, I think mobile and local

22 search and maps, I think.

23 MR. SCHMIDTLEIN: Your Honor, we would move into evidence

24 DX673.

25 MR. SEVERT: No objection.

 3648

1 THE COURT: It will be admitted.

2 (Exhibit DX673 admitted into evidence)

3 BY MR. SCHMIDTLEIN:

4 Q. Now, Mr. Nadella, this is another document that your

5 counsel has marked confidential in full, so I'm going to try to

6 navigate around it. Hopefully you can help me do that. If you

7 look at the page, it is -- the DX number is 015. It's 14 on

8 the document, but it's 015 on the number. And this is in a

9 report that talks about consumer insights and implications.

10 One of the insights here is that search is becoming

11 susceptible to competition with vertical search sites, correct?

12 A. Yeah, I think in 2010 -- that's when this document is

13 from I'm assuming, that was one of the thesis or theories or

14 hopes we had.

15 Q. Well, it doesn't talk about hopes, there's a line in

16 there -- the second paragraph says: "Today, vertical domain

17 specialization" -- and it talks about in the past, talking

18 about the emergence -- has led to the emergence of, and then it

19 talks about a variety of vertical search competitors, right?

20 A. Yeah, it talks about sort of a lot of the vertical

21 search folks who have -- I mean, the reality of vertical search

22 unfortunately is that none of them have become -- other than

23 perhaps in travel, which is a little bit of metasearch.

24 Economically, they come nowhere close to the success even of

25 Bing with its low share. I mean, think about it this way:
 3649

1 Bing, with its meager share, is a bigger business than most of

2 these vertical search engines put together.

3 Q. Well, sir, Bing is attempting to deal with all search

4 queries, and each of these is attempting to deal with a narrow

5 slice of the search queries, right?

6 A. That's correct.

7 Q. And each of these vertical search engines also has

8 advertising on them, correct?

9 A. That is correct.

10 Q. They're competing to get the same advertiser to

11 advertise on their sites that you're trying to get to advertise

12 on Bing?

13 A. Again, the ad auction that is there in AdCentral or

14 AdWords is very different for those commercial domains. In

15 many cases, the most active advertisers on Bing or in Google

16 will be vertical search engines.

17 Q. I understand the vertical search engines also

18 advertise on general search engines, I understand that. But

19 other -- the people who are advertising on the local search

20 providers or a local or an Amazon or people like that, they're

21 also advertising on Bing and Google?

22 A. That's correct.

23 Q. And you understand that if you -- if Bing does a lousy

24 job of answering vertical search queries, people will start

25 going and searching on Amazon directly?

 3650

1 A. Yeah, I mean, Amazon's a unique case in this. But,

2 yes, if Microsoft -- if Bing or Google were not doing vertical

3 searches well, or at least not having organic results that

4 people could click to get to vertical search engines, yeah. In

5 fact, vertical search engines do a good job of SEO

6 optimization.

7 Q. Vertical search engines have, for many years, been

8 getting free traffic from Google and Bing, right?

9 A. Yeah, that's sort of the fundamental point of our SEO.

10 Q. And if you look at the next to last sentence in

11 this -- on the paragraph that began: "Today, vertical domain

12 specialization," the one that begins: "In these categories" --

13 A. This is in 015?

14 Q. Yes, sir. If you look at the -- the paragraph that

15 starts "today."

16 A. Yeah.

17 Q. And if you go to the next to last sentence that starts

18 with "in these categories."

19 A. Yes.

20 Q. And the categories that are being referred to here are

21 things like local search, shopping, social, things like that,

22 correct?

23 A. That's correct.

24 Q. And here, what's being reported is that consumers are

25 starting to bypass general search engines?

 3651

1 A. Yeah. And, again, you're a hundred percent right that

2 these are the other places people go to. But they have not

3 been substitutes to general search engines. Social sites

4 people go for social, people go for e-commerce to commerce

5 sites, and they've not been somehow taking away from what

6 happens in general search as a category and its growth.

7 Q. You're not aware of people searching for products or

8 searching for other things on Facebook?

9 A. I mean, there are different modalities. It's like in

10 Instagram, they have e-commerce or what have you, but it's not

11 the same thing as general search.

12 Q. People aren't searching for products on TikTok?

13 A. Just very different products.

14 Q. Are you aware that TikTok has launched a search

15 advertising business?

16 A. I'm not aware of TikTok's search advertising business,

17 but I'm aware of TikTok having searches inside of TikTok which

18 is a different auction, different market. For example, I'm not

19 interested in competing with TikTok.

20 Q. Well, we'll see how that goes. Is it your testimony

21 that Microsoft's advertising business -- well, let me strike

22 that.

23 You would agree Microsoft's advertising business competes

24 with Facebook?

25 A. Yeah, again, broad advertising business, very

 3652

1 different. I'm talking about search advertising auction has

2 nothing to do with the auction on TikTok or auction on any

3 other social site.

4 Q. You don't believe the same advertisers that are

5 participating in general search engines are also --

6 A. In general --

7 Q. Let me just finish my question, then I'll give you a

8 chance to answer. It's your testimony that advertisers who are

9 participating in Google search or Bing search ad auctions are

10 not moving campaigns back and forth with -- to Facebook or

11 Instagram or Amazon?

12 A. I would say predominantly advertisers are doing more

13 digital, and in that context they're moving mostly offline to

14 online. And they look at all of these things, and with very

15 different campaign objectives.

16 Q. If you'll look in your binder at DX501.

17 A. 501?

18 Q. Yes. Oh, sorry, it's not in your binder, sorry about

19 that. I'll hand it up to you.

20 Your Honor, we'd move the admission of DX501.

21 MR. SEVERT: Objection on that. There's a number of

22 third-party sources throughout the document, Amazon.

23 THE COURT: It looks like it's a market evaluation done by

24 Microsoft.

25 MR. SEVERT: It says what it says, but there's --

 3653

1 THE COURT: Okay, we'll admit it. I understand what it's

2 built for, and probably based upon public source reporting, so

3 go ahead.

4 (Exhibit DX501 admitted into evidence)

5 BY MR. SCHMIDTLEIN:

6 Q. Mr. Nadella, this is a cover e-mail that circulates a

7 presentation from December 2018. Do you see that?

8 A. Yes, I do.

9 Q. And you're not copied on this, but this appears to be

10 circulated to a group of Microsoft executives, including people

11 involved with the Bing search engine, correct?

12 A. That's correct.

13 Q. And the subject of this market assessment has to do

14 with Amazon's advertising business, right?

15 A. That's correct.

16 Q. And people within Microsoft do types of analyses like

17 this on a fairly regular basis, right?

18 A. Yeah, sure.

19 Q. I mean, Microsoft is, again, analyzing and trying to

20 keep up with the competition that Microsoft faces in search

21 advertising, right?

22 A. That's correct.

23 Q. And this is just one example of the type of

24 competitive analysis that Microsoft does, this one having to do

25 with Amazon, right?

 3654

1 A. That's correct.

2 Q. And Microsoft has similar types of documents like this

3 for Facebook or Instagram or others in the digital advertising

4 space, correct?

5 A. Sure.

6 Q. And you're aware that in the digital advertising

7 space, Amazon has grown to become the third largest advertiser

8 in all of digital advertising, right?

9 A. That's right.

10 Q. And Facebook is number two?

11 A. That's correct.

12 Q. And Bing trails both of them, right?

13 A. That's right.

14 Q. You testified earlier today about some conversations

15 Microsoft has had with Apple over time?

16 A. Yes.

17 Q. You recall that. Now, since May of 2010, Microsoft

18 has had a search distribution agreement with Apple, correct?

19 A. We had, for a period of time, some search deal around

20 I think Spotlight and Siri backfill. I forget the exact dates,

21 but I think it went away when they renewed with Google in 2018,

22 I think.

23 Q. Microsoft, though, today has an agreement in

24 connection with Microsoft being provided as an alternative to

25 be selected as the default search engine?

 3655

1 A. Oh, so in the dropdown in Safari, yeah.

2 Q. Bookmarks and things like that, correct?

3 A. I'm assuming, yeah.

4 Q. And under that agreement, Apple provides Bing a

5 readily available and easily discoverable option for the user

6 to change the user's default search engine?

7 A. Yeah, I mean, we can contest, you know, how easy

8 defaults are to change. But, yeah, we are one of the

9 alternatives, but we are not the default.

10 Q. You've been an option to change the default on Macs,

11 iPhones and iPads since 2010?

12 A. Sure.

13 Q. And is it fair to say that a very, very low percentage

14 of Apple users have chosen to search using Bing on Apple

15 devices?

16 A. That's correct.

17 Q. And even though there's been no appreciable demand for

18 Bing on Apple devices, Microsoft has tried to engage Apple on a

19 number of occasions to persuade Apple to change the default on

20 Safari from Google to Bing?

21 A. Yeah. And, as I said, we also had some modicum of

22 success with these alternative entry points like Siri and

23 Spotlight.

24 Q. And one topic that arose frequently during these

25 conversations with Apple has been Bing's search quality, right?

 3656

1 A. That's right.

2 Q. And Apple executives have raised concerns about Bing's

3 search quality, right?

4 A. That's correct.

5 Q. And they've raised concerns about Bing's brand with

6 consumers, right?

7 A. That's correct.

8 Q. And I take it Microsoft has never been able to

9 persuade Apple to switch the default on Safari from Google to

10 Bing?

11 A. That's correct.

12 Q. I assume you're going to keep on trying?

13 A. We will.

14 Q. I want to go back and just ask you about DX502. That

15 was the single document that Mr. Severt asked you about.

16 MR. LARRABEE: Your Honor, sorry to interrupt. If there's

17 anything the Court can do to help the parties finish with

18 Mr. Nadella by 2:30...

19 (Inaudible)

20 MR. SCHMIDTLEIN: Two more questions.

21 THE COURT: Okay, two more questions.

22 THE WITNESS: 502?

23 THE COURT: Yes, 502.

24 MR. SCHMIDTLEIN: This is the --

25 THE COURT: I don't think it's in the binder.

 3657

1 MR. SCHMIDTLEIN: No, this was the spare one that they

2 used. If you need another one, I'm happy to --

3 THE WITNESS: Oh, is this the same as 736?

4 THE COURT: It's the same one, it appears to be. It's the

5 same as 736.

6 THE WITNESS: 736, yep.

7 BY MR. SCHMIDTLEIN:

8 Q. My apologies.

9 A. No problem.

10 Q. And you were asked some questions -- and again, this

11 is a document that has been marked highly confidential, so I

12 want to be very careful and respectful here.

13 A. Yes.

14 Q. But I believe you were asked some questions, and there

15 was some at least recognition of some of the language that I

16 think you recognized wasn't overly sensitive with respect to

17 Windows is absolutely open. Do you remember that?

18 A. Yes.

19 Q. And it then goes on to say: "Our OEMs are all free to

20 sell themselves to Google"?

21 A. Yep.

22 Q. "And yet we have a 24 percent O&O share." And that's

23 owned and operating, right?

24 A. That's right.

25 Q. And there you're communicating that you believe, at

 3658

1 least as of this time, Bing's share on Windows is roughly

2 24 percent?

3 A. That's right.

4 Q. And although you say: "OEMs are free to sell

5 themselves to Google," Microsoft was selling itself to Google,

6 right?

7 A. Selling itself to the OEMs you mean?

8 Q. To the OEMs, right?

9 A. Yes.

10 Q. Not to Google, that would be interesting.

11 A. We'd be happy to.

12 Q. Yes, well, I think they'd -- I think we'd have another

13 problem.

14 So you say here: "OEMs are free to sell themselves to

15 Google." You've got a 24 percent share on the device, and

16 you've got a hundred percent distribution share on Windows,

17 right?

18 A. Yep.

19 Q. So you thought that the fact that with a hundred

20 percent distribution on Windows, and a 24 percent share for

21 Bing, that was a lead selling point to persuade Apple that it

22 would be a great idea to switch the Safari browser from Google

23 to Bing?

24 A. So the main argument I was making is the fact that on

25 an open platform, we have a search product that can hold -- the

 3659

1 thing that was debated was can we hold on to defaults. Even

2 let's say I had a hundred percent. I have 24 percent on an

3 open platform. So the point I was making to Apple -- which, by

4 the way, that is the only reason why they kept engaging, is

5 with the Apple brand, it was not going to be called Bing. We

6 had all kinds of strategic flexibility. It was going to be

7 just like Apple Maps, that was the idea. This was not about

8 trying to put Bing and Bing brand front and center.

9 Q. Oh, I see, you were going to hide the Bing brand?

10 A. Yeah, of course. We were going to take whatever Apple

11 felt was their chance as a success with the technology.

12 Q. So you were going to try to hide the Bing brand behind

13 the Apple brand and fool all the Google users into staying with

14 Bing instead of switching back to Google, which is what they

15 did on Windows in droves, right?

16 A. That is not what I said.

17 THE COURT: Hang on, hang on, I'll sustain the objection.

18 I think it's a little argumentative.

19 BY MR. SCHMIDTLEIN:

20 Q. You're saying here a hundred percent open platform.

21 It was a hundred percent controlled with Bing defaults?

22 A. And Google has the ability to deliver their payload,

23 and so, yes, they want share on it. So, to me, us being able

24 to go to Apple and say hey, just look at our technology

25 independent of our brand and make your choice.

 3660

1 Q. The Apple Maps analogy --

2 THE COURT: Why don't we --

3 MR. SCHMIDTLEIN: Can I just have --

4 THE COURT: One more question.

5 MR. SCHMIDTLEIN: I mean, he's talked about Apple Maps.

6 THE COURT: No, I know. One more question. We're sort of

7 treading the same territory we have already.

8 BY MR. SCHMIDTLEIN:

9 Q. The Apple Maps incident happened in 2012, correct?

10 A. That's correct.

11 Q. Tim Cook issued a public letter apology to Apple users

12 about that incident, right?

13 A. That's correct.

14 Q. It was a huge embarrassment to Apple to have switched

15 from Google Maps to Apple Maps when it was an inferior product?

16 A. Okay.

17 Q. You don't disagree with that?

18 A. Look, I think it worked out beautifully as a strategy

19 for them, and I congratulate them on that.

20 Q. And that's the strategy you had proposed to Apple,

21 that they suffer through that embarrassment with Bing to get

22 through --

23 A. That's not what I said. I said that they have -- if

24 they wanted to have a second supplier and have competition in

25 search and innovation in search -- which they wanted to take

 3661

1 some special angles around privacy, we'd be willing to do the

2 deal.

3 Q. Apple wasn't willing to make that --

4 THE COURT: I think we know what Apple did or didn't

5 decide. Let's move on.

6 MR. SCHMIDTLEIN: Thank you, Your Honor.

7 THE COURT: Thank you, Mr. Schmidtlein. Let's get to the

8 redirect.

9 REDIRECT EXAMINATION OF SATYA NADELLA

10 BY MR. SEVERT:

11 Q. Thank you, Your Honor. Mr. Nadella, in addition to a

12 general search engine, does Bing have specialized vertical

13 search engines?

14 A. We do have vertical scopes like Shopping inside and

15 Bing Travel inside of Bing. But they are more about sort of, I

16 would call it, just general search with vertical scopes. And

17 we know that we need to do search transfers to vertical search

18 engines as well.

19 Q. And does Bing compete with vertical search engines for

20 default placement in, say, web browsers?

21 A. Not really.

22 Q. And earlier you were asked about AI and unusual

23 behavior and hallucinations. Do you recall that?

24 A. Yes.

25 Q. And is that feature or occurrence, is that unique to

 3662

1 Bing Chat?

2 A. No, not that I know. I think both when Bing came out

3 and Bard came out, both had hallucinations. Sometimes

4 hallucinations are actual features, it's called creativity, and

5 sometimes it's a bug. And we're all trying to improve how to

6 get better at this on both sides.

7 Q. And I know there was sort of some initial interest in

8 ChatGPT, Bing Chat when it was released. But has the

9 introduction of Bing Chat resulted in a substantial increase in

10 Bing's market share?

11 A. No.

12 Q. I want to talk to you about -- you were asked about

13 investments in counsel's examination. Do you recall that?

14 A. Yes, I do recall that.

15 Q. And if we look across Microsoft's various businesses,

16 would it be fair to say that there are certain products where

17 Microsoft has been very successful?

18 A. I would say so, yes.

19 Q. And if we were to take a close look at even the

20 successful products, would we see Microsoft made mistakes or

21 misjudgments in those products from time to time?

22 A. Many.

23 Q. And is there any product within Microsoft where the

24 company just makes every investment that someone might suggest?

25 A. Never.
 3663

1 Q. Why not?

2 A. That would be silly to do. I mean, like the point is

3 management teams exist to make judgment calls on what

4 investments make sense. You will get lots of proposals, and

5 nine out of 10 them you reject, and you pick the ones that

6 matter.

7 Q. And then you were asked about mobile -- search on

8 mobile in the 2008 to 2013 time period. Do you recall that?

9 A. Yes, absolutely.

10 Q. Is the mobile industry different today than it was

11 back then?

12 A. Very much so.

13 Q. How so?

14 A. It's evolved to a place where the mobile phone itself,

15 both because of the connectivity, the form factor, it's the

16 main computing device. And in the early sort of 2010s it was

17 still growing into that position, whereas now it is the

18 dominant platform.

19 Q. And did Microsoft or Bing have any more mobile scale

20 back then than it does today?

21 A. Back then, in the way the market was -- it's kind of

22 like taking a very small market at that time, and yes, we had

23 high share of a very small market called mobile at one point.

24 And it's when the mobile market became the dominant one we had

25 very low share or zero share.

 3664

1 Q. Sure. And then in more recent discussions with Apple

2 about the Safari default, was there a premise that Bing could

3 hold on to more users than we saw in those older documents?

4 A. Absolutely, that was -- the entire case being made and

5 their evaluation was all around that.

6 Q. Was Microsoft willing to bet billions of dollars on

7 that proposition?

8 A. Yes.

9 Q. And if you were wrong and Google would end up with

10 most of the Safari queries, even with Bing as the default, do

11 you have any explanation for why Google pays Apple so much

12 money?

13 A. That's a great question.

14 MR. SCHMIDTLEIN: Objection.

15 THE WITNESS: I mean, I don't know why anybody would

16 pay if you are confident --

17 THE COURT: I'm sorry, there's an objection. He can opine

18 on it, that's fine. I understand he doesn't have any firsthand

19 knowledge. Go ahead, sir.

20 THE WITNESS: I mean, I think the questions that are asked

21 of me was why on Windows you don't have to pay and yet have

22 high share, even though we have the defaults for which we are

23 paying. The question then is why does Google have to pay

24 Apple. I would love an opportunity to sort of not have them

25 pay, maybe on behalf of the Google shareholders.

 3665

1 MR. SEVERT: No further questions, Your Honor.

2 THE COURT: Any redirect on behalf of the states?

3 MR. CAVANAUGH: No, Your Honor.

4 THE COURT: Mr. Nadella, thank you very much for your time

5 and your testimony, sir. Thank you for being here, and safe

6 travels home.

7 THE WITNESS: Thank you so much.

8 THE COURT: Are we ready with our next witness, once we

9 change places here?

10 MR. DINTZER: Your Honor, the DOJ plaintiffs call

11 Mr. Ramaswamy to the stand. And while he's being got, I'd like

12 to introduce the Court to his counsel. We have Thomas Mueller,

13 Nana Wilberforce and Tyler Shearer from WilmerHale.

14 THE COURT: I'm sorry, can we start again. This is

15 Mr. Ramaswamy -- oh, from Neeva?

16 MR. DINTZER: Yes, sir.

17 THE COURT: Correct, okay. And the name of counsel from

18 WilmerHale again?

19 MR. DINTZER: Thomas Mueller, Nana Wilberforce and Tyler

20 Shearer.

21 THE COURT: Mr Dintzer, I should have asked, what's the

22 expectation with respect to any closed session?

23 MR. DINTZER: I expect the whole thing to be open, Your

24 Honor.

25 THE COURT: Terrific, thank you.

 3666

1 MR. DINTZER: If I may approach -- well, actually, I've

2 got the wrong binder.

3 (Brief interruption)

4 DEPUTY CLERK: Can you please raise your right hand. Do

5 solemnly swear or affirm that the testimony you'll provide to

6 the Court will be the truth, the whole truth, and nothing but

7 the truth?

8 THE WITNESS: I do.

9 DEPUTY CLERK: Thank you.

10 THE COURT: Mr. Ramaswamy, thank you for being here.

11 Please have a seat.

12 DIRECT EXAMINATION OF SRIDHAR RAMASWAMY

13 BY MR. DINTZER:

14 Q. Sir, we met in the hall. My name is Kenneth Dintzer,

15 and I'm with the Department of Justice. If you could please

16 state and spell your name.

17 A. Sure. My name is Sridhar Ramaswamy; S-R-I-D-H-A-R,

18 R-A-M-A-S-W-A-M-Y.

19 Q. Sir, we're going to begin with your background. If

20 you could describe your education.

21 A. Yes, I grew up in India. I have a bachelor of

22 technology in computer science from IDE Madras in Chennai,

23 India. I got a masters and PhD from Brown University in

24 Providence, Rhode Island.

25 Q. And do you remain active at Brown?

 3667

1 A. I'm on the board of trustees at Brown, yes, I'm still

2 active at Brown.

3 Q. And have you ever been employed by Google?

4 A. Yes, I worked for Google between April 2003 and

5 October 2018.

6 Q. What was your first job title at Google?

7 A. I joined Google as a software engineer. I worked in

8 the search ads infrastructure team, which was then the

9 monetization engine for Google.

10 Q. And can you just briefly trace your subsequent

11 positions and responsibilities at Google?

12 A. As I said, I joined in 2003 as a software engineer. I

13 worked on the infrastructure that powered search ads, the ads

14 that get shown when people search on Google. I grew with the

15 team. I had leadership positions first with that

16 infrastructure team which powered, as I said, Google's

17 monetization. And then I gradually -- over the next three or

18 four years, took over running the advertiser interfaces team,

19 which is called the AdWords team, as well as the team that

20 works on what's called ads quality. This is the team that runs

21 the auction that essentially operates the heart of the Google

22 search advertising system.

23 I was promoted to vice president in 2008, if I recall

24 correctly. By the time I was running most of search ads. And

25 then in 2011, I was given broader responsibility as part of the

 3668

1 reorganization of Google into a more GM-like organization that

2 was then called the L Team. And then I became part of the CEO

3 Larry Page's staff in 2013, by which time I was the SVP of ads

4 and commerce. This is a job that Susan Wojcicki and I shared

5 for a year. And then in 2014, Susan went over to run YouTube,

6 and I was the sole lead for the Google ads and commerce team.

7 And by that time, this job encompassed all of search

8 advertising, all of what are called Google display products --

9 products for publishers, advertisers, as well as things like

10 Google Shopping and Payments. And then Sundar effectively

11 became first the lead for Google in 2014, and eventually the

12 CEO, so I was on his staff from 2014 through 2018. And then in

13 2016, in addition to ads and commerce, I also assumed

14 management of the search infrastructure team as well as the

15 privacy teams at Google.

16 Q. And did ad revenues change during your time at Google?

17 A. Yes. Google made most of its money early on from

18 search advertising, and I think it continues to be a majority

19 of revenue to this day. But certainly it was Google's most

20 profitable division by a long shot. We, actually in 2003

21 itself, launched a product called AdSense, which was really the

22 beginning of Google's foray into serving ads beyond its own

23 properties. And that culminated with the acquisition of

24 DoubleClick in 2007, and so Google of course added revenues

25 from other products, like showing ads on third-party sites as

 3669

1 well as sprouting both publishers and advertisers. Then came

2 specialized products on the search surface itself, things like

3 shopping and travel ads.

4 And then finally I would call out the rise of mobile

5 operating systems, mobile devices as another very, very large

6 scale change to how Google revenue, Google advertising worked.

7 And during my time there, I was also one of the folks in charge

8 of transitioning Google's advertising products from being

9 primarily desktop-oriented to a combination of desktop, mobile

10 along with multiple product lines. So a fair amount of change

11 through the years.

12 Q. And when did you leave Google?

13 A. I left Google late September, early October 2018,

14 about five years ago.

15 Q. And where did you go from there, sir?

16 A. I became a venture partner at a partnership called

17 Greylock Partners, and they're based in Menlo Park in Silicon

18 Valley. A venture partner is a part-time role, it tends to be

19 somewhat of a transitional role. At the time I left Google, I

20 wasn't sure about whether I wanted to do investing as a

21 full-time living. Soon after, I decided that what I really

22 wanted to work on was an alternative to search. I loved search

23 the product, and I thought that the way in which Google

24 operated its search, you know, didn't leave sufficient avenues

25 for exploring what a great product could be.

 3670

1 So along with two other colleagues that I had worked with

2 earlier, we launched a company called Neeva that was about

3 creating a new search engine from scratch. That started -- the

4 company was incorporated in late 2017, but it really got off

5 the ground in early 2019.

6 Q. And just so we're using the term, what is a general

7 search engine, briefly?

8 A. Yeah, I mean, it's what we would turn to when we are

9 in need, when we have like any information task generally in

10 mind. This can be the weather, this can be things around us,

11 this can be directions, you know, things that we are curious to

12 learn about, products. So a general search engine is I guess

13 best defined in contrast to a specialized search engine, which

14 is everything that you are looking for, that's the first place

15 that you can turn to. While for very specific things like you

16 want to buy soap, you're likely going to a different place.

17 But a general search engine is a place that you go to for the

18 vast majority of your information needs.

19 Q. And general search engines provide a specific

20 convenience?

21 A. Yes, they do. It's a little bit of a one-stop shop

22 for all information needs. In this day and age, you know, we

23 have lots of questions all the time.

24 Q. So we talked about Neeva, and that was a -- Neeva

25 developed a general search engine; is that right?

 3671

1 A. That is correct.

2 Q. And can you briefly walk me through the steps you took

3 to start Neeva.

4 A. The product thesis behind Neeva, as I said earlier,

5 was that we could create a better search engine. And a core

6 part of that product thesis involved not being beholden to an

7 advertiser's and ad supported search engine. So my co-founder,

8 Vivek, and I -- who I worked with him for -- I had worked with

9 him even before the start of Neeva for close to a decade. We

10 felt that there was both a strong need as well as a technical

11 insight that could create this better experience.

12 So in the first two-ish years of its existence, the way we

13 positioned Neeva to potential users and customers was that it

14 was an ads-free private search engine. Building search engines

15 is a difficult job, it's one of the most technically difficult

16 problems that there is today. And it took us two plus years to

17 build enough of the infrastructure that would be needed to

18 power this search engine. We bootstrapped on top of Bing API,

19 which is a programmatic way in which Microsoft makes Bing's

20 results available to third parties like Neeva.

21 But our sort of big breakthrough realization came in early

22 2022 where we realized that the power of generative AI models

23 could really create a much, much better search experience where

24 instead of providing a set of links, we could provide concise

25 and useful answers to many queries that users could put in.
 3672

1 And that's what we worked through through last year. So Neeva

2 transformed itself from being an ads-free private search

3 engine -- we kept all of it -- to being more of an answer

4 engine powered by AI working on top of the search

5 infrastructure that we had built in the early years of Neeva.

6 Q. And where did the funding come from for Neeva?

7 A. So Neeva had two substantial funding rounds. The

8 first funding round was in early 2019, and that round was split

9 equally between Greylock Partners, Sequoia Capital -- which is

10 another venture firm based in Silicon Valley, and myself. The

11 second funding round, which was in 2021, came from Greylock,

12 Sequoia and another venture firm called Inovia Ventures which

13 is based in Canada and the UK.

14 Q. Are you familiar with a gentleman named Udi Manber?

15 A. Yes, Udi and I were colleagues and peers at Google.

16 Udi ran the search team while at that time I was the lead for

17 search ads. That's what I mean by we were peers, we worked on

18 the same product. You know, we kept in touch, and he actually

19 worked part-time with Neeva, roughly 50 percent time, for about

20 two years. So, yes, I know Udi really well.

21 Q. And you've touched on this, but just so that we get a

22 focus. What type of search engine was envisioned by Neeva,

23 what were they aiming for in their search engine?

24 A. As I said, our thesis with Neeva was that we could

25 create a much better search experience by having a relentless

 3673

1 focus on just the users of the search engine. Being ad

2 supported imposes many natural constraints on how a search

3 engine can operate. To give you a concrete example, people

4 dislike certain sites in their search results because they

5 don't want to, say, buy products from some large companies that

6 they don't like. In a commercial search engine, it's really

7 not feasible to give options to users where you say -- where

8 they're able to say I don't want to see, take your pick,

9 xyz.com from the search results.

10 It's quite hard to personalize search even more depending

11 on the preferences that you have. When you search for terms,

12 you might -- all things being equal, you might prefer more

13 legal results say compared to some other field, even if the

14 query might pertain to either field. And so we focused a lot

15 on this kind of personalization where we could work on creating

16 a product that our users would want. And we also supported

17 features like being able to search over your personal data.

18 More and more, all of our lives are online. You know, I

19 have everything from my tax returns to copies of my passport,

20 and so on, on cloud storage, and it's often difficult to find

21 these. So we would support searching for that, searching

22 through your e-mail, all from the same interface. Generally

23 speaking, commercial search engines, which serve ads and by

24 necessity have to do a lot of work to track what users are

25 doing online to show the efficacy of advertising, are reluctant

 3674

1 to work on features like this because they end up appearing

2 quite creepy to end users. But because we had a business model

3 and a maniacal focus on what our users would want, we were free

4 to innovate.

5 And then the final point is that, as I said, for us AI was

6 a real breakthrough. And I continue to think that generative

7 AI is going to be a big game changer for search. So we could

8 switch from showing links. People only reluctantly click or

9 tap on links, but we could begin to show answers to questions

10 that people asked. So, in many ways, AI search was the

11 culmination of our vision to truly create a better search

12 experience for our users. But hopefully you begin to get

13 the -- get an idea of the kinds of things that we were able to

14 pursue by both creating a product and a business model that was

15 much more aligned with what users would like to see.

16 Q. And is Neeva still available to the public?

17 A. No. We launched AI-powered search earlier this year,

18 and we are -- as I said, we were venture funded, we had raised

19 a substantial amount of money. But Neeva also struggled with

20 user growth. And with the changes in just the macro

21 environment of things like interest rates, venture funding has

22 changed pretty substantially in the past year. My co-founder,

23 Vivek, and I came to the reluctant conclusion that we would not

24 be able to build up a business fast enough to be able to

25 continue raising capital to support the growth of the product

 3675

1 and the team. So earlier this year, in May -- we actually

2 started potential acquisition conversations in March. But

3 earlier this year, in May, we shut down the consumer search

4 engine, refunded the money that customers had paid us and got

5 acquired by Snowflake, which is an enterprise data company. So

6 no, Neeva, the consumer search engine no longer exists.

7 Q. Is it run as a public available search engine by

8 Snowflake?

9 A. It is not. As I was saying earlier, we shut down the

10 search engine first, even before the acquisition consummated.

11 And the reasoning behind that roughly was Snowflake is a

12 company that sells to other companies, it's what's called an

13 enterprise software company. And it made little sense for us

14 to have Snowflake run a consumer search engine, so we actually

15 shut down the product. Snowflake absorbed the technology that

16 we had built for searching, the technology that we had built

17 around AI into its product's offering. But no, the consumer

18 search engine no longer exists, and it's most definitely not a

19 product of Snowflake's.

20 THE COURT: Sorry to interrupt.

21 MR. DINTZER: Please.

22 THE COURT: Mr. Ramaswamy, from what I understand, one of

23 the other business propositions for Neeva was that it was a

24 subscription service?

25 THE WITNESS: That's correct.

 3676

1 THE COURT: And can you just give me a general sense of,

2 for example, how much it costs on an annual basis to purchase a

3 subscription to Neeva, and ultimately how many users you were

4 able to acquire through subscriptions?

5 THE WITNESS: Yeah, so we set up the business model of

6 Neeva as a subscription search engine, because we wanted a

7 product that was directly aligned with our customers. It's

8 what provided us the freedom, beyond the initial venture

9 funding, to be customer focused in everything that we did. But

10 free products exert a very powerful influence on all of us.

11 Most people are reluctant to pay for something that they

12 consider to be free, even if it comes with various other

13 disadvantages. So we wanted it to have affordable pricing, so

14 we priced Neeva at roughly $5 a month, or $50 a year. And the

15 majority of Neeva's customers -- I think 65, 70 percent, ended

16 up purchasing the yearly subscription for $50.

17 And through much of Neeva's existence as a search engine

18 that was generally available, we came to recognize that getting

19 people to use Neeva as their primary search engine was the key

20 unlock. Once people tried Neeva for -- you know, call it

21 dozens of queries, 50-ish queries or a few days, they were

22 likely to stick with Neeva. And people that stuck with Neeva,

23 several percentage points of them, were perfectly happy to

24 become Bing subscribers. And so our early focus was less on

25 let's maximize further number of subscribers that we got, and

 3677

1 more on -- we basically had a premium model where people could

2 try a free version of the product. So much of the early focus

3 of Neeva was really on acquiring users.

4 And then in terms of the actual numbers, we were at a

5 pretty modest place when it came to subscription revenue. It

6 was actually less than a million dollars earlier this year. It

7 was growing robustly, but it was quite small. That was all a

8 part of our calculus for where we needed to get to in a world

9 that had gone back to pricing companies as a small two digit

10 multiple of their revenue.

11 THE COURT: Okay, thank you.

12 BY MR. DINTZER:

13 Q. Sir, in the search context, are you familiar with the

14 use of the word privacy in the search context?

15 A. Yes, absolutely.

16 Q. Could you just give us a working -- I want to ask you

17 some questions, I just want to make sure we're on the same

18 page, just a short working definition?

19 A. Privacy tends to be a remarkably squishy concept. It

20 means many different things to many people. But one simple way

21 to think about it in an online context is whether a user can

22 have a reasonable expectation that their interaction with the

23 site is limited to that site. All of us draw distinctions

24 between, say, going to CNN.com or Google.com. One way to think

25 about privacy is what happens between you and a site is between

 3678

1 you and the site. And so, again, this becomes one of these

2 concepts that is sometimes easier to define by its absence.

3 And so a reasonable person can say that a lack of privacy is a

4 site knowing about everything you did everywhere. So it's like

5 a useful counterpoint. But the simplest definition that I

6 would give, in a general online interaction, is do you have a

7 reasonable expectation that there is a boundary to your

8 interaction with a given entity, with a given site.

9 Q. And was privacy a focus for Neeva?

10 A. It definitely was a focus for Neeva, because in ad

11 supported search engines like a Google, but also a Bing, what

12 happens is that it is ad supported. One of the most important

13 ways in which a search engine knows that its ads are effective

14 is by observing your behavior on other sites, especially sites

15 that you have clicked on the ads for. In other words, if you

16 do a search on Google and you go to some shopping site,

17 smallmerchant.com, Google is very interested in knowing that

18 you actually bought something on smallmerchant.com. And so

19 there's a lot of information that flows around, that flows back

20 to Google. And, as I said, it's also true of Bing. And this

21 will often manifest itself as things like ads that you see on

22 one site being affected by your behavior on other sites. More

23 generally, it leads to people having this vague sense of unease

24 about their behavior being watched, or people use words like

25 tracked. And so we wanted to create an experience that was

 3679

1 worry-free.

2 Search, as everybody here knows, is deeply personal. We

3 search for all manners of things, headaches we have, fears we

4 have, things that we want to do. And the fact that commercial

5 search experiences sort of literally look at everything that we

6 do or use the contents of our searches to show ads in other

7 unexpected places is a constant source of unease and distrust.

8 And so with Neeva, we wanted to create a service that was all

9 about hey, you search on Neeva, it sort of stays there. And,

10 in fact, the default was that there would be no search history

11 meaning that Neeva didn't have a list of the searches that you

12 did. And so Neeva was all about we show you the best results,

13 we personalize those results for you. We give you a worry-free

14 service. And, yes, we have a subscription model in which, if

15 you wanted some premium features, you'd have to pay the $50 a

16 year. It felt like a reasonable trade-off.

17 THE COURT: I'm sorry, how has Neeva been able to

18 personalize search, if it was not collecting search history?

19 How was it able to get the right signals from a user as to what

20 they would like to see?

21 THE WITNESS: So we always made our personalization very

22 explicit. We didn't rely on implicit things like search

23 history. So you are able to say hey, these are the new sources

24 that I prefer. And it's stored sort of at that level, it's not

25 stored at the level of what are the things you searched for
 3680

1 yesterday. And this way, we didn't really have to do the

2 implicit personalization that comes from oh, so and so did not

3 click on any results that were shown for a particular site, and

4 therefore we can infer something from that. And so it's really

5 like personalization is very explicit, the user is always in

6 charge. And we would also show when we did any personalization

7 so that people were always aware.

8 And it's a tricky trade-off -- you know, to your question,

9 but the promise that we always gave our users was that anything

10 we did would be on -- sort of on the user's directive. And we

11 didn't collect location history, we didn't collect search

12 history. We wanted it to be a very straight forward,

13 predictable interaction with the search engine.

14 THE COURT: And so if you didn't collect location history,

15 for example, how did that impact your ability on local search?

16 THE WITNESS: So we would get location -- for example, IP

17 addresses are a pretty predictable source of location on

18 desktop computers. And similarly, our phone scans send

19 location over when searches are being done. And so first of

20 all, we would like blur the location before we used it. But we

21 were also very careful to never log location anywhere in our

22 systems. And we use partners like Apple Maps for things like

23 local results. But we would take privacy safeguards like

24 remove the last eight bits of an IP address so you could not

25 identify somebody's computer, or we would blur out location so

 3681

1 you could -- so like these APIs could not tell where a user was

2 located. So we would operate the service with location

3 information that came over in the request, but we would never

4 log it for ourselves or record it against the user.

5 THE COURT: Thank you.

6 BY MR. DINTZER:

7 Q. And do you believe that search engine users care about

8 privacy?

9 A. Again, privacy is one of these topics -- as I said,

10 it's a little squishy to define. But people are at least --

11 for example, for certain surveys and in any casual

12 conversations that you are going to have, will express a degree

13 of unease about how much their behavior is tracked. We live in

14 a world where unless you take a fair number of precautions, and

15 sometimes painful steps, it is hard to avoid this data

16 collection. And, yes, in all of our surveys, many tens of

17 percentages of people -- definitely a majority -- have concerns

18 about a lack of privacy online, definitely with search engines.

19 But part of what we relied on was the fact that if a fraction

20 of them also expressed a willingness to pay for a commercial

21 service that was as good, if not better, than the best ones

22 that were out there for search and offered things like much

23 better privacy.

24 Q. Now, I've put a binder in front of you, sir, and it

25 has some documents in it, and they're organized by they are

 3682

1 exhibit number. I'm going to ask you if you can turn to

2 UPX940. Just let me know when you're there.

3 A. 940?

4 Q. Yes, sir.

5 A. Yes, sir.

6 Q. Do you recognize this document?

7 A. Yes.

8 Q. What is it?

9 A. So it's one of -- it's almost certainly a presentation

10 that we had made. Obviously I can't tell the date from here,

11 but if I had to guess, I would guess 2019 in terms of when this

12 was prepared.

13 MR. DINTZER: Your Honor, we'd offer it.

14 MR. SMURZYNSKI: Your Honor, we have no objection to the

15 document in toto. There's some surveys and the like embedded

16 in it that we don't have the methodology behind, et cetera, so

17 we would object to those. But I don't know how it's going to

18 be used. And that would be the limitation of our objection.

19 THE COURT: Mr. Dintzer, is that something you'll be

20 bringing to the witness' attention?

21 MR. DINTZER: Briefly, Your Honor. My understanding was

22 that the only objection they had on this document was

23 relevance, which we would suggest that this is clearly

24 relevant. We are checking --

25 MR. SMURZYNSKI: Your Honor, could I address that? That's

 3683

1 not true. We have an objection beyond relevance, and so just

2 to be clear --

3 MR. DINTZER: And we're checking this. I did not assert

4 it as fact, but as -- oh, fair enough, they have logged a

5 hearsay objection.

6 THE COURT: Let's take a look and see what is being

7 brought to the witness' attention, and figure out what it's

8 being offered for and how it can be used. So as a general

9 proposition, 940 will be admitted, and then we can discuss the

10 admissibility of any particular aspect of it.

11 (Exhibit UPX940 admitted into evidence)

12 BY MR. DINTZER:

13 Q. Thank you, Your Honor. Sir, I'm going to ask you to

14 go to the third page, and at the bottom it's Bates numbered

15 481. And just let me know when you're there. It's also going

16 to be on the screen, if that's easier.

17 A. Yeah, I'll use the screen, it's easier.

18 Q. Okay, that's great. And the heading on this page is

19 search, right?

20 A. Yes.

21 Q. And I assume this is you talking about the product

22 that Neeva is in the process of developing; is that right?

23 A. You know, on this slide, yes, this refers to search

24 the product, but it also refers to its embodiment in Google.

25 Q. And it says under bullet one: "One of the most

 3684

1 revolutionary products ever invented." What's that reference

2 to, sir?

3 A. This simply refers to the fact that, you know, every

4 web page that is out there, many of the books that humanity has

5 written is all at our fingertips. I'm old enough to remember

6 the time when I've actually gone to libraries to look at

7 encyclopedias, and certainly, compared to that, it's magical.

8 Q. And then the second bullet says: "One of the most

9 profitable businesses in history." Do you see that?

10 A. Yes, sir.

11 Q. And are you referring to a specific business in that

12 bullet?

13 A. Yeah, this is roughly Google.

14 Q. So let's go to slide four, and that's the next page.

15 It says Consumer Perception of Search. Do you see that

16 heading?

17 A. Yes, sir.

18 Q. And then below that it says: "Many internet services

19 are free to consumers but still valuable." Do you see that?

20 A. That's right.

21 Q. Could you explain that?

22 A. So this is actually an excerpt from an Economist

23 article that we can look up. As I said, I suspect that this

24 was -- this deck was part of the pitch that we showed our

25 investors. And this slide refers to a particular methodology

 3685

1 for trying to value a -- value-free products. So this article,

2 which in turn is based on a paper that was published, consists

3 of asking people for replacement value for free services that

4 they get. These are questions framed along the lines of how

5 much would you, a particular survey taker, how much would you

6 need to get paid in order to not use search or not use free

7 e-mail. And the table on the right shows their responses.

8 This is mostly meant to indicate that even though it is free, a

9 lot of users place an enormous dollar value on what they get

10 from a search engine.

11 Q. And if you could turn now to slide 14, and that's

12 Bates number 492. And, again, it will appear on the screen.

13 Just let me know when you're there, sir.

14 A. Yes.

15 Q. And the heading is User Surveys, and then it says:

16 "In the last seven days (week), which of the following search

17 engines have you used?" Do you see that?

18 A. I do.

19 Q. And the first one is Google search; is that right?

20 A. That's correct.

21 Q. And the second one is browser's default. Do you see

22 that?

23 A. That's correct.

24 Q. What did you understand that to mean, sir?

25 A. Part of the anecdotal feedback that we heard from

 3686

1 users was that they would use whatever was the default on a

2 browser, and not really pay attention to what the search engine

3 was. You know, somewhat to our dismay, we found that many

4 perfectly tech-savvy people could not tell the difference

5 between a browser and a search engine. So we devised this

6 survey to understand consumer perceptions for how people used

7 search, what they really even thought search was. And these

8 surveys, you should use them as, you know, these are

9 directional results generally speaking. The confidence

10 intervals are sort of pretty sound, because we're only looking

11 at broad things.

12 But part of what one can take away from this is that a

13 fairly large fraction of people -- this was a multi-choice

14 meaning that you could pick more than one, a fairly large

15 fraction of people simply use whatever was default in their

16 browser, and didn't really pay attention to what their search

17 engine was. This was early. This led us down the path of

18 focusing on things like creating an extension that made it easy

19 for people to switch their default search over to Neeva.

20 Q. Are you familiar with the term commercial intent?

21 A. Sure.

22 Q. And briefly, sir, what does commercial intent mean?

23 A. We are switching to a very different topic, but

24 commercial intent, in the context of a search engine, refers to

25 queries that show ads. This is a long-studied topic. The

 3687

1 funny answer for commercial intent is queries that show ads

2 have commercial intent, and commercial intent is those queries

3 that show ads. So it's basically commercial intent covers

4 queries that people are willing to pay money for. Political

5 queries become very commercially-oriented right before an

6 election, so it's a little bit of a loose term. But queries

7 that look for products or services, for example, you can

8 generally think of as having a strong commercial intent.

9 Q. And what did Neeva do if it received a query that

10 would otherwise be considered commercial intent in another

11 search engine?

12 A. So we would try and get you the best results like

13 portrayed in the best way that would satisfy your commercial

14 intent. So among the things that we did in our quality system

15 was if you looked for products, we would make sure that the

16 results were also tailored towards informing you about the

17 product. In a commercial search engine, as soon as you type in

18 a commercial query, you almost get the feeling that like, you

19 know, there's like this giant software system that jumps on you

20 and tells you where you should buy this product right this

21 instance. At Neeva, we took a more leisurely approach of let's

22 tell you about the product that you're looking for, its

23 characteristics, its features. Of course, if people wanted to

24 buy it, it was our job to show them those sites as well.

25 So, again, by virtue of being a user-focused search

 3688

1 engine, we were free to innovate. So, as I said, we focused on

2 sort of reviews a lot more. When it came to product-oriented

3 queries, we would surface information like pictures, like

4 prices in a way that was meant to be most helpful for you when

5 you were looking for products, for example.

6 Q. And if you could go to page 49 of this document, and

7 it's Bates number 527. We're still on UPX --

8 A. Yeah, I see it in front of me.

9 Q. And is this -- you were talking about sort of how

10 Neeva addressed it. Is that what we're seeing here?

11 A. These are some examples. The idea, again, is that if

12 you're looking for a product, we focused on giving you as much

13 information as you needed about the product. We used a variety

14 of techniques to extract this information and show it to you.

15 And the example on the right hand side is a travel query.

16 That's a little bit of a simple example of how whenever people

17 are looking for locations that they want to travel to, they're

18 interested in a more visual experience. We did a lot more with

19 travel queries, but travel and products are among the largest

20 commercial verticals in search engines. And having done both

21 the shopping and product team at Google, as well as the travel

22 search team, I was pretty familiar with the kinds of things

23 that users wanted in these queries as opposed to what

24 advertisers wanted. And we tried to focus Neeva very much on

25 the things that users would care about.

 3689

1 Q. No further questions on that document, sir.

2 THE COURT: Mr. Dintzer, if I could interrupt you. Why

3 don't we go ahead and take our afternoon break, if you're

4 switching topics or documents.

5 MR. DINTZER: Absolutely.

6 THE COURT: It's 10 after 3:00, we'll resume about 3:25.

7 I'll see everybody shortly.

8 Mr. Ramaswamy, I'll ask you not to discuss your testimony

9 during the break. Thank you, sir.

10 (Recess taken at 3:09 p.m.)

11 (Back on the record at 3:27 p.m.)

12 MR. DINTZER: May I proceed, Your Honor?

13 THE COURT: You may.

14 BY MR. DINTZER:

15 Q. Thank you, Your Honor. Sir, changing gears, what is

16 the most efficient way to distribute a search product?

17 A. If you're talking about a general purpose search

18 engine, at current times, having users use a browser -- which

19 they're quite used to bringing up, and being the default on the

20 browser is the most efficient way to get it into the hands of

21 your users. So the convenience of easy accessibility and

22 tapping into, you know, by now engrained default behaviors are

23 the deciding factors when it comes to whether a search engine

24 gets lots of usage.

25 Q. Is habit related to the use of default at all?

 3690

1 A. Yes -- well, to a certain extent they're synonymous,

2 right. Almost by definition, a default says that it is used

3 out-of-the-box or it is used in the ordinary course of action.

4 And so in many situations, defaults become habits.

5 Q. Was Neeva able to get default distribution with a

6 browser?

7 A. So there are nuances to this question. Users can take

8 steps, for example, in a freely available browser like a

9 Firefox or a Chrome in order to set Neeva as the default search

10 engine. For example, on Chrome desktop, they could install an

11 extension that would change the default search for Chrome to be

12 Neeva. This meant that every time they opened a new tab on

13 Chrome to search -- which is the habitual way in which people

14 access search on Chrome on desktop, they get access to Neeva.

15 But this is a complicated step, and not everybody is capable of

16 or comfortable with installing an extension.

17 The story is different on mobile. On mobile Chrome, for

18 example, you have to click on menus and follow a fairly

19 intricate set of options before you're able to set Neeva as the

20 default search. Now, a certain number of search engines --

21 some of whom -- you know, many of whom you know like Google and

22 Bing, are part of what you can call the default shipped search

23 engines. Meaning that it is easier to search -- to set Google

24 or Bing as the search engine on Chrome or Safari compared to

25 the steps that you'd have to take with Neeva. So being part of
 3691

1 that default search engine set makes it easier for people to

2 pick you as the default.

3 And, finally, in some browsers like Safari, unless you are

4 in this default set -- and as of right now, it's only Google,

5 Bing, Yahoo!, DuckDuckGo and Ecosia, if I remember correctly,

6 these are the only five available options on Safari. And it is

7 incredibly difficult for somebody to set Neeva as their default

8 search engine on Safari. Hopefully this gives you some

9 context.

10 Q. Are you aware that Google has contracts for -- to be

11 the out-of-the-box default for certain browsers?

12 A. Yes, I do. You know, we've talked about how I worked

13 for 16 years at Google, and so these contracts were very much

14 things that came my way. I was also part of what was and is

15 called the Business Council, which effectively approved all

16 contracts across the board that Google struck. So, yes, I'm

17 intimately familiar with them.

18 Q. And at Neeva, did you feel that its ability to

19 distribute its product was affected by those contracts?

20 A. This is a tricky question with not sort of clear cut

21 answers. Certainly we didn't get far in discussions with Apple

22 about being one of the defaults in their list of search engines

23 on Safari. We had several conversations with their exec staff

24 as well as folks in the marketing team that ultimately control

25 that list of search engines, but those conversations didn't go

 3692

1 anywhere. On the other hand, when we had conversations with

2 mobile carriers, many of whom -- mobile carriers are

3 effectively the buyers of Android phones. They are the ones

4 that spend the money to buy Android phones that then is sold to

5 you and me as consumers and customers. We had some early

6 promising discussions, but they would always end up petering

7 out. Anecdotal, not definitive, feedback that we got from some

8 of the execs was that the complex nature of the contracts

9 with -- you know, that they had with Google made it very, very

10 difficult for them to really offer Neeva as even an option --

11 not the default search, but a default search option on their

12 phones.

13 Q. Was there a Neeva app?

14 A. Partly because it was next to impossible to have Neeva

15 set as the default search engine on iOS on Safari, we decided

16 to go down the route of making a mobile app. You know, this

17 was not a light choice meaning that it takes a fair number of

18 engineers -- close to eight people between Android and iOS out

19 of a 50-person team, which is a substantial investment, to

20 create these apps. But startups are often forced to create

21 apps, because that is one way to get their product into the

22 hands of consumers. And post pandemic, getting people on

23 mobile devices to use Neeva was a pretty high priority, so we

24 ended up making both iOS as well as Android apps.

25 It came with some positives. Because the app is a more

 3693

1 natively controlled experience on the part of Neeva developers,

2 we could offer innovations like showing you search results as

3 you're typing into the autocomplete box. To take a small

4 digression there, you type queries into a search engine by

5 typing out a few letters. It shows you a list of completions

6 that people often pick from. And there's no reason why you

7 cannot show links to sites or the actual search result itself

8 as people are typing if you know what it is that they have in

9 mind.

10 But browsers, partly driven by commercial interests, don't

11 really make it easy for you to show actual search results as a

12 person is typing. But in an app, developers have full control

13 over the experience. So we launched something called FastTap,

14 which is basically how can we send you to the sites that you

15 want to go to, even without having to see a search page at all.

16 So we did create apps. They came with benefits, but they also

17 came with considerable burden when it came to creating and

18 maintaining them.

19 Q. Were apps a meaningful alternative channel of

20 distribution for Neeva beyond browsers?

21 A. Yes, they were a meaningful mechanism. Something like

22 40 percent of all our traffic, if I remember correctly, came

23 from mobile applications. Apps also have potential other

24 benefits even though there are trade-offs. On iOS, for

25 example, many users, especially in places like the U.S., UK,

 3694

1 Germany, they have payment instruments stored in the phone, so

2 it makes it easier for them to subscribe. But on the other

3 hand, if you want lots of installs, the makers of the operating

4 system, Apple and Google, also force you to purchase ads on

5 their platform. Yes, it was a meaningful amount of users and

6 usage overall, but again, it came with cost benefit trade-offs.

7 Q. Did you conclude that defaults would be a better means

8 of distribution?

9 A. It's not an either/or in the sense that you can get

10 somebody to install the iOS Neeva app, but they would still

11 need to remember to open the app to search rather than open

12 Safari. Both of the major operating systems offer an

13 additional entitlement called default browser that you could

14 explicitly ask the user permission for. That effectively made

15 the Neeva app the browser of choice on the phone. But to the

16 extent that Neeva's success depended on having a lot of

17 users -- and then because the only way to have a reliable large

18 number of users consistently use your product is to be the

19 default, I would say that continued to be one of the biggest

20 priorities and worries for the team, for the investors and for

21 me throughout Neeva's existence.

22 Q. Are you familiar with the term click and query

23 information?

24 A. Could you repeat your question?

25 Q. Sure. Click and query information?

 3695

1 A. Yes, it is a -- yes, I am. It's a rough area that

2 talks about what sites users clicked on after querying for

3 something. It's a pretty large field.

4 Q. And are you familiar with the term scale?

5 A. Yes.

6 Q. And what do you understand scale to mean, sir?

7 A. In this context you mean?

8 Q. Yes.

9 A. You know, search engines are in the business of

10 showing you the best sites, the best answers than your type-in

11 queries. Obviously, the list of queries that humans can think

12 up is potentially unbounded. And, in fact, this is true for

13 Neeva, this is true for Google as well as Bing, something like

14 15 to 20 percent of queries that are typed into a search engine

15 tend to be unique period, they have never been seen before.

16 And the job of a search engine is to provide you with the best

17 sites for any query, including ones the search engine has never

18 seen before. So it's a formidable challenge.

19 And one of the biggest signals that all search engines

20 have relied on for the past 20 years is this thing that you

21 talk about, which is query click information. Search engines

22 are -- sort of capture the paradoxical concept that the best

23 results for a particular query are ones that people have

24 clicked on before and have liked. And so scale here refers to

25 how much query click information is one able to collect.

 3696

1 Because not only does it make your search engine product

2 better, it also gives you very powerful insights into figuring

3 out how to answer queries you have never seen. You can do

4 things like figure out the queries that are semantically

5 closest to the new query that you've never seen, and use that

6 context to figure out what results to show. And so the more

7 users you have, the more queries you have seen over time, the

8 more click behavior you've observed over time, the more

9 effective you can be in creating the -- creating a higher

10 quality search engine.

11 Q. Now, I think you mentioned that Neeva started using

12 artificial intelligence to improve its results; is that right?

13 A. That's correct, in many different ways. I'm happy to

14 go into them.

15 Q. Does AI eliminate the importance of behavioral data in

16 trying to figure out what the user is looking for?

17 A. I don't think it eliminates it, it makes certain

18 problems easier. To sort of coarsely break search down, when

19 you type in a query into a search engine, you first go through

20 a process by which semantically-related queries or misspellings

21 that there might be in query-related concepts and so on are

22 added to the context of a query. And this -- with this

23 additional context, search engines do a first round of

24 retrieval of potential search results. And then they use query

25 click information to help figure out the best among these

 3697

1 results.

2 And what AI offers in addition is you're then able to take

3 the contents of the best pages that come back after the

4 retrieval to craft, for example, an AI-powered answer. In

5 other words, AI enables search engines to do things that are

6 not really conceivable in a return-a-set-of-links model, which

7 is what commercial search engines generally do today. AI is a

8 significant help in the first problem, because large language

9 models, as they are known, are trained on -- you know, most

10 pages that exist, their ability to handle spell corrections or

11 figure out semantically-related concepts, you know, for a

12 query, that is -- you know, you have a lot more ability in a

13 fairly small model that previously used to be powered by a lot

14 of query and click history. So AI is very helpful there.

15 AI let's you do, as I said, things like summarization,

16 presenting a single answer in ways that, honestly, search

17 engines of old could not do. But the middle problem of

18 figuring out what are the most relevant pages for a given query

19 in a given context still benefits enormously from query click

20 information. And it's absolutely not the case that AI models

21 eliminate that need or supplant that need.

22 Q. If we could please go to -- we're going to turn now to

23 your time at Google, and I just want to show you a few

24 documents. If you could go to UPX475 in your binder.

25 THE COURT: Mr. Dintzer, could I interrupt you for a

 3698

1 moment?

2 MR. DINTZER: Of course.

3 THE COURT: Mr. Ramaswamy, fairly simple question: Why

4 did you think you could compete with Google?

5 THE WITNESS: Because of a hunch that starting over at a

6 different point with a different business model with a set of

7 new insights about how products could work often create

8 opportunities. Now, I'm not claiming I'm smart enough to have

9 figured out that the AI wave would come. But part of our bet

10 was that if you had a clean slate, that we could visualize

11 something much, much better. It was also driven by our

12 intuitions around various other large scale technical changes

13 that had taken place. If you look back, for example, to the

14 history of Android, which came via a 40 percent acquisition

15 that Google did, no one would have said in -- call it circa

16 2007-8, that Google could have created an operating system that

17 would rival Nokia and Microsoft. It was an unthinkable concept

18 at the time. But yet Android exists and is easily the largest

19 operating system on the planet. As a technologist, you often

20 look for seams and discontinuities in experiences. And there

21 is also a lot of -- you know, you have a certain direction and

22 you can't figure out all the details, there's very much an

23 element in startups of figuring out things as you go along.

24 And the final point was also that unlike an ads model

25 where economies of scale and getting bigger tend to be very

 3699

1 pronounced -- because it is more of a two-sided marketplace

2 where you need users, you also need advertisers and the

3 platform. A consumer subscription product is a much simpler

4 proposition. You know, a lot of Neeva was predicated on the

5 belief that we could get modest escape velocity, which is

6 millions of paying subscribers which can most definitely

7 support a team that can create a good product. Again, this is

8 not dissimilar to, say, how Netflix happened. No one in their

9 right mind in 2000 would have said oh, let's move away from

10 ad-supported TV that existed over to a streaming model. As I

11 said, as technologists, you look for seams. Sometimes you get

12 them, sometimes you don't.

13 THE COURT: So I may be preempting further questioning,

14 but these were the seams you thought you could exploit or take

15 advantage of or build on, but ultimately were not successful --

16 THE WITNESS: That is correct.

17 THE COURT: -- in doing so? And so can you help me, what

18 is your explanation of Neeva's inability to ultimately be able

19 to compete in the search market?

20 THE WITNESS: The problems that we would face were

21 predictable four years ago, meaning that acquiring users,

22 getting them into a habit is something that is tricky. Every

23 startup founder has sort of dealt -- has dealt with this. And

24 there is always a race to create a better product fast enough

25 that you get enough users before your employees tire or your
 3700

1 money runs out. These are all very, very real things. Also,

2 the underlying market which -- you know, around things like how

3 many people can you convince to install your extension that

4 changes their default search, or even things like the cost to

5 acquire a user, say, commercially by paying for ads, whether

6 it's for desktop extensions or for mobile apps, is a dynamic

7 changing thing. What was true last year, in terms of how many

8 dollars it would cost to acquire a user, might not be true this

9 year.

10 And so through Neeva's four years, the core thesis around

11 we can create a much better experience, we can create a search

12 engine from scratch. We ran a pretty by scale system. We did

13 our own crawl, we did our own search index. These are things

14 that most even -- like I would say most good engineers would

15 basically give up on before they start, because it is a

16 Herculean problem. Unlike, say, creating an e-mail client on a

17 phone, that it's a relatively predictable surface area. You

18 need a place to show the inbox, to be able to compose mails,

19 manage an address book. Search, you have to sort of understand

20 the whole web, figure out how to make a copy of it, run a

21 system that figures out the best pages. Our ultimate take was,

22 yes, we could build it. And especially this year, in practice,

23 people that tried things like our AI experience genuinely loved

24 it, because it was just a better, easier, sleeker experience.

25 But, as I said earlier, part of what was weighing on me

 3701

1 was our ability to grow our user base fast enough, and on that

2 user base grow the subscriber base fast enough. And, as I

3 said, especially in a world of very changed expectations around

4 valuation and revenue, we sort of -- and our inability to be

5 even a default provider on things like the major browsers or

6 operating systems was what made us conclude that it was hard to

7 have Neeva consumer search as a viable business. The

8 technology pieces that we had built, whether it was crawl or

9 whether it was the software serving system, or even our

10 expertise around using AI on top of data, those are all going

11 to be very useful. And that was part of the reason why we

12 decided to pursue an acquisition.

13 THE COURT: Thank you.

14 BY MR. DINTZER:

15 Q. Thank you, Your Honor. Sir, if you could turn to

16 UPX475 in the binder. And, again, it will appear on the screen

17 if you'd rather just do it that way. Oh, actually it won't,

18 because Google has asked that this not be shown.

19 Do you have is, sir?

20 A. Yes.

21 Q. And you can see that this is an e-mail; is that right?

22 A. Yes.

23 Q. And it's dated July 2018?

24 A. Yes.

25 Q. And you're one of the recipients; is that right?

 3702

1 A. That's correct.

2 Q. And the subject is Q2 ads financials recap?

3 A. Yes.

4 Q. If you could just read the first sentence to yourself,

5 not out loud, and then just let me know when you're done

6 reading it.

7 (Witness reads document)

8 A. Yes.

9 Q. So what is this -- who is Dominic Tang?

10 A. To give folks a little bit of context, it's okay if I

11 mention the recipients of the e-mail, correct, or no?

12 Q. I'm sorry, just a little louder.

13 A. Can I mention who this e-mail was sent to in addition

14 to me?

15 Q. Yes, that you can.

16 A. So Phillipp Schindler is -- was, is the global head of

17 business. And he and I, as I said, ran the ads and commerce

18 product and engineering teams, so we were peers when it came to

19 running the ads business. There was an independent team that

20 reported up to the CFO that would send us summaries about the

21 state of consumer behavior as well as the state of the ads

22 business. And so the sender of this e-mail is somebody that

23 worked on the team updating us about the state of Q2. Google

24 runs on a calendar quarter -- so this e-mail's roughly like mid

25 July, and so a couple of weeks after the quarter had closed.

 3703

1 Q. And it says: "Attachments - Global Ads Financials

2 Fact Pack." What's a fact pack?

3 A. Typically, Google's ad business, like any other

4 business, has a number of dimensions to it. These dimensions

5 include the various products that we had, things like search,

6 YouTube advertising, revenue from publishers, revenue from ad

7 exchange and so on. It also had other geographical dimensions,

8 money made from United States, Americas, Europe and so on. And

9 so a fact pack is both a summary and a set of detailed

10 breakdowns along many of these axes, and is meant to convey the

11 state of the business.

12 Q. And what unit within Google prepared these numbers?

13 A. So this was a team that reported up to the CFO.

14 Q. So now --

15 A. And several people on the CC list of this e-mail are

16 all from that team.

17 Q. Now, I'm going to ask you to turn to -- it's page --

18 it's numbered page 14 in the attachment, it's Bates numbered

19 X744. And a redacted version appears on the screen. Just let

20 me know when you're there.

21 A. Uh-huh.

22 Q. Do you see, sir, that the top says user metrics?

23 A. Yes.

24 Q. And then in the first -- in the fourth column, the

25 heading is Desktop Search Query Share. Do you see that?

 3704

1 A. Yes.

2 Q. And to the left it has -- in the top row it has the

3 United States; is that right?

4 A. That is correct.

5 Q. And then there's a number that we won't read out loud,

6 it's a percentage, that is -- do you understand that to be

7 Google's U.S. desktop query share in Q2 2018?

8 A. Yes, that's correct.

9 Q. And then if you move -- if you slide over two columns,

10 it says Mobile Search Query Share. Do you see that?

11 A. Yes, sir.

12 Q. And, again, there's a percentage that we won't read

13 out loud, but did you understand that to be Google's U.S.

14 mobile search query share in Q2 2018?

15 A. Yes, sir.

16 Q. And what explains the difference between these two

17 figures? Obviously they're different numbers. What explains

18 the difference between the two of them?

19 A. Can you give a little more color to your question?

20 Q. Sure, of course. Why are they broken out desktop

21 versus mobile?

22 A. You know, these are obviously different platforms.

23 Users interact with their desktop computers in ways that are

24 very different from how they interact with their phones. You

25 know, a lot of running a business is understanding the

 3705

1 components that make up the whole. And so what you see here,

2 in the context of search, is a breakdown by geography as well

3 as the different platforms that there are. Obviously the

4 history of these platforms is also very different.

5 I'll keep it short, but search came of age at a time --

6 Google search came of age at a time when desktop was already an

7 established platform. This is early 2000s. But, you know,

8 computers were around, lots of people used them. There were

9 existing search engines like a Yahoo!, like Ask.com, so on and

10 so forth, while the mobile world converged rather quickly on

11 essentially being a bipolar world between iOS and Android. And

12 so they have very different histories. And Google's role in

13 search also is vastly different in the rough decade that -- you

14 know, when desktop came of age versus when mobile came of age.

15 And so these are some of the factors that can cause these

16 numbers to be very different.

17 Q. Your Honor -- no further questions on this document,

18 sir. This document is already in evidence, UPX475. We have a

19 set of these that we would like to move into evidence, some of

20 which we are -- we have faced objections over. And so we'd

21 like to resolve those while we have the witness on the stand

22 and move the set of them -- obviously we believe we've laid the

23 foundation, move the set of them into evidence. And I can --

24 they're in the binder -- I believe they're in the binder.

25 THE COURT: I'm sorry, let me cut to the chase here. So

 3706

1 there are multiple of these stat packs you'd like to move into

2 evidence?

3 MR. DINTZER: Yes, Your Honor. Basically we have a full

4 set of them for the Court.

5 THE COURT: And by that, you mean over some temporal

6 period to establish the shares that you just identified?

7 MR. DINTZER: That's exactly right, Your Honor.

8 THE COURT: And there's an objection to that?

9 MR. SMURZYNSKI: Your Honor, if we're talking about a

10 collection of documents that are identical in nature to these,

11 then we have no objection. But I don't have a list -- I

12 haven't been presented with a list or a package. So on the

13 representation that they are the same thing, just for different

14 time periods, we have no objection.

15 MR. DINTZER: Okay. And that's fair, Your Honor, we'll

16 get them a list and a set, and get them into evidence.

17 THE COURT: You'll get me a stat pack binder.

18 MR. DINTZER: I'm sorry, what, Your Honor?

19 THE COURT: Get me a stat pack binder.

20 BY MR. DINTZER:

21 Q. Let's go to UPX93 and UPX123. So we're going to look

22 at two documents, because the second was the attachment to the

23 first -- and these are in evidence, Your Honor. Let me know

24 when you have --

25 A. Which is the first one, sir?

 3707

1 Q. 123 is the cover letter.

2 A. Yep.

3 Q. I'm sorry, 93 is the cover letter, and we can start

4 there.

5 A. Yep.

6 Q. And you see an e-mail chain that includes you, sir?

7 A. Yes, sir.

8 Q. And who is Diane Tang?

9 A. Diane is -- well, so we are talking about 2007, so

10 everything I say will be in reference to that. So Diane was

11 one of the key engineers in what's called the -- what was

12 called the ads quality team, which was the team that was

13 primarily responsible for figuring out what ads got shown and

14 where on Google.com.

15 Q. And if you can remind me, what was your role in 2007,

16 sir?

17 A. So I ran the ads quality team. Diane reported to me.

18 I was also running the advertiser-facing and the infrastructure

19 teams in ads at the time.

20 Q. And on the bottom e-mail -- you know, read upward, so

21 the bottom most one, on March 22nd, 2007, Ms. Tang writes --

22 actually, let's do the subject first. The subject of this

23 e-mail chain is Nitin on home pages. Do you see that?

24 A. Yes.

25 Q. Do you know who Nitin was?

 3708

1 A. Nitin Sharma at that time was part of a group we

2 called the metrics team. These are people that we would

3 variously describe as data analysts or data scientists today.

4 These were the people that dived into different topics

5 typically around search ads, and did exploratory projects and

6 presented us with insight and ideas.

7 Q. So on the bottom e-mail, Ms. Tang writes: "You guys

8 have seen most of this says Nitin, so should be about half an

9 hour, and then rest of time for discussion." Do you see that?

10 A. Yes, sir.

11 Q. Do you remember -- and you can look at the deck that

12 we're going to be looking at in just a second which is the next

13 document UPX123. But do you remember this general discussion

14 back in 2007?

15 A. Yes, I do.

16 Q. And the next thing, she shares the title on the

17 strategic value of default home page to Google -- which is the

18 title I believe of the deck at UPX123, and then the abstract

19 which says:

20 "One of the most fundamental questions about web

21 search, on which rests a $15 billion search business,

22 is what influences the choice of a search engine.

23 Several factors are believed to affect the choice,

24 including the quality of search results, brand

25 strength, search features, the quality of user

 3709

1 experience, the presence of local competitors, etc.

2 In this talk, we present some evidence that seems to

3 suggest that one factor surprisingly trumps them all:

4 the default home page setting. Using data from

5 Google logs, we show that users who have home page

6 set to Google do 50 percent more searches on Google

7 compared to those that don't. Google's market share

8 by region seems directly correlated with the share of

9 default home page in those regions."

10 Do you see that?

11 A. I do.

12 Q. Do you remember the presentation and the discussion of

13 the importance of defaults to Google's market share?

14 A. I do.

15 Q. So let's go to the deck UPX123. The Court's already

16 seen much of it, so I'm not going to take a lot of time. We're

17 going to go to page 21 on the deck, and that's Bates number

18 487. Just let me know when you're there.

19 A. Yeah.

20 Q. At the top it says key takeaways. Do you see that,

21 sir?

22 A. Please go ahead.

23 Q. And the second bullet says: "HP sets could be a

24 powerful strategic weapon for Google." Do you see that?

25 A. I do.
 3710

1 Q. And then it says: "Could be an easy way to grow and

2 defend search market share." Do you see that?

3 A. I do.

4 Q. And did you understand that to be search market share

5 in the general purpose search engine?

6 A. That's right.

7 Q. And then let's go back now to Exhibit UPX93, so we're

8 going back to the e-mail chain.

9 THE COURT: I'm sorry, maybe I'm missing something. HP

10 sets -- oh, is it home page?

11 THE WITNESS: Home page sets.

12 THE COURT: Okay. I just wanted to make sure I understood

13 that.

14 BY MR. DINTZER:

15 Q. Those are home page defaults; is that right, sir?

16 A. That's correct. Some context is helpful. If I can

17 just do that for two minutes, and if you have a follow up

18 questions, we can go to that.

19 Q. Of course, please.

20 A. So, first of all, we are talking 16 years ago. What

21 is important to understand is that most of us -- you know, the

22 ones that were around, would get to a search engine or would do

23 anything on the computer literally by opening a browser and

24 then doing something with it. This was not the time of

25 trackpads, this was not the time of computers being on

 3711

1 perpetually with a browser always running. The default is

2 somebody opens up a computer, like even log into it -- because

3 computers were quite often shared. They would bring up Chrome,

4 they would bring up Internet Explorer, and then they would do a

5 specific task with it. They would shut the browser and then go

6 on. So that's the mental framework you need to keep in mind.

7 What this effectively says is that there was enormous

8 value in getting Google.com to be the home page that comes up

9 whenever somebody opened the browser. In retrospect, this is a

10 perfectly obvious sort of conclusion where if you make it easy

11 for somebody to get to something, they're going to do more of

12 it. Except that 16 years ago, it was not that -- you know, it

13 was not that obvious. There were people that would be like oh,

14 people will elect to use Google.com by typing in Google.com

15 into the URL bar. This was not a time when like the URL bar

16 was used as the place for typing in a search query, as it is

17 today. You would have to actually type in some site, go there

18 and do something with it.

19 So what you see here is really the beginnings of Google

20 taking these defaults very, very seriously, starting with this;

21 and doing a lot of commercial deals to have Google.com be the

22 default home page on a lot of PCs that got sold, for example.

23 So Google soon struck deals with like every PC manufacturer

24 that there was to make Google the home page, all driven by a

25 simple but powerful observation that as humans, like we do the

 3712

1 most efficient thing possible. And the way you get people to

2 use a lot of search is by being the home page as much as

3 possible, so we at Google did this.

4 We also went one step further and said why do you need a

5 home page at all, you can simply type into the URL bar. And

6 that is now the de facto standard for the world. As I said,

7 whoever controls that search box gets a lot of usage

8 independent of the merits of the search engine. And so you get

9 enormous usage simply by the power of the default. You know,

10 when we talk about sort of why Neeva failed, it was simply

11 because we could not get enough users to be in that state where

12 they regularly used Neeva.

13 So hopefully this gives you -- it's an evolution. It

14 started with things like setting the home page at a time when

15 going to that home page was the dominant way in which many

16 people got going with the internet, tracing the arc all the way

17 to whoever controls the search box, the URL -- I mean, it's

18 called the omnibar where people enter queries or people enter

19 sites, they're going to be the winners of the search game

20 overall.

21 Q. And --

22 THE COURT: Sorry, Mr Dintzer. You said that Google would

23 strike deals to make Google the default home page. Who was

24 Google striking deals with, a browser company --

25 THE WITNESS: All of the above.

 3713

1 THE COURT: -- that was not Google?

2 THE WITNESS: It's Dell, it's Compaq, it's IBM, it's

3 Firefox, it's everybody.

4 THE COURT: What was the deal with the OEMs? In other

5 words, the OEMs, they would have a browser preset, but you had

6 the ability -- would the OEM have the ability to default to a

7 home page?

8 THE WITNESS: That's right, and so Microsoft would be

9 pretty adamant about things like the role that Internet

10 Explorer would play on Dell PCs. But at the time, they were

11 not as adamant about who got the home page. And so these

12 were -- you know, Microsoft has large and complicated

13 agreements with all PC manufacturers about what they can and

14 cannot do. But things like setting the home page at the time

15 were not really covered by this. And so there were basically a

16 variety of commercial agreements, many of which are well known;

17 everything from a bounty program, meaning that the manufacturer

18 got paid a certain number of dollars if a user had Google as

19 the home page and stayed on for a certain period of time, to

20 more complicated ones like an ongoing revenue share which was

21 more the agreement with people like Firefox. Those details are

22 sort of -- they're market-driven and not all that relevant.

23 But the point is everybody that made browsers or machines,

24 Google went and struck business deals to be the home page or

25 the default search on those browsers.

 3714

1 THE COURT: Thank you.

2 BY MR. DINTZER:

3 Q. And you're familiar with the term search access point?

4 A. It's what we just talked about, it's the browsers,

5 it's the home page. The search access points have changed over

6 time, but it's roughly the places where people get to a search

7 engine using whatever device they have.

8 Q. And as search access points, as you said, have changed

9 over time and have multiplied, the basic concepts of the

10 importance of defaults has remained?

11 A. That's correct.

12 Q. And so if we can just go back to UPX93. At the top,

13 you send that e-mail; is that right?

14 A. Yes, sir.

15 Q. And if you could read that.

16 A. Sure. "This study is very cool. We should definitely

17 put some marketing push behind it, will make some inquiries."

18 Q. And with that, Your Honor -- and Google did put some

19 marketing push behind it; is that right?

20 A. That's right.

21 MR. DINTZER: With that, Your Honor, we have no further

22 questions, and we pass the witness.

23 THE COURT: Thank you. Okay, Mr. Cavanaugh.

24 DIRECT EXAMINATION OF SRIDHAR RAMASWAMY

25 BY MR. CAVANAUGH:
 3715

1 Q. Good afternoon, sir. My name is Bill Cavanaugh, and I

2 represent the states of Colorado and Nebraska. I just have a

3 few questions for you today. If you could start by looking in

4 your binder at PX1022.

5 A. Yes, sir.

6 Q. Which is already in evidence, Your Honor. Are you

7 familiar with this document?

8 A. Yes, I think I wrote it.

9 Q. And what did you write it in connection with? Was

10 this in connection with the introduction of Neeva?

11 A. I think this is the time at which we made Neeva

12 available to anyone in the United States.

13 Q. Let me draw your attention on the first page at the

14 bottom. And this has your name on it, so did you draft this or

15 work on it?

16 A. I did.

17 Q. You state at the bottom: "I have come to believe that

18 ads detract from a good search experience and have also had

19 many unintended side effects that have large social

20 consequences."

21 Do you see that?

22 A. I do.

23 Q. You then go on to describe what your concerns are

24 regarding ad-based searches?

25 A. Yep.
 3716

1 Q. The first one you talk about, you say: "First, the

2 very existence of an online ad right on the top of the search

3 results pushes down and deprioritizes the information we are

4 searching."

5 What did you mean by deprioritizing the information we are

6 searching for?

7 A. Most of us scan like pages top to bottom, whether it's

8 on a desktop or on a mobile phone. When we type something into

9 the top of an app, the app renders results top to bottom.

10 That's the scan order. Having the ability to show stuff that

11 people have to look at is what I mean when I talk about the

12 deprioritization. Don't get me wrong, there are some ads that

13 can be relevant. I helped build a very large business on that

14 basis. But it also can take a life of its own, a backdoor

15 point about how commercial is something that someone is willing

16 to pay ads money for. There can be situations in which like

17 unpleasant, dangerous ads can be up there. But the general

18 point is that an over reliance on the ads model can take away

19 from what I consider to be the primary purpose -- and what

20 everybody at Google and elsewhere would agree with is a primary

21 purpose of a search engine, which is to give you relevant

22 results as quickly as possible.

23 Q. Thank you. You then go on to say: "Second,

24 ad-supported search engines face the daily pressures of

25 returning value to their shareholders by prioritizing

 3717

1 advertisers and ads revenue."

2 Can you explain to the Court what your concern was there?

3 A. I mean, this is just saying that all companies have

4 expectations of growth, and the relationship between quality

5 and revenue is not cut and dry, it's a trade-off. And over

6 time, especially, you can end up making a series of decisions,

7 all of which appear to be just fine in isolation. But driven

8 by things like the need to make more revenue, you can end up

9 creating overall experiences that are not so positive for

10 users.

11 Q. You then go on in the next paragraph to say: "These

12 problems" -- which you've just referred to, "are further

13 exacerbated by the fact that there tends to be only one or two

14 viable search engines in most countries."

15 How is that exacerbated?

16 A. It just means that there's effectively no

17 counterweight.

18 THE COURT: I'm sorry, no what?

19 THE WITNESS: No counterweight, because there are only two

20 providers. You earlier asked me why I thought I could compete

21 or why I started Neeva at all. And one of the reasons was

22 because there needed to be a product that just thought about

23 search very differently. And if it were to take off, then

24 there would be a counterweight to how much people could push

25 monetization. It is this lack of competition, this lack of

 3718

1 alternative that makes for a situation in which experiences get

2 worse and worse over time. Plus, you also end up creating

3 environments where there effectively is going to be no

4 competition.

5 BY MR. CAVANAUGH:

6 Q. Were the views you expressed in this article driven at

7 all by your experiences at Google?

8 A. Sorry, can you repeat that?

9 Q. Sure. The views you expressed in this article, were

10 they shaped by your experiences at Google?

11 A. I mean, yes, but we are continuously informed by the

12 experiences that we go through. You know, I will sort of

13 openly admit, confess, state that I was the final approver in

14 many launches that increased ad load. We tried to do the best

15 job that we can. We looked at all kinds of trade-offs. Diane

16 and I probably made many of the pivotal decisions for search

17 ads for the better part of a decade. But you can also realize,

18 especially over time, that there need to be alternatives, that

19 there need to be different ways of looking at the same problem.

20 These are the principles that Google famously thought of for

21 itself. And so, yes, everything that -- like many things that

22 I did at Neeva, including its conception, were informed by my

23 personal experiences of search, of search ads, of where it can

24 go and where things needed to be.

25 Q. In creating Neeva, did you make efforts to

 3719

1 differentiate it -- beyond simply being ad-free, did you make

2 efforts to differentiate it from other general search engines?

3 A. Yes. We talked about privacy and just the feeling of

4 futility. We also talked about how we created the experience

5 to be far more personalized. And I would really say that the

6 culmination of our efforts to create a better search experience

7 was what we did on a shoestring budget with a team of 50 people

8 to be able to launch a generative AI experience that gave

9 concise answers to complicated questions that it had references

10 for, every sentence that it wrote was realtime. These are

11 things that companies with billion-dollar budgets like still

12 have not rolled out to the mainstream even to this day. Both

13 Sydney and Bard, the generative AI experiences from Bing and

14 Google respectively, are trials that you have to enroll into.

15 And so, yes, I think we -- you know, in many ways we showed

16 what was possible in the search experience.

17 Q. Let me ask you to turn to the second document in the

18 binder I handed you, and that's PSX1023 which is already in

19 evidence, Your Honor. This is another blog post from Neeva,

20 correct?

21 A. Uh-huh.

22 Q. The title is FastTap Search?

23 A. Yep.

24 Q. I don't think you've used that term today. What is --

25 what was FastTap?

 3720

1 A. I tried to avoid using too many -- too much jargon,

2 but I talked about this earlier. I said how, when we created

3 an iOS mobile app, we had more control over the search

4 experience including what happened in the autocomplete box.

5 FastTap was just the name that we gave for it. Essentially

6 today, for most queries, especially commercial queries, as you

7 start typing into a browser, you're going to get essentially

8 like text completions. Because there is an incentive for the

9 search engine to take you to a page that has ads where it can

10 monetize. Instead, with FastTap, we essentially came up with a

11 scheme where, for over half the queries, we could take you to

12 the site that you wanted without needing to go to a search

13 results page. Because our monetization was tied just to

14 creating a great search experience, not necessarily on driving

15 traffic to our search pages. FastTap was just a name that I

16 think my son came up with for this feature.

17 Q. I'm sure you gave him credit. How did Neeva select

18 the results that show up through FastTap?

19 A. It's the same underlying algorithm. So, in other

20 words, if you hit return, you would get these varied results.

21 But basically we are able to show it to you faster so that you

22 didn't have to go through the trouble of getting a page and

23 then parsing it and then tapping it. And certainly with Neeva,

24 you didn't have to then like parse some ads. But the idea

25 basically is that you use the same search algorithm. You run
 3721

1 it quickly, because the latency expectations of an autocomplete

2 are an order of magnitude smaller, it has to be faster than a

3 regular search result. But we would basically run the same

4 algorithm and show you results faster.

5 Q. On PX1023, on that first page, there's a heading

6 Innovation that Could Not Exist with an Ad-Driven Business

7 Model. Can you explain what you meant in connection with

8 FastTap?

9 A. It's just what I said, which is that anything that

10 reduces the number of search result pages, especially ones with

11 ads, is something that a commercial search engine cannot afford

12 to do without. So it's not the case that Google or Bing

13 engineers are not clever enough to come up with a feature like

14 this, they are, but it's just unlaunchable; while we, with a

15 focus on consumers and customers, could launch this feature.

16 Q. Thank you, sir. I'm done with 1023. What is Neeva

17 Scope?

18 A. Neeva Scope was another feature that we launched in

19 our mobile app, and his the idea was that you would get

20 additional context into any page that you were looking at. The

21 idea was to surface things like conversations centered around a

22 product. Let's say you're looking at a product page in the

23 Neeva browser. Neeva Scope would, for example, show you the

24 Reddit pages for people who are having conversations about the

25 product. What has happened with product pages on the internet

 3722

1 is that there is this unending drive to monetize them. So most

2 sites are only going to show links to other products on their

3 own site with very little interest in surfacing information

4 that might be useful to you as a user.

5 Using the same technology of search that we had built, we

6 could show you things that were of interest to you with respect

7 to a page that you were browsing. You can think of it as

8 basically the other side of the search coin whereas opposed to

9 telling you where to go, we look at a page and give you

10 additional context for what else you might need to know about

11 that page.

12 Q. Are you also familiar with the term tracking

13 preventer?

14 A. Yes, sir.

15 Q. Can you explain to the Court what tracking preventer

16 was on Neeva?

17 A. One of the distressing things about the internet today

18 is that any site that you go to has software from several

19 hundred companies that are effectively constructing profiles of

20 everything that you do. This powers online advertising, and

21 this is the one that leads to things like -- you know, you

22 looked at, take your pick, a shoe or a bag once, and now the

23 entire world seems to show you ads for those products

24 everywhere that you go. And this is done with small pieces of

25 code that run on these pages. They are often -- very, very
 3723

1 often not particularly useful or necessary when it came to how

2 that page operates.

3 And so we launched this technology called tracking

4 prevention that would essentially block out companies that were

5 in the business of constructing profiles. We shipped that as

6 part of our desktop extension, but we also shipped it into our

7 mobile apps. And the idea -- you know, going back to our

8 earlier discussion about privacy, is what you do on a site

9 should be between you and a site, and should not be observable

10 by thousands of companies that want to build profiles on you.

11 This is also the kind of thing that is very difficult for

12 commercially-supported search engines, commercially-supported

13 browsers to do, even though I think the tide is turning and

14 there are more and more browsers that are making things like

15 tracking prevention more of a default in how they operate.

16 Q. Let me turn to a different topic. Are you familiar

17 with the phrase specialized search engine or specialized

18 vertical provider?

19 THE COURT: Mr. Cavanaugh, can I interrupt you before you

20 go to another topic?

21 MR. CAVANAUGH: Sure.

22 THE COURT: How would you respond to the following: Which

23 is that some of the concerns that you've had with commercial

24 search that you thought made a good business case to build a

25 different type of search engine, privacy being the primary --

 3724

1 one of the primary concerns, what would your reaction be to

2 somebody who says Neeva's inability to stay in business is just

3 a feature of the market; people don't really care about privacy

4 as much as you might think, and so if they really cared, a

5 company like Neeva would have been more successful?

6 THE WITNESS: You know, I'll readily admit that that is a

7 conclusion that you can draw from Neeva's story. But there are

8 potentially other conclusions -- and I've written about it, so

9 I'm not saying anything new here, which is that the lack of

10 access to consumers serves as a self-fulfilling prophecy for

11 why there can never be change. In my mind, that is equally

12 possible. I'm not an oracle, I'm not claiming one is true over

13 the other, they're both true. I think the -- you know, people

14 have busy lives. Yes, they care about privacy, but if you're

15 forced to take a lot of action to defend your privacy, most

16 people get on with their lives and forget about the privacy.

17 That is the reality of how all of us live.

18 On the other hand, if choices were easy, if a well-funded

19 and exceptionally talented team like Neeva could not even be a

20 provider on most of the browsers, I don't see that as the

21 market working. I see that as a bunch of commercial interests

22 that are -- you know, as they should, I guess, defending how

23 they operate.

24 THE COURT: Thank you.

25 BY MR. CAVANAUGH:
 3725

1 Q. Based on your experience, do you believe users care

2 about privacy?

3 A. Everybody will assure you that they care about

4 privacy, there's just no question about it, mostly because they

5 don't even know what privacy is. It just sounds like a

6 reasonable thing to say. It's like dieting, everybody likes

7 dieting. The question is are there easy ways to -- are there

8 easy ways to get to it, and what are like simple concrete

9 steps. And people who talk to me about privacy and security,

10 you know, I would tell them like use a password manager, use a

11 safe browser, use a tracking-free search engine. There are

12 simple steps that you can take. And so I think the question of

13 whether people care about privacy needs to be answered in the

14 context of even if you were to, what could you do about it.

15 Q. Let me go back to my question before. Are you

16 familiar with the phrase specialized search engine or

17 specialized vertical provider?

18 A. Yes, sir.

19 Q. Did Neeva utilize data from what I'll refer to as SVPs

20 to assist in providing query results?

21 A. A search engine is by now, in 2023, a pretty complex

22 piece of software. Not even the almighty Google can source all

23 of the information needed to serve a search engine by itself.

24 And so, yes, we had relationships with a number of companies

25 for local business information, for things like stock

 3726

1 information, for weather. There were a variety of things that

2 were and are needed to operate a commercial search engine. We

3 took great pains to ensure that the promises that we made to

4 our users with respect to how they use Neeva and what

5 guarantees we offered them were applicable, even when we showed

6 results from some of these specialized providers, for example,

7 with local.

8 As I talked about earlier, whenever we would make requests

9 on behalf of our users, we would make sure to not send IP

10 information, because IP can pretty easily be traced, at least

11 to a household. So we took great care to ensure the privacy of

12 our users. But, yes, we had relationships with specialized

13 people. Sometimes it was for data that they would give to us

14 once a month or once every two weeks, other times it was for

15 making life calls during the life cycle of a search request; in

16 other words, within 800 milliseconds or so.

17 Q. Did you have an arrangement with Yelp?

18 A. We did have an arrangement with Yelp, yes.

19 Q. And how did that arrangement work?

20 A. It's a commercial arrangement. We paid them a certain

21 number of dollars every month. It was a combination of

22 datasets and online calls. They had requirements for how their

23 data was presented, what it could be commingled -- you know,

24 usual commercial stuff.

25 Q. Would Yelp get credit on the results page?

 3727

1 A. Yeah, yeah. At the bottom of every page, we would

2 credit the people that we got information from. On local

3 queries, Yelp would be mentioned as one of the people that we

4 got information from for that query.

5 Q. Was that mutually beneficial for both Yelp and Neeva?

6 A. You know, not really. I mean, the arrangement was

7 mutually beneficial: They made money, we got local result.

8 But if you're talking about the attribution at the bottom, it's

9 one of these pro forma things that doesn't actually do anything

10 for anyone.

11 Q. Are you familiar with Google's Search Ads 360?

12 A. Very different topic, yes.

13 Q. Yes, sorry, I should have told you. We refer to it as

14 SA360. Are you comfortable with that reference?

15 A. Yes, I am. I think I ran the team for close to a

16 decade, so yes.

17 Q. Well, then you certainly know more than me. Is SA360

18 what evolved out of what was the original DoubleClick

19 acquisition?

20 A. Yes, sir.

21 Q. When Google acquired DoubleClick search, did

22 DoubleClick have a search ad platform?

23 A. Yes, it was called DART for Search. It was an

24 acquisition that the original DoubleClick company had done

25 before Google acquired DoubleClick.

 3728

1 Q. Did it have -- let me ask it differently. When Google

2 acquired DoubleClick, Google already had a search ad platform,

3 correct?

4 A. So DART for Search is what is referred to as an SEM,

5 search engine marketing, platform. It is meant to be a

6 one-stop shop by which an advertiser could buy advertising

7 across Google search, Bing search, and then over time things

8 like Yandex. It was meant to be a one-stop shop. You know,

9 Google and Microsoft could never agree to have their main

10 search engines be represented directly in like the premier

11 buying experience for those platforms. In other words, you

12 could not buy ads on Bing search from Google AdWords, because

13 Microsoft would never agree to it and vice versa for good

14 reason. So there were a number of companies: DART for Search

15 which then became DoubleClick Search which then became Search

16 Ads 360 is one of them.

17 Q. When you had Google Ads, and Google then had SA360 --

18 A. Can you be a little bit more precise. Do you mean

19 Google AdWords?

20 Q. Yes.

21 A. Okay.

22 Q. And Google then had SA360?

23 A. Yes.

24 Q. Were there any concerns about a conflict of interest?

25 A. That is a great question. Because Google was both a

 3729

1 big platform for search, but then now was creating a product

2 that was meant to represent advertiser interests across a set

3 of search platforms, there definitely was concern on the part

4 of lots of people -- advertisers for a start, about fairness

5 and who had access to what. And we had good internal rules for

6 who could get access to DoubleClick search. And certainly the

7 tone that I set for that team -- which was initially based in

8 Chicago and then in Seattle and Chicago, was that they needed

9 to represent advertiser interests first and foremost. And the

10 fact that a good bit of the spend went through Google needed to

11 be justified in terms of benefit to advertisers and not in

12 terms of them -- of the team being a part of Google. I'd like

13 to think that we did a good job with that. The product

14 definitely grew quite a lot.

15 But there was a very clear distinction when it comes to

16 like essentially obligations and responsibilities when it came

17 to the product. And this also extended out to the sales team.

18 Access to Search Ads 360 -- as I said, it was called

19 DoubleClick for Search for a good chunk of time, was also

20 tightly restricted in terms of who could see what. There were

21 good processes in place, both on the product side as well as

22 the business side.

23 Q. What was the message to advertisers from Google in

24 terms of how SA360 would be operated?

25 A. This product operates in your best interest, and

 3730

1 things like distribution of spend across platforms was --

2 again, those decisions had to be based on sort of what was best

3 for advertisers. This is not to say that you don't have

4 practical problems, you know, like there's often conflict about

5 what features AdWords had versus what features Bing had, and

6 how do you represent them, and like what sequence do you do

7 product development in and things like that. So I won't

8 pretend that these trade-offs are easy, but we really tried

9 hard to make sure that we always did the right thing for the

10 advertiser and clearly represented to them that that was our

11 obligation.

12 MR. CAVANAUGH: Thank you. Nothing further, Your Honor.

13 THE COURT: We'll start with Google's examination.

14 CROSS-EXAMINATION OF SRIDHAR RAMASWAMY

15 BY MR. SMURZYNSKI:

16 Q. Good afternoon, Dr. Ramaswamy. Do you still have in

17 front of you the white binder from the United States?

18 A. Yes, sir.

19 Q. And one of the documents in that binder is behind

20 UPX123?

21 A. Yes, sir.

22 Q. That was a 2007 document I think we identified?

23 A. That's right.

24 Q. Could you turn to slide 18, which is on the page

25 ending at 484. Do you have that in front of you?

 3731

1 A. Yes, sir.

2 Q. And this slide shows with regards to the three

3 different search engines -- Google, Yahoo! and MSN -- the

4 percentage of searches that are coming through the home page;

5 is that right?

6 A. Uh-huh.

7 Q. And so, for example, this shows that -- going from

8 right to left, for MSN, 71 percent of all their searches came

9 through MSN being set as a home page; is that right?

10 A. Uh-huh.

11 Q. And then Yahoo! 72 percent, and then Google

12 24 percent; is that right?

13 A. Uh-huh.

14 Q. And did that indicate to you at that time that most of

15 Google's traffic was coming through some other mechanism as

16 opposed to relying on a home page default?

17 A. Yeah, absolutely. That means that a good chunk of

18 traffic was coming from people typing Google.com into the

19 browser bar. What is hard to tell from here is whether this is

20 24 percent of users or 24 percent of queries. What this slide

21 definitely shows is that a number of people explicitly did type

22 Google.com into the URL bar and then did their search.

23 Q. And whether the metric was users or number of queries,

24 the majority of that traffic measured either way came to Google

25 through somebody taking an affirmative act as opposed to being

 3732

1 the default?

2 A. That is correct.

3 Q. And in contrast, Yahoo! and MSN were relying on these

4 home pages for the majority of the traffic that they received,

5 again, whether measured through queries or users?

6 A. That's correct.

7 Q. Okay, thank you. You can put that one away. You

8 mentioned earlier in your testimony that Neeva raised capital

9 from, among other institutions, Greylock and Sequoia Ventures;

10 is that right?

11 A. That's correct.

12 Q. And you would consider those to be some of the most

13 sophisticated venture capital funds in Silicon Valley?

14 A. Definitely.

15 Q. And you invested your own money as well in Neeva,

16 correct?

17 A. That's correct.

18 Q. And not only your own money, but your time?

19 A. That's right.

20 Q. At the outset, when Neeva was considering what to

21 build, one of the things Neeva considered was enterprise

22 search; is that right?

23 A. That's correct.

24 Q. And could you explain what enterprise search means to

25 the Court?
 3733

1 A. So an enterprise search product is providing a search

2 facility over the different applications that are used by

3 business users. This could be Dropbox, this could be Google

4 Enterprise, things like Google Docs. It can also be the

5 Microsoft Office Suite which is very popular. And so the idea

6 behind an enterprise search engine was to provide a single

7 aggregation point that you could use to search across these

8 different applications. More and more companies, over the last

9 10 years, end up using many different tools with overlapping

10 needs. We thought there would be benefit to having a single

11 place in which these different applications could be searched.

12 Q. And the technology that you began to develop at Neeva

13 in 2019 could be used either for that enterprise search

14 function or for a consumer search experience?

15 A. Yes and no. There were some technologies that were

16 applicable, but there was quite a bit that was not. But the

17 work that we did in -- this was early 2019, we certainly kept

18 on and offered it as a feature for consumers also. As I said,

19 like you and I have the same problem finding things -- our

20 personal things online as the average enterprise user, so we

21 kept that technology. But things like running a crawl at scale

22 or running a search index at scale often has like -- you know,

23 it's a thousand times larger. So some of it is applicable, but

24 we also had to build a lot of new things starting from 2019.

25 Q. Certainly. And at some point, Neeva made the decision

 3734

1 to focus on consumer search as opposed to enterprise search?

2 A. That's correct.

3 Q. And that was late 2019 or so?

4 A. That is correct.

5 Q. Now, Neeva launched to the public, I think we saw

6 earlier, in June of 2020; is that right?

7 A. That's right.

8 Q. And the concept was to allow users to try Neeva for

9 free and then encourage them to purchase a subscription; is

10 that right?

11 A. That's correct.

12 Q. What was the peak number of unique monthly users of

13 Neeva during the course of its existence?

14 A. I think we especially -- sorry, we started initially

15 with needing to sign up for an account meaning that you needed

16 to create -- have like a user ID and a password, which is

17 modestly high bar for a new user that wants to try a search

18 engine. And our initial expectation -- well, our sort of

19 initial positioning of the product was that you could get a

20 certain number of months free, and that you would need to pay

21 after that. But the more we looked at our data, the more we

22 came to the conclusion that having users sort of stay on on a

23 semi-permanent basis was sort of the thing that really would

24 most indicate that, and how many of them did things like set us

25 as the default search.

 3735

1 But those were the most dominant factors in our success.

2 So we progressively made the product more and more accessible,

3 including without setting up -- you know, without setting up an

4 account. I would -- but a side effect of allowing anonymous

5 usage, for example, is that it becomes pretty tricky to define

6 precisely a user, because a user can come use the product for

7 some time, go and then come back. And we didn't set long term

8 cookies, and it would not be all that easy. Anyway, with the

9 caveats on methodology, I would say that we peaked at like

10 several million users a month; we never crossed 10,000,000.

11 Q. Now, the Neeva business model depended on overcoming

12 something you've described as the power of zero; is that right?

13 A. That is correct. That is a phrase popularized by Dan

14 Ariely who is a behavioral economist.

15 Q. And the power of zero refers to the fact that people

16 like stuff that's free?

17 A. Yes.

18 Q. And so you not only had at Neeva to get people to look

19 at your search engine and try it, you had to get them to pay

20 for it to overcome that power of zero, yes?

21 A. Yes and no. As I said earlier, we got pretty

22 comfortable with how we were able to convert users -- convert

23 free users to paying ones. The dominant factor really was how

24 many free users we had effectively. And unsurprisingly, the

25 people that turned into paying subscribers were the ones that
 3736

1 cared about privacy, cared about things like not having ads in

2 their results or sometimes not exposing their children to an

3 ads-free -- a world full of ads. And as I said, our

4 fundamental business problem was getting a lot of people to try

5 it. And we were very clear, in the last year of Neeva's

6 existence, that we were a premium product meaning there was no

7 obligation to pay ever, if that is what you wanted to do. But

8 our conversion rates were in the several digit percentage

9 points. And we could work out the math and say like even at a

10 4 or 5 percent conversion from free to paid, we could have a

11 sustainable business.

12 THE COURT: Could or could not? I'm sorry, you could?

13 THE WITNESS: We could have a sustainable business,

14 because at scale, the cost of serving search can be lowered

15 quite a lot. So something like one to 20 paid users is a

16 stable equilibrium in terms of trying to run a business.

17 BY MR. SMURZYNSKI:

18 Q. Now, somebody who subscribed to Neeva would create an

19 account; is that right?

20 A. Yes, sir.

21 Q. And you offered those users the ability to connect

22 their Neeva account to various other types of accounts. I

23 think you may have mentioned their Google account and the like?

24 A. So we actually offered this to all account users.

25 There were additional products, like Notion and Figma,

 3737

1 specialized ones that were reserved for the premium paying

2 users. But we actually -- I mean, these are lines that are

3 informed by data, and so we would let even free users connect

4 Google Drive and Office 365 to Neeva.

5 Q. And you also spoke about personalization. Did you

6 believe that personalization improved the quality of the search

7 experience for users?

8 A. Absolutely. So to give you concrete examples, we had

9 a feature called preferred providers by which people could pick

10 their news sources. Among other things, I have subscriptions

11 to like New York Times and Economist, and I prefer those

12 results. All other things being equal, a search engine just

13 becomes more useful with information like that. And the user

14 feedback that we got was definitely that users liked features

15 like that.

16 Q. Neeva was positioned as a private search engine,

17 correct?

18 A. Startups go through iterations, but certainly that's

19 the positioning that we started with.

20 Q. You're familiar with DuckDuckGo?

21 A. I am.

22 Q. Is it your understanding that DuckDuckGo also

23 positions itself as a private search engine?

24 A. Yes, they do.

25 THE COURT: I'm sorry, when you say private search

 3738

1 engine --

2 MR. SMURZYNSKI: Private, yes.

3 THE COURT: -- what do you mean by private search engine?

4 As opposed to a public company? I'm not sure I understand.

5 BY MR. SMURZYNSKI:

6 Q. Well, fair enough, Your Honor. Let me ask a better

7 question. Do you understand that DuckDuckGo positions itself

8 as offering its users privacy?

9 A. They don't describe themself as a private search

10 engine, so he's absolutely correct. But that word, when

11 attached to a search engine, is meant to indicate a search

12 engine that protects your privacy.

13 Q. Does DuckDuckGo's conception of the privacy that it

14 offers impact the quality of the search experience on

15 DuckDuckGo, in your view?

16 A. You know, I'm not sure I should be in the business of

17 criticizing DuckDuckGo. I don't understand the question. I

18 mean, yes, I can talk about it. How can I be helpful without

19 like...

20 Q. Well, let me ask you this question and see if you can

21 answer it. Are there things that are out of scope for

22 DuckDuckGo because of the approach it takes to privacy that

23 impact its search quality, in your view?

24 MR. DINTZER: Objection, Your Honor, relevancy.

25 THE COURT: I think the question is relevant, I just don't

 3739

1 know that -- do you understand what he's asking?

2 THE WITNESS: I do.

3 THE COURT: Okay, go ahead and answer it.

4 THE WITNESS: I'll give you a short answer. As I said, I

5 don't think I should be in the business of commenting on

6 DuckDuckGo's search. But having said that, if they are showing

7 ads within their search experience -- you cannot show ads today

8 without also being in the business, directly or indirectly, of

9 making sure that somebody knows whether those ads were

10 successful or not. While DuckDuckGo can reasonably say that

11 they don't maintain profiles of users, somebody is on the basis

12 of the ads that are being shown on them.

13 BY MR. SMURZYNSKI:

14 Q. Now, at the outset, when Neeva launched, it used Bing

15 to provide its search results; is that right?

16 A. Yes, sir.

17 Q. And in the course of your work, you became familiar

18 with the quality of Bing's search results?

19 A. Yes, sir.

20 Q. And in 2020, when Neeva was using Bing's search

21 results, it's correct that you believed that Bing's search

22 quality did not match Google's?

23 A. That is correct. There was that, and the fact that we

24 felt that we could not innovate as much as we wanted to on top

25 of an API were major factors in us deciding to essentially

 3740

1 build our scale search ourselves. And that effort started in

2 late 2019.

3 THE COURT: I'm sorry, late 2019?

4 THE WITNESS: Late 2019.

5 THE COURT: And that was essentially the building of your

6 own search engine?

7 THE WITNESS: Yes.

8 BY MR. SMURZYNSKI:

9 Q. As you were modeling Neeva's future, Neeva believed

10 that because of the egregious fails in Bing's search results,

11 that users would churn off Neeva at a rate of at least

12 10 percent in a few weeks because of that quality problem at

13 Bing, correct?

14 A. I don't recall the specific numbers, but I can talk

15 generally about egregious fails and churn, if you'd like.

16 Q. Well, let me refresh your recollection.

17 Dr. Ramaswamy, you have in front of you a document that

18 indicates it's an e-mail exchange between you and a gentleman

19 at Microsoft from August of 2020. Do you see that?

20 A. Yes, sir.

21 Q. And I won't ask you to read it aloud, but if you could

22 read the portion there in the third paragraph referring to the

23 churn model, and let me know when you've finished reading that.

24 (Witness reads document)

25 A. Uh-huh.
 3741

1 Q. Does that refresh your recollection that in 2020 --

2 A. Yes, it does.

3 Q. It does, okay. And in 2020, what was Neeva's belief

4 as to the amount of churn it would experience because it was

5 relying on Bing's search results that you had characterized as

6 including egregious fails?

7 A. So the context to have when you're dealing with search

8 experiences are one dominant player that has a reputation for

9 excellence, which is Google search. To this day, it has a good

10 reputation for excellence. What happens is that whenever there

11 is an egregious fail, you can think of it as a confidence

12 busting result page where you try a certain query and you don't

13 think it has the results that you want. What happens to

14 smaller players -- Bing, definitely Neeva, is that that user is

15 likely to go try that search on Google. The important thing to

16 understand here is that if Google were to have similar

17 egregious experiences, even at the same ratio, people generally

18 blame themselves and try and reformulate their query. This is

19 a part of being a dominant player.

20 And so what I was pointing out, we constructed churn

21 models. There were a number of things that caused churn,

22 including the infamous did you want to switch it back prompt

23 that Google Chrome would throw up. But what we came -- the

24 conclusion that we came to here is -- I don't remember the

25 exact numbers -- is that if you assume -- if you make

 3742

1 assumptions about how many queries users are likely to do, and

2 then you measure the number of egregious fails on Bing and ergo

3 on Neeva as well, these egregious fails would often result in

4 churn because that user was likely to go and try the same query

5 on Google and perhaps decide to switch over.

6 The point that I am making here, if you look at the last

7 sentence in this e-mail that you showed me, is that at the same

8 level of egregiousness, at the same level of good and bad

9 quality on Bing and Google, us relying on Bing would cause a

10 certain amount of -- would cause 10 percent of churn. And my

11 point in constructing this model and then explaining it to

12 Rajesh, who was the EVP running search -- he might still be --

13 at Microsoft, is that they were leaving behind a

14 multibillion-dollar revenue opportunity because they were

15 churning users by virtue of being number two and having these

16 egregious fails.

17 Q. And in this e-mail, you're politely saying if their

18 level of quality is the same. But you testified five minutes

19 ago you believed that Bing's quality was inferior to Google's

20 in 2020, correct?

21 A. No, no, my point is -- this is a hypothetical, and the

22 hypothetical is that even at the same level of quality at Bing,

23 between Bing and Google, they were going to lose users on a

24 steady basis because of this user effect of if I get a bad

25 result on a Google or a Neeva, I'm going to go try it on -- if

 3743

1 I were to get a bad result on Bing or Neeva, I'm going to go

2 try out Google. It doesn't happen the other way. I stand

3 behind my saying that there were cases, especially when it came

4 to queries around software engineering, deeply technical

5 queries, where we were pretty clear -- we sent them

6 spreadsheets -- where we said Bing quality has to get higher.

7 THE COURT: Mr. Smurzynski, we're a few minutes past 5:00.

8 What's your timing?

9 MR. SMURZYNSKI: This is a reasonable place to stop. I

10 have a decent amount of material left, it's not five or 10

11 minutes or something like that.

12 THE COURT: So, Mr. Ramaswamy, we're going to conclude for

13 the day which means you'll have to return in the morning.

14 THE WITNESS: Okay.

15 THE COURT: So we appreciate that. So we'll start

16 tomorrow at 9:30. I will just ask you to heed the same

17 instruction, which is not to discuss your testimony with anyone

18 overnight, and we'll see you in the morning. Thank you very

19 much.

20 THE WITNESS: Thank you, sir.

21 THE COURT: So we'll conclude with Mr. Ramaswamy tomorrow

22 morning, and then we're on track to continue to move forward,

23 correct?

24 MR. DINTZER: I'm sorry, Your Honor, I couldn't make

25 out --
 3744

1 THE COURT: I'm sorry, I was both talking and searching at

2 the same time. I'm sorry.

3 MR. DINTZER: Could we hand up another schedule, would

4 that help?

5 THE COURT: I was just looking for it, it was buried under

6 the binders. So then we've got Mr. Lowcock and then Mr. Juda,

7 is it?

8 MR. DINTZER: Juda, yes, Your Honor.

9 THE COURT: So that's where we are tomorrow. So just a

10 couple housekeeping matters. I think last Thursday, I talked

11 to the parties about proposing redactions to the Weinberg,

12 Giannandrea and Cue transcripts that are under seal. Where do

13 we stand with that?

14 MS. BELLSHAW: Yes, Your Honor. We've exchanged proposed

15 redactions with Google and with Apple, and we're just waiting

16 for final sign off, but we're prepared to file tonight.

17 THE COURT: So you just want to submit them to chambers,

18 and we'll obviously take a look at them, and if I have any

19 questions about it, I'll bring it to you all before.

20 MS. BELLSHAW: Yes, we can do that. Thank you.

21 THE COURT: So in terms of timing, then, you all tell me

22 what's reasonable. I'll give you a few days on this one, but

23 I'd like to sort of turn to the next set. And just going in

24 order, we have Mr. Kolotouros, Mr. Higgins and then

25 Mr. Dischler. Could I get proposals by week's end for those

 3745

1 three?

2 MR. DINTZER: Your Honor, we're looking at about -- give

3 us like a three-day turnaround time for us to figure it out,

4 communicate. So if the Court would -- I understand, I can see

5 where this is headed, and we're happy of course to -- but three

6 days per set would be helpful for us.

7 THE COURT: Right. So I had proposed week's end, so on

8 Thursday instead?

9 MR. DINTZER: Friday works, Your Honor. I didn't catch

10 the week's end. Friday works, Your Honor. We appreciate that.

11 THE COURT: So we'll do Friday. So we'll continue to get

12 these on a rolling basis and then unseal those transcripts. Is

13 there anything else we need to talk about before we adjourn for

14 the evening?

15 MR. CAVANAUGH: Your Honor, one small matter. When

16 Mr. Sallet was examining Mr. Tinter, he showed him Exhibit 715

17 and offered it into evidence, but I don't think Your Honor said

18 it was admitted.

19 THE COURT: Okay.

20 MR. SCHMIDTLEIN: This was the draft e-mail problem at the

21 top. I'm not sure if the witness was able to the confirm what

22 the sender unidentified actually meant.

23 THE COURT: My recollection is -- and maybe I'm wrong. My

24 recollection is he thought -- or somebody thought, maybe it was

25 either him or counsel, that what it meant is that he had

 3746

1 drafted it but not sent it.

2 MR. CAVANAUGH: Yes, and I think the only questioning went

3 to Mr. Tinter's e-mail below.

4 MR. SCHMIDTLEIN: If that's the case, then we don't have

5 an objection to the first e-mail in time.

6 MR. CAVANAUGH: And as to the second one, I think

7 Mr. Tinter said it was drafted, but I don't think -- we have

8 not represented that it was sent.

9 THE COURT: Right, okay. So with that caveat, 715 will

10 come in.

11 (Exhibit UPX715 admitted into evidence)

12 MR. SCHMIDTLEIN: Thank you.

13 MR. CAVANAUGH: Thank you, Your Honor.

14 THE COURT: Anything else, folks?

15 MR. DINTZER: Not from the DOJ plaintiffs, Your Honor.

16 THE COURT: Mr. Cavanaugh?

17 MR. CAVANAUGH: No, Your Honor.

18 MR. SCHMIDTLEIN: No, Your Honor.

19 THE COURT: Thank you, everybody. We'll see you in the

20 morning.

21 (Proceedings adjourned at 5:09 p.m.)

22

23

24

25
BY MR. CAVANAUGH: - absolutely 3748

3656/24 3657/3 0 2017 [1] 3670/4 492 [1] 3685/12

3659/16 3660/1 2018 [8] 3653/7
BY MR. 015 [3] 3648/7 3648/8 5
3660/3 3660/5 3661/3 3654/21 3667/5
CAVANAUGH: [3] 3650/13
3661/6 3664/16 3668/12 3669/13 5 percent [1] 3736/10
3714/24 3718/4
3724/24
3665/1 3665/3 3665/7 1 3701/23 3704/7 50 [2] 3709/6 3719/7
3665/13 3665/16 3704/14 50 percent [1] 3672/19
BY MR. DINTZER: [9] 10 [4] 3663/5 3689/6
3665/20 3665/24 2019 [12] 3637/4 50-ish [1] 3676/21
3666/12 3677/11 3733/9 3743/10
50-person [1] 3692/19
3681/5 3683/11
3666/9 3675/19 10 percent [3] 3643/15 3640/23 3670/5
3675/21 3675/25 3672/8 3682/11 501 [1] 3652/17
3689/13 3701/13 3740/12 3742/10
502 [2] 3656/22
3706/19 3710/13
3677/10 3679/16 10,000,000 [2] 3636/1 3733/13 3733/17
3680/13 3681/4 3733/24 3734/3 3656/23
3714/1 3735/10
3682/18 3683/5 3740/2 3740/3 3740/4 524 [1] 3637/14
BY MR. 100 [1] 3634/12
3689/1 3689/5 2020 [6] 3734/6 527 [1] 3688/7
SCHMIDTLEIN: [7] 10036 [2] 3631/25
3689/12 3697/24 3739/20 3740/19 5:00 [1] 3743/7
3634/3 3637/15 3632/12
3698/2 3699/12 3741/1 3741/3 5:09 p.m [1] 3746/21
3648/2 3653/4 3657/6 1023 [1] 3721/16
3699/16 3701/12 3742/20
3659/18 3660/7
3705/24 3706/4
1095 [1] 3632/11
2021 [1] 3672/11 6
BY MR. SEVERT: [1] 1100 [1] 3631/14
3706/7 3706/16 2022 [2] 3635/10 600 [1] 3631/20
3661/9 1133 [1] 3631/24
3706/18 3710/8 3671/22 60604 [1] 3631/20
BY MR. 123 [1] 3707/1
3710/11 3712/21 2023 [2] 3631/5 65 [1] 3676/15
SMURZYNSKI: [5] 1300 [1] 3632/4
3712/25 3713/3 3725/21 680 [2] 3632/8 3637/9
3730/14 3736/16 14 [4] 3631/7 3648/7
3713/25 3714/22 209 [1] 3631/20
3738/4 3739/12
3717/17 3723/18
3685/11 3703/18
20s [1] 3643/23 7
3740/7 15 [1] 3695/14 70 percent [1] 3676/15
3723/21 3724/23 16 [4] 3644/19 21 [1] 3709/17
DEPUTY CLERK: [2] 71 percent [1] 3731/8
3730/12 3736/11 3691/13 3710/20 2200 [2] 3631/24
3666/3 3666/8 715 [2] 3745/16
3737/24 3738/2 3711/12 3631/24
MR. CAVANAUGH: [8] 3746/9
3738/24 3739/2 18 [1] 3730/24 22nd [1] 3707/21
3665/2 3723/20 72 percent [1] 3731/11
3740/2 3740/4 3743/6 1:20-cv-3010 [1] 24 percent [8] 3657/22
3730/11 3745/14 736 [3] 3657/3 3657/5
3743/11 3743/14 3631/4 3658/2 3658/15
3746/1 3746/5 3657/6
3743/20 3743/25 1:36 [1] 3631/6 3658/20 3659/2
3746/12 3746/16 7th [1] 3632/4
3744/4 3744/8 3731/12 3731/20
MR. DINTZER: [24]
3665/9 3665/15
3744/16 3744/20 2 3731/20
8
3745/6 3745/10 20 [3] 3645/21 27th [1] 3632/12
3665/18 3665/22 800 [1] 3726/16
3745/18 3745/22 3695/20 3736/15 2:30 [1] 3656/18
3665/25 3675/20 80203 [1] 3632/4
3746/8 3746/13 20 percent [2] 3643/15 3
3682/12 3682/20
3746/15 3746/18
3683/2 3689/4
THE WITNESS: [27]
3695/14
30 percent [2] 3643/17
9
3689/11 3698/1 2000 [1] 3699/9 93 [1] 3707/3
3656/21 3657/2 20001 [2] 3631/18 3645/21
3706/2 3706/6 940 [2] 3682/3 3683/9
3657/5 3664/14 3632/25 3010 [1] 3631/4
3706/14 3706/17 9:30 [1] 3743/16
3664/19 3665/6 20005 [1] 3631/15 333 [1] 3632/24
3714/20 3738/23
3666/7 3675/24 360 [3] 3727/11
3743/23 3744/2 2000s [1] 3705/7
3728/16 3729/18
A
3676/4 3679/20 20024 [1] 3632/8
3744/7 3745/1 3745/8 ability [11] 3643/2
3680/15 3698/4 2003 [3] 3667/4 3634 [1] 3633/4
3746/14 3659/22 3680/15
3699/15 3699/19 3667/12 3668/20 3637 [2] 3633/15
MR. LARRABEE: [1] 3691/18 3697/10
3710/10 3712/24 2007 [6] 3668/24 3633/16
3656/15 3697/12 3701/1
3713/1 3713/7 3707/9 3707/15 3648 [1] 3633/17
MR. SCHMIDTLEIN: 3713/6 3713/6
3717/18 3724/5 3707/21 3708/14 365 [1] 3737/4
[13] 3637/10 3647/22 3716/10 3736/21
3736/12 3739/1 3730/22 3653 [1] 3633/18
3656/19 3656/23 able [23] 3643/5
3739/3 3740/3 3740/6 2007-8 [1] 3698/16 3661 [1] 3633/5
3656/25 3660/2 3643/9 3656/8
3743/13 3743/19 2008 [2] 3663/8 3666 [1] 3633/8
3660/4 3661/5 3659/23 3673/8
3667/23 3683 [1] 3633/19
3664/13 3745/19 $ 3715 [1] 3633/9 3673/17 3674/13
3746/3 3746/11 2009 [5] 3641/3 3674/24 3674/24
$15 [1] 3708/21 3641/21 3642/13 3730 [1] 3633/10
3746/17 3676/4 3679/17
$5 [1] 3676/14 3642/25 3643/25 3746 [1] 3633/20
MR. SEVERT: [6] 3679/19 3679/23
$50 [3] 3676/14 2010 [4] 3647/14 3:00 [1] 3689/6
3637/7 3637/12 3690/5 3690/19
3676/16 3679/15 3648/12 3654/17 3:09 p.m [1] 3689/10
3647/24 3652/20 3695/25 3697/2
3655/11 3:25 [1] 3689/6
3652/24 3664/25 ' 3:27 p.m [1] 3689/11 3699/18 3700/18
MR. SMURZYNSKI: 2010s [1] 3663/16 3719/8 3720/21
'09 [1] 3644/4
[5] 3682/13 3682/24 '16 [1] 3644/15 2011 [1] 3667/25 4 3735/22 3745/21
3706/8 3738/1 3743/8 2012 [1] 3660/9 above [2] 3712/25
40 [1] 3636/2
MS. BELLSHAW: [2] . 2013 [2] 3663/8
40 percent [2] 3693/22 3747/5
3744/13 3744/19 3668/3 above-entitled [1]
......Admitted [6] 3698/14
THE COURT: [83] 2014 [3] 3668/5 3747/5
3633/15 3633/16 450 [1] 3631/17
3634/1 3637/8 3668/11 3668/12 absence [1] 3678/2
3633/17 3633/18 481 [1] 3683/15
3637/13 3647/25 2015 [1] 3644/15 absolutely [9] 3657/17
3633/19 3633/20 484 [1] 3730/25
3652/22 3652/25 2015-16 [1] 3644/19 3663/9 3664/4
487 [1] 3709/18
3656/20 3656/22 2016 [1] 3668/13 3677/15 3689/5
49 [1] 3688/6
absolutely... - Apple 3749

A ad-supported [2] 3699/15 aggregation [1] 3733/7 3632/11 3703/8

3699/10 3716/24 advertise [4] 3646/11 ago [5] 3669/14 AMIT [1] 3631/10
absolutely... [4]
ADAM [1] 3631/16 3649/11 3649/11 3699/21 3710/20 among [5] 3687/14
3697/20 3731/17
adamant [2] 3713/9 3649/18 3711/12 3742/19 3688/19 3696/25
3737/8 3738/10
3713/11 advertiser [8] 3649/10 agree [4] 3651/23 3732/9 3737/10
absorbed [1] 3675/15
AdCentral [1] 3649/13 3654/7 3667/18 3716/20 3728/9 amount [7] 3636/6
abstract [1] 3708/18
add [1] 3643/22 3707/18 3728/6 3728/13 3669/10 3674/19
access [9] 3690/14
added [2] 3668/24 3729/2 3729/9 agreement [5] 3641/3 3694/5 3741/4
3690/14 3714/3
3696/22 3730/10 3654/18 3654/23 3742/10 3743/10
3714/5 3714/8
addition [4] 3661/11 advertiser's [1] 3671/7 3655/4 3713/21 analogy [1] 3660/1
3724/10 3729/5
3668/13 3697/2 advertiser-facing [1] agreements [2] analyses [1] 3653/16
3729/6 3729/18
3702/13 3707/18 3713/13 3713/16 analysis [1] 3653/24
accessibility [1]
additional [6] 3642/17 advertisers [13] ahead [5] 3653/3 analysts [1] 3708/3
3689/21
3694/13 3696/23 3649/15 3652/4 3664/19 3689/3 analyzed [1] 3646/21
accessible [1] 3735/2
3721/20 3722/10 3652/8 3652/12 3709/22 3739/3 analyzing [1] 3653/19
account [6] 3734/15
3736/25 3668/9 3669/1 AI [31] 3638/8 3638/11 Android [7] 3692/3
3735/4 3736/19
address [3] 3680/24 3688/24 3699/2 3638/15 3639/8 3692/4 3692/18
3736/22 3736/23
3682/25 3700/19 3717/1 3729/4 3639/11 3639/24 3692/24 3698/14
3736/24
addressed [1] 3688/10 3729/11 3729/23 3640/5 3640/12 3698/18 3705/11
accounts [2] 3635/5
addresses [1] 3680/17 3730/3 3640/15 3640/25 anecdotal [2] 3685/25
3736/22
adjourn [1] 3745/13 advertising [24] 3642/7 3661/22 3671/22 3692/7
accumulate [1] 3636/1
adjourned [1] 3746/21 3649/8 3649/19 3672/4 3674/5 3674/7 angles [1] 3661/1
acquire [3] 3676/4
admissibility [1] 3649/21 3651/15 3674/10 3674/17 announcement [1]
3700/5 3700/8
3683/10 3651/16 3651/21 3675/17 3696/15 3635/8
acquired [4] 3675/5
admission [1] 3652/20 3651/23 3651/25 3697/2 3697/4 3697/5 annual [1] 3676/2
3727/21 3727/25
admit [3] 3653/1 3652/1 3653/14 3697/7 3697/14 anonymous [1] 3735/4
3728/2
3718/13 3724/6 3653/21 3654/3 3697/15 3697/20 answered [1] 3725/13
acquiring [2] 3677/3
admitted [11] 3637/9 3654/6 3654/8 3698/9 3700/23 Antitrust [1] 3632/3
3699/21
3637/10 3637/14 3667/22 3668/8 3701/10 3719/8 API [2] 3671/18
acquisition [7] 3668/23
3637/15 3648/1 3668/18 3669/6 3719/13 3739/25
3675/2 3675/10
3648/2 3653/4 3683/9 3669/8 3673/25 AI-powered [2] APIs [1] 3681/1
3698/14 3701/12
3683/11 3745/18 3703/6 3722/20 3674/17 3697/4 apologies [1] 3657/8
3727/19 3727/24
3746/11 3728/6 aiming [1] 3672/23 apology [1] 3660/11
across [6] 3662/15
ads [60] 3645/14 AdWords [5] 3649/14 al [1] 3631/3 app [21] 3634/24
3691/16 3728/7
3667/8 3667/13 3667/19 3728/12 algorithm [3] 3720/19 3635/6 3635/6 3635/9
3729/2 3730/1 3733/7
3667/13 3667/20 3728/19 3730/5 3720/25 3721/4 3635/14 3635/22
act [1] 3731/25
3667/24 3668/3 affect [3] 3640/1 aligned [2] 3674/15 3635/24 3636/10
action [3] 3631/3
3668/6 3668/13 3640/6 3708/23 3676/7 3636/12 3636/15
3690/3 3724/15
3668/22 3668/25 affected [2] 3678/22 allow [1] 3734/8 3692/13 3692/16
active [5] 3634/12
3669/3 3671/14 3691/19 allowing [1] 3735/4 3692/25 3693/12
3634/17 3649/15
3672/2 3672/17 affecting [1] 3640/2 almighty [1] 3725/22 3694/10 3694/11
3666/25 3667/2
3673/23 3678/13 affirm [1] 3666/5 almost [4] 3635/9 3694/15 3716/9
activity [1] 3636/9
3678/15 3678/21 affirmative [1] 3731/25 3682/9 3687/18 3716/9 3720/3
actual [4] 3662/4
3679/6 3686/25 afford [1] 3721/11 3690/2 3721/19
3677/4 3693/7
3687/1 3687/3 3694/4 affordable [1] 3676/13 along [5] 3669/10 app's [1] 3635/2
3693/11
3698/24 3700/5 afternoon [4] 3631/7 3670/1 3685/4 appear [3] 3685/12
actually [16] 3666/1
3702/2 3702/17 3689/3 3715/1 3698/23 3703/10 3701/16 3717/7
3668/20 3672/18
3702/19 3702/21 3730/16 aloud [1] 3740/21 APPEARANCES [2]
3675/1 3675/14
3703/1 3707/12 again [25] 3637/6 alternative [5] 3654/24 3631/12 3632/1
3677/6 3678/18
3707/13 3707/17 3638/20 3639/19 3655/22 3669/22 appearing [1] 3674/1
3684/6 3684/22
3707/19 3708/5 3640/23 3642/13 3693/19 3718/1 appears [3] 3653/9
3701/17 3707/22
3715/18 3716/12 3647/7 3647/9 alternatives [2] 3655/9 3657/4 3703/19
3711/17 3727/9
3716/16 3716/17 3649/13 3651/1 3718/18 Apple [35] 3635/6
3736/24 3737/2
3716/18 3717/1 3651/25 3653/19 although [1] 3658/4 3654/15 3654/18
3745/22
3718/17 3718/23 3657/10 3665/14 always [8] 3679/21 3655/4 3655/14
ad [18] 3649/13
3720/9 3720/24 3665/18 3678/1 3680/5 3680/7 3680/9 3655/14 3655/18
3652/9 3668/16
3721/11 3722/23 3681/9 3685/12 3692/6 3699/24 3655/18 3655/19
3671/7 3673/1
3727/11 3728/12 3687/25 3688/11 3711/1 3730/9 3655/25 3656/2
3678/10 3678/12
3728/16 3728/17 3694/6 3699/7 Amazon [8] 3646/4 3656/9 3658/21
3699/10 3703/3
3729/18 3736/1 3701/16 3704/12 3646/17 3649/20 3659/3 3659/5 3659/7
3703/6 3715/24
3736/3 3736/3 3739/7 3730/2 3732/5 3649/25 3652/11 3659/10 3659/13
3716/2 3716/24
3739/7 3739/9 against [4] 3645/8 3652/22 3653/25 3659/24 3660/1
3718/14 3719/1
3739/12 3645/14 3646/3 3654/7 3660/5 3660/9
3721/6 3727/22
ads-free [3] 3671/14 3681/4 Amazon's [3] 3647/3 3660/11 3660/14
3728/2
3672/2 3736/3 age [5] 3670/22 3650/1 3653/14 3660/15 3660/20
ad-based [1] 3715/24
AdSense [1] 3668/21 3705/5 3705/6 AMERICA [1] 3631/3 3661/3 3661/4 3664/1
Ad-Driven [1] 3721/6
advantage [2] 3640/16 3705/14 3705/14 Americas [3] 3631/24 3664/11 3664/24
ad-free [1] 3719/1
Apple... - books 3750

A assumptions [1] 3632/24 beholden [1] 3671/6 3641/9 3641/11

3742/1 bar [6] 3711/15 belief [2] 3699/5 3641/11 3641/11
Apple... [4] 3680/22
assure [1] 3725/3 3711/15 3712/5 3741/3 3641/19 3642/13
3691/21 3694/4
attached [1] 3738/11 3731/19 3731/22 Belknap [1] 3631/23 3643/12 3643/16
3744/15
attachment [2] 3734/17 BELLSHAW [1] 3643/20 3643/25
applicable [3] 3726/5
3703/18 3706/22 Bard [2] 3662/3 3631/16 3644/2 3644/4 3644/4
3733/16 3733/23
Attachments [1] 3719/13 below [2] 3684/18 3644/8 3644/10
applications [4]
3703/1 base [4] 3635/21 3746/3 3644/13 3645/2
3693/23 3733/2
attempting [2] 3649/3 3701/1 3701/2 3701/2 BENCH [1] 3631/10 3645/4 3645/8
3733/8 3733/11
3649/4 based [9] 3653/2 beneficial [2] 3727/5 3645/11 3645/12
apply [1] 3640/15
attention [5] 3682/20 3669/17 3672/10 3727/7 3645/14 3645/14
appreciable [1]
3683/7 3686/2 3672/13 3685/2 benefit [3] 3694/6 3646/3 3646/15
3655/17
3686/16 3715/13 3715/24 3725/1 3729/11 3733/10 3646/21 3646/23
appreciate [2] 3743/15
attractive [1] 3640/15 3729/7 3730/2 benefits [3] 3693/16 3647/9 3648/25
3745/10
attribution [1] 3727/8 basic [1] 3714/9 3693/24 3697/19 3649/1 3649/3
approach [3] 3666/1
auction [6] 3649/13 basically [12] 3635/12 best [15] 3670/13 3649/12 3649/15
3687/21 3738/22
3651/18 3652/1 3646/10 3677/1 3679/12 3681/21 3649/21 3649/23
approved [1] 3691/15
3652/2 3652/2 3687/3 3693/14 3687/12 3687/13 3650/2 3650/8 3652/9
approver [1] 3718/13
3667/21 3700/15 3706/3 3695/10 3695/10 3653/11 3654/12
apps [9] 3635/17
auctions [1] 3652/9 3713/15 3720/21 3695/16 3695/22 3655/4 3655/14
3692/20 3692/21
August [1] 3740/19 3720/25 3721/3 3696/25 3697/3 3655/18 3655/20
3692/24 3693/16
autocomplete [3] 3722/8 3700/21 3718/14 3656/10 3658/21
3693/19 3693/23
3693/3 3720/4 3721/1 basis [9] 3634/20 3729/25 3730/2 3658/23 3659/5
3700/6 3723/7
available [8] 3655/5 3635/20 3653/17 bet [2] 3664/6 3698/9 3659/8 3659/8 3659/9
April [1] 3667/4
3671/20 3674/16 3676/2 3716/14 better [17] 3662/6 3659/12 3659/14
April 2003 [1] 3667/4
3675/7 3676/18 3734/23 3739/11 3671/5 3671/11 3659/21 3660/21
arc [1] 3712/16
3690/8 3691/6 3742/24 3745/12 3671/23 3672/25 3661/12 3661/15
area [2] 3695/1
3715/12 Bates [5] 3683/14 3674/11 3681/21 3661/15 3661/19
3700/17
Avenue [4] 3631/24 3685/12 3688/7 3681/23 3694/7 3662/1 3662/2 3662/8
areas [2] 3637/20
3632/8 3632/11 3703/18 3709/17 3696/2 3698/11 3662/9 3663/19
3640/15
3632/24 beautifully [1] 3660/18 3699/24 3700/11 3664/2 3664/10
argument [1] 3658/24
avenues [1] 3669/24 became [8] 3644/11 3700/24 3718/17 3671/18 3676/24
argumentative [1]
average [1] 3733/20 3663/24 3668/2 3719/6 3738/6 3678/11 3678/20
3659/18
avoid [2] 3681/15 3668/11 3669/16 beyond [6] 3645/19 3690/22 3690/24
Ariely [1] 3735/14
3720/1 3728/15 3728/15 3668/22 3676/8 3691/5 3695/13
arose [1] 3655/24
aware [7] 3651/7 3739/17 3683/1 3693/20 3719/13 3721/12
around [19] 3634/14
3651/14 3651/16 become [5] 3648/22 3719/1 3728/7 3728/12
3642/6 3648/6
3651/17 3654/6 3654/7 3676/24 big [4] 3647/8 3671/21 3730/5 3739/14
3654/19 3661/1
3680/7 3691/10 3687/5 3690/4 3674/7 3729/1 3740/13 3741/14
3664/5 3670/10
away [6] 3651/5 becomes [3] 3678/1 bigger [2] 3649/1 3742/2 3742/9 3742/9
3675/17 3678/19
3654/21 3686/12 3735/5 3737/13 3698/25 3742/22 3742/23
3698/12 3700/2
3699/9 3716/18 becoming [1] 3648/10 biggest [2] 3694/19 3743/1 3743/6
3700/10 3701/3
3732/7 began [2] 3650/11 3695/19 Bing's [15] 3642/19
3701/10 3705/8
axes [1] 3703/10 3733/12 Bill [1] 3715/1 3643/2 3645/2
3708/5 3710/22
begin [4] 3646/13 billion [2] 3708/21 3655/25 3656/2
3721/21 3743/4 B 3666/19 3674/9 3719/11 3656/5 3658/1
arrangement [9]
bachelor [1] 3666/21 3674/12 billion-dollar [1] 3662/10 3671/19
3641/5 3641/6
back [22] 3643/11 beginning [1] 3668/22 3719/11 3739/18 3739/20
3642/25 3643/24
3645/10 3647/7 beginnings [1] 3711/19 billions [4] 3644/18 3739/21 3740/10
3726/17 3726/18
3652/10 3656/14 begins [1] 3650/12 3644/21 3644/23 3741/5 3742/19
3726/19 3726/20
3659/14 3663/11 begun [1] 3647/5 3664/6 bipolar [1] 3705/11
3727/6
3663/20 3663/21 behalf [3] 3664/25 binder [17] 3636/19 bit [8] 3648/23
article [4] 3684/23
3677/9 3678/19 3665/2 3726/9 3647/13 3652/16 3670/21 3687/6
3685/1 3718/6 3718/9
3689/11 3697/3 behavior [8] 3647/11 3652/18 3656/25 3688/16 3702/10
artificial [1] 3696/12
3698/13 3708/14 3661/23 3678/14 3666/2 3681/24 3728/18 3729/10
Ask.com [1] 3705/9
3710/7 3710/8 3678/22 3678/24 3697/24 3701/16 3733/16
aspect [1] 3683/10
3714/12 3723/7 3681/13 3696/8 3705/24 3705/24 bits [1] 3680/24
assert [1] 3683/3
3725/15 3735/7 3702/21 3706/17 3706/19 blame [1] 3741/18
assessment [1]
3741/22 behavioral [2] 3696/15 3715/4 3719/18 block [1] 3723/4
3653/13
backdoor [1] 3716/14 3735/14 3730/17 3730/19 blog [1] 3719/19
assist [1] 3725/20
backfill [1] 3654/20 behaviors [1] 3689/22 binders [1] 3744/6 blur [2] 3680/20
assume [4] 3645/16
background [1] behind [11] 3640/25 Bing [101] 3634/7 3680/25
3656/12 3683/21
3666/19 3659/12 3671/4 3634/10 3634/11 board [2] 3667/1
3741/25
bad [3] 3742/8 3675/11 3682/16 3634/15 3634/19 3691/16
assumed [1] 3668/13
3742/24 3743/1 3714/17 3714/19 3634/19 3634/24 book [1] 3700/19
assuming [4] 3643/14
bag [1] 3722/22 3730/19 3733/6 3635/2 3635/9 3635/9 Bookmarks [1] 3655/2
3647/17 3648/13
Bankruptcy [1] 3742/13 3743/3 3636/15 3641/7 books [1] 3684/4
3655/3
bootstrapped - Chrome 3751

B 3720/7 3721/23 bypass [1] 3650/25 3702/15 3704/19 3690/1 3690/20

3725/11 3731/19 3705/15 3705/23 3691/11 3696/17
bootstrapped [1] C
browser's [1] 3685/21 3707/3 3707/15 3698/21 3713/18
3671/18
browsers [12] 3661/20 calculus [1] 3677/8 3708/11 3710/16 3713/19 3726/20
both [20] 3638/9
3691/3 3691/11 calendar [1] 3702/24 3710/18 3712/5 3734/20 3741/12
3646/3 3654/12
3693/10 3693/20 call [6] 3661/16 3713/13 3714/12 3742/10
3662/2 3662/3 3662/6
3701/5 3713/23 3665/10 3669/4 3716/13 3716/14 certainly [11] 3635/3
3663/15 3669/1
3713/25 3714/4 3676/20 3690/22 3716/16 3716/17 3668/19 3682/9
3671/10 3674/14
3723/13 3723/14 3698/15 3716/18 3717/2 3684/7 3691/21
3688/20 3692/24
3724/20 called [24] 3638/4 3717/6 3717/8 3718/8 3720/23 3727/17
3694/12 3703/9
browsing [1] 3722/7 3659/5 3662/4 3718/15 3718/17 3729/6 3733/17
3719/12 3724/13
budget [1] 3719/7 3663/23 3667/19 3718/23 3720/9 3733/25 3737/18
3727/5 3728/25
budgets [1] 3719/11 3667/20 3668/2 3721/7 3722/7 certify [1] 3747/4
3729/21 3744/1
bug [1] 3662/5 3668/8 3668/21 3722/15 3723/19 cetera [1] 3682/16
bottom [11] 3639/2
build [11] 3643/9 3669/16 3670/2 3724/7 3724/11 CFO [2] 3702/20
3683/14 3707/20
3671/17 3674/24 3672/12 3675/12 3725/12 3725/22 3703/13
3707/21 3708/7
3699/15 3700/22 3691/15 3693/13 3726/10 3728/18 chain [3] 3707/6
3715/14 3715/17
3716/13 3723/10 3694/13 3707/11 3732/7 3733/4 3735/6 3707/23 3710/8
3716/7 3716/9 3727/1
3723/24 3732/21 3707/12 3708/2 3736/14 3738/18 challenge [1] 3695/18
3727/8
3733/24 3740/1 3712/18 3723/3 3738/18 3738/20 chambers [1] 3744/17
bought [1] 3678/18
building [2] 3671/14 3727/23 3729/18 3739/10 3740/14 chance [2] 3652/8
boundary [1] 3678/7
3740/5 3737/9 3741/11 3744/20 3659/11
bounty [1] 3713/17
built [7] 3643/6 3653/2 calls [3] 3663/3 3745/4 change [11] 3647/10
box [6] 3690/3
3672/5 3675/16 3726/15 3726/22 Canada [1] 3672/13 3655/6 3655/8
3691/11 3693/3
3675/16 3701/8 came [34] 3639/14 capable [1] 3690/15 3655/10 3655/19
3712/7 3712/17
3722/5 3662/2 3662/3 3669/1 capital [4] 3672/9 3665/9 3668/16
3720/4
bullet [4] 3683/25 3671/21 3672/11 3674/25 3732/8 3669/6 3669/10
Brain [3] 3638/4
3684/8 3684/12 3674/23 3676/18 3732/13 3690/11 3724/11
3638/6 3638/22
3709/23 3677/5 3681/3 3688/2 capture [1] 3695/22 changed [5] 3634/21
brand [8] 3656/5
bunch [1] 3724/21 3691/14 3692/25 care [8] 3681/7 3674/22 3701/3
3659/5 3659/8 3659/9
burden [1] 3693/17 3693/16 3693/17 3688/25 3724/3 3714/5 3714/8
3659/12 3659/13
buried [1] 3744/5 3693/17 3693/22 3724/14 3725/1 changer [1] 3674/7
3659/25 3708/24
business [48] 3640/10 3694/6 3698/14 3725/3 3725/13 changes [3] 3674/20
brands [1] 3645/7
3642/3 3644/14 3702/18 3705/5 3726/11 3698/12 3700/4
break [4] 3634/6
3644/19 3644/25 3705/6 3705/14 cared [3] 3724/4 changing [2] 3689/15
3689/3 3689/9
3647/17 3649/1 3705/14 3720/10 3736/1 3736/1 3700/7
3696/18
3651/15 3651/16 3720/16 3723/1 careful [2] 3657/12 channel [1] 3693/19
breakdown [1] 3705/2
3651/21 3651/23 3729/16 3731/8 3680/21 characteristics [1]
breakdowns [1]
3651/25 3653/14 3731/24 3734/22 carriers [2] 3692/2 3687/23
3703/10
3674/2 3674/14 3741/23 3741/24 3692/2 characterized [1]
breakthrough [2]
3674/24 3675/23 3743/3 case [6] 3650/1 3664/4 3741/5
3671/21 3674/6
3676/5 3684/11 campaign [2] 3647/8 3697/20 3721/12 charge [2] 3669/7
Brief [1] 3666/3
3691/15 3695/9 3652/15 3723/24 3746/4 3680/6
briefly [5] 3667/10
3698/6 3701/7 campaigns [1] 3652/10 cases [2] 3649/15 chase [1] 3705/25
3670/7 3671/2
3702/17 3702/19 can [95] 3634/22 3743/3 Chat [4] 3634/15
3682/21 3686/22
3702/22 3703/3 3636/25 3638/23 casual [1] 3681/11 3662/1 3662/8 3662/9
bring [3] 3711/3
3703/4 3703/11 3640/19 3644/1 catch [1] 3745/9 ChatGPT [8] 3634/7
3711/4 3744/19
3704/25 3708/21 3645/13 3646/13 categories [3] 3650/12 3634/11 3635/24
bringing [2] 3682/20
3713/24 3716/13 3648/6 3655/7 3650/18 3650/20 3636/1 3636/6
3689/19
3721/6 3723/5 3656/17 3658/25 category [3] 3640/6 3636/10 3636/15
broad [2] 3651/25
3723/24 3724/2 3659/1 3660/3 3640/6 3651/6 3662/8
3686/11
3725/25 3729/22 3664/17 3665/14 cause [3] 3705/15 checking [2] 3682/24
broader [3] 3640/4
3733/3 3735/11 3666/4 3667/10 3742/9 3742/10 3683/3
3645/6 3667/25
3736/4 3736/11 3670/10 3670/10 caused [1] 3741/21 Chennai [1] 3666/22
Broadway [1] 3632/4
3736/13 3736/16 3670/11 3670/15 CAVANAUGH [6] Chicago [3] 3631/20
broken [1] 3704/20
3738/16 3739/5 3671/2 3673/3 3676/1 3631/23 3633/9 3729/8 3729/8
brought [1] 3683/7
3739/8 3677/21 3678/3 3714/23 3715/1 Chief [1] 3637/2
Brown [4] 3666/23
businesses [2] 3680/4 3682/1 3683/8 3723/19 3746/16 children [1] 3736/2
3666/25 3667/1
3662/15 3684/9 3683/9 3684/23 caveat [1] 3746/9 choice [6] 3659/25
3667/2
busting [1] 3741/12 3686/12 3687/7 caveats [1] 3735/9 3686/13 3692/17
browser [20] 3658/22
busy [1] 3724/14 3690/7 3690/22 CC [1] 3703/15 3694/15 3708/22
3686/2 3686/5
buy [7] 3670/16 3693/14 3694/9 center [1] 3659/8 3708/23
3686/16 3689/18
3673/5 3687/20 3695/11 3696/3 centered [1] 3721/21 choices [1] 3724/18
3689/20 3690/6
3687/24 3692/4 3696/9 3699/6 3699/7 CEO [3] 3644/11 chosen [1] 3655/14
3690/8 3694/13
3728/6 3728/12 3699/17 3700/3 3668/2 3668/12 Chrome [9] 3690/9
3694/15 3710/23
buyers [1] 3692/3 3700/11 3700/11 certain [14] 3662/16 3690/10 3690/11
3711/1 3711/5 3711/9
buying [1] 3728/11 3701/21 3702/13 3673/4 3681/11 3690/13 3690/14
3712/24 3713/5
Chrome... - COURT 3752

C combined [2] 3644/3 competition [8] congratulate [1] continuously [1]

3644/4 3638/15 3639/24 3660/19 3718/11
Chrome... [4] 3690/17
comfortable [3] 3640/1 3648/11 connect [2] 3736/21 contracts [5] 3691/10
3690/24 3711/3
3690/16 3727/14 3653/20 3660/24 3737/3 3691/13 3691/16
3741/23
3735/22 3717/25 3718/4 connection [4] 3654/24 3691/19 3692/8
chunk [2] 3729/19
coming [3] 3731/4 competitive [1] 3715/9 3715/10 contrast [2] 3670/13
3731/17
3731/15 3731/18 3653/24 3721/7 3732/3
churn [8] 3740/11
comment [1] 3639/10 competitor [1] 3643/8 connectivity [1] control [3] 3691/24
3740/15 3740/23
commenting [1] 3739/5 competitors [3] 3645/2 3663/15 3693/12 3720/3
3741/4 3741/20
commerce [7] 3651/4 3648/19 3709/1 Connolly [1] 3632/7 controlled [2] 3659/21
3741/21 3742/4
3651/4 3651/10 completely [1] 3635/23 consequences [1] 3693/1
3742/10
3668/4 3668/6 completions [2] 3693/5 3715/20 controls [2] 3712/7
churning [1] 3742/15
3668/13 3702/17 3720/8 consider [3] 3676/12 3712/17
circa [1] 3698/15
commercial [36] complex [2] 3692/8 3716/19 3732/12 convenience [2]
circulated [1] 3653/10
3636/9 3645/25 3725/21 considerable [1] 3670/20 3689/21
circulates [1] 3653/6
3646/5 3646/6 3646/7 complicated [5] 3693/17 converged [1] 3705/10
Civil [1] 3631/3
3646/13 3649/14 3642/10 3690/15 considered [2] 3687/10 conversations [9]
claiming [2] 3698/8
3673/6 3673/23 3713/12 3713/20 3732/21 3654/14 3655/25
3724/12
3679/4 3681/20 3719/9 considering [1] 3675/2 3681/12
clean [1] 3698/10
3686/20 3686/22 components [1] 3705/1 3732/20 3691/23 3691/25
clear [5] 3683/2
3686/24 3687/1 comports [1] 3635/12 consistently [1] 3692/1 3721/21
3691/20 3729/15
3687/2 3687/2 3687/3 compose [1] 3700/18 3694/18 3721/24
3736/5 3743/5
3687/8 3687/10 computer [4] 3666/22 consists [1] 3685/2 conversion [2] 3736/8
clearly [2] 3682/23
3687/13 3687/17 3680/25 3710/23 constant [1] 3679/7 3736/10
3730/10
3687/18 3688/20 3711/2 Constitution [1] convert [2] 3735/22
clever [1] 3721/13
3693/10 3697/7 computers [5] 3680/18 3632/24 3735/22
click [13] 3642/23
3711/21 3713/16 3704/23 3705/8 constraints [1] 3673/2 convey [1] 3703/10
3650/4 3674/8 3680/3
3716/15 3720/6 3710/25 3711/3 constructed [1] convince [1] 3700/3
3690/18 3694/22
3721/11 3723/23 computing [1] 3663/16 3741/20 Cook [1] 3660/11
3694/25 3695/21
3724/21 3726/2 conceivable [1] 3697/6 constructing [3] cookies [1] 3735/8
3695/25 3696/8
3726/20 3726/24 concept [4] 3677/19 3722/19 3723/5 cool [1] 3714/16
3696/25 3697/14
commercially [4] 3695/22 3698/17 3742/11 copied [1] 3653/9
3697/19
3687/5 3700/5 3734/8 consumer [13] copies [1] 3673/19
clicked [3] 3678/15
3723/12 3723/12 conception [2] 3718/22 3641/13 3648/9 copy [1] 3700/20
3695/2 3695/24
commercially-oriented 3738/13 3675/3 3675/6 core [2] 3671/5
client [1] 3700/16
[1] 3687/5 concepts [4] 3678/2 3675/14 3675/17 3700/10
close [5] 3648/24
commercially-supporte 3696/21 3697/11 3684/15 3686/6 corrections [1] 3697/10
3662/19 3671/9
d [2] 3723/12 3723/12 3714/9 3699/3 3701/7 correctly [3] 3667/24
3692/18 3727/15
commingled [1] concern [2] 3717/2 3702/21 3733/14 3691/5 3693/22
closed [2] 3665/22
3726/23 3729/3 3734/1 correlated [1] 3709/8
3702/25
communicate [1] concerns [7] 3656/2 consumers [8] 3650/24 cost [4] 3694/6 3700/4
closest [1] 3696/5
3745/4 3656/5 3681/17 3656/6 3684/19 3700/8 3736/14
cloud [3] 3640/5
communicating [1] 3715/23 3723/23 3692/5 3692/22 costs [1] 3676/2
3640/10 3673/20
3657/25 3724/1 3728/24 3721/15 3724/10 Council [1] 3691/15
CNN.com [1] 3677/24
companies [11] 3673/5 concise [2] 3671/24 3733/18 counsel [6] 3637/14
co [3] 3632/4 3671/7
3675/12 3677/9 3719/9 consummated [1] 3637/17 3648/5
3674/22
3717/3 3719/11 conclude [4] 3694/7 3675/10 3665/12 3665/17
co-founder [2] 3671/7
3722/19 3723/4 3701/6 3743/12 CONT [1] 3632/1 3745/25
3674/22
3723/10 3725/24 3743/21 contents [2] 3679/6 counsel's [1] 3662/13
coarsely [1] 3696/18
3728/14 3733/8 conclusion [5] 3674/23 3697/3 counterpoint [1]
code [1] 3722/25
company [10] 3662/24 3711/10 3724/7 contest [1] 3655/7 3678/5
coin [1] 3722/8
3670/2 3670/4 3675/5 3734/22 3741/24 context [18] 3652/13 counterweight [3]
colleagues [2] 3670/1
3675/12 3675/13 conclusions [1] 3724/8 3677/13 3677/14 3717/17 3717/19
3672/15
3712/24 3724/5 concrete [3] 3673/3 3677/21 3686/24 3717/24
collect [4] 3680/11
3727/24 3738/4 3725/8 3737/8 3691/9 3695/7 3696/6 countries [1] 3717/14
3680/11 3680/14
Compaq [1] 3713/2 confess [1] 3718/13 3696/22 3696/23 couple [2] 3702/25
3695/25
compared [4] 3673/13 confidence [2] 3686/9 3697/19 3702/10 3744/10
collecting [1] 3679/18
3684/7 3690/24 3741/11 3705/2 3710/16 course [12] 3646/7
collection [2] 3681/16
3709/7 confident [2] 3634/22 3721/20 3722/10 3646/9 3659/10
3706/10
compete [8] 3640/17 3664/16 3725/14 3741/7 3668/24 3687/23
color [1] 3704/19
3645/7 3645/8 3646/9 confidential [4] continue [7] 3634/22 3690/3 3698/2
Colorado [4] 3631/22
3661/19 3698/4 3637/18 3645/17 3635/18 3645/11 3704/20 3710/19
3632/2 3632/3 3715/2
3699/19 3717/20 3648/5 3657/11 3674/6 3674/25 3734/13 3739/17
COLUMBIA [1] 3631/1
competes [3] 3645/4 confirm [1] 3745/21 3743/22 3745/11 3745/5
column [1] 3703/24
3646/3 3651/23 confirming [1] 3640/23 continued [3] 3633/4 COURT [13] 3631/1
columns [1] 3704/9
competing [2] 3649/10 conflict [2] 3728/24 3634/3 3694/19 3632/23 3632/23
combination [2] 3669/9
3651/19 3730/4 continues [1] 3668/18 3636/25 3656/17
3726/21
COURT... - distribute 3753

C 3721/15 3692/15 3701/12 demand [1] 3655/17 3698/6 3698/6

cut [3] 3691/20 deciding [2] 3689/23 Denver [1] 3632/4 3704/17 3704/22
COURT... [8] 3665/12
3705/25 3717/5 3739/25 Department [5] 3704/24 3705/3
3666/6 3706/4 3717/2
cv [1] 3631/4 decision [1] 3733/25 3631/14 3631/17 3705/4 3705/12
3722/15 3732/25
cycle [1] 3726/15 decisions [3] 3717/6 3631/19 3632/3 3705/13 3705/16
3745/4 3747/3
3718/16 3730/2 3666/15 3706/13 3708/4
Court's [1] 3709/15 D deck [5] 3684/24 depended [2] 3694/16 3718/19 3723/16
Courts [1] 3632/24
DAHLQUIST [1] 3708/11 3708/18 3735/11 3723/25 3727/12
cover [3] 3653/6
3631/19 3709/15 3709/17 dependent [1] 3646/18 3731/3 3733/2 3733/8
3707/1 3707/3
daily [5] 3634/12 deeply [2] 3679/2 depending [1] 3673/10 3733/9 3733/11
coverage [2] 3645/18
3634/17 3634/24 3743/4 deprioritization [1] differentiate [2] 3719/1
3645/20
3636/1 3716/24 DeepMind [3] 3638/4 3716/12 3719/2
coverages [1] 3645/21
Dan [1] 3735/13 3638/6 3638/22 deprioritizes [1] 3716/3 differently [2] 3717/23
covered [1] 3713/15
dangerous [1] 3716/17 default [49] 3636/15 deprioritizing [1] 3728/1
covers [1] 3687/3
DART [3] 3727/23 3654/25 3655/6 3716/5 difficult [6] 3671/15
CPS [1] 3632/3
3728/4 3728/14 3655/9 3655/10 describe [4] 3666/20 3671/15 3673/20
CPS/Antitrust [1]
data [15] 3642/22 3655/19 3656/9 3708/3 3715/23 3691/7 3692/10
3632/3
3647/2 3673/17 3661/20 3664/2 3738/9 3723/11
craft [1] 3697/4
3675/5 3681/15 3664/10 3679/10 described [1] 3735/12 digit [2] 3677/9 3736/8
crawl [3] 3700/13
3696/15 3701/10 3685/21 3686/1 describing [1] 3638/20 digital [4] 3652/13
3701/8 3733/21
3708/3 3708/3 3709/4 3686/15 3686/19 designated [1] 3637/17 3654/3 3654/6 3654/8
crawling [1] 3641/22
3725/19 3726/13 3689/19 3689/22 desktop [17] 3634/21 digression [1] 3693/4
create [18] 3671/5
3726/23 3734/21 3689/25 3690/2 3643/4 3643/8 3669/9 dimensions [3] 3703/4
3671/11 3671/23
3737/3 3690/5 3690/9 3669/9 3680/18 3703/4 3703/7
3672/25 3674/11
datasets [1] 3726/22 3690/11 3690/20 3690/10 3690/14 DINTZER [8] 3631/13
3678/25 3679/8
date [2] 3682/10 3690/22 3691/1 3700/6 3703/25 3633/8 3665/21
3692/20 3692/20
3747/10 3691/2 3691/4 3691/7 3704/7 3704/20 3666/14 3682/19
3693/16 3698/7
dated [1] 3701/23 3691/11 3692/11 3704/23 3705/6 3689/2 3697/25
3699/7 3699/24
dates [1] 3654/20 3692/11 3692/15 3705/14 3716/8 3712/22
3700/11 3700/11
DAVID [1] 3631/19 3694/13 3694/19 3723/6 Direct [4] 3633/8
3719/6 3734/16
day [9] 3631/7 3700/4 3701/5 desktop-oriented [1] 3633/9 3666/12
3736/18
3645/19 3645/19 3708/17 3709/4 3669/9 3714/24
created [3] 3698/16
3668/19 3670/22 3709/9 3711/1 detailed [1] 3703/9 direction [1] 3698/21
3719/4 3720/2
3719/12 3741/9 3711/22 3712/9 details [2] 3698/22 directional [1] 3686/9
creating [13] 3670/3
3743/13 3745/3 3712/23 3713/6 3713/21 directions [1] 3670/11
3673/15 3674/14
day-to-day [1] 3645/19 3713/25 3723/15 detract [1] 3715/18 directive [1] 3680/10
3686/18 3693/17
days [6] 3636/2 3731/16 3732/1 develop [1] 3733/12 directly [5] 3649/25
3696/9 3696/9
3636/13 3676/21 3734/25 developed [1] 3670/25 3676/7 3709/8
3700/16 3717/9
3685/16 3744/22 defaults [11] 3655/8 developers [2] 3693/1 3728/10 3739/8
3718/2 3718/25
3745/6 3659/1 3659/21 3693/12 disadvantages [1]
3720/14 3729/1
DC [5] 3631/5 3631/15 3664/22 3690/4 developing [1] 3683/22 3676/13
creativity [1] 3662/4
3631/18 3632/8 3691/22 3694/7 development [4] disagree [1] 3660/17
credible [1] 3643/7
3632/25 3709/13 3710/15 3638/11 3642/19 Dischler [1] 3744/25
credit [3] 3720/17
de [1] 3712/6 3711/20 3714/10 3647/17 3730/7 discontinuities [1]
3726/25 3727/2
deal [9] 3641/21 defend [2] 3710/2 developments [1] 3698/20
creepy [1] 3674/2
3642/10 3643/13 3724/15 3638/21 discoverable [1]
critical [1] 3640/24
3643/16 3649/3 Defendant [2] 3631/7 device [3] 3658/15 3655/5
criticizing [1] 3738/17
3649/4 3654/19 3632/6 3663/16 3714/7 discuss [4] 3640/9
Cross [4] 3633/4
3661/2 3713/4 defending [1] 3724/22 devices [4] 3655/15 3683/9 3689/8
3633/10 3634/3
dealing [1] 3741/7 define [3] 3678/2 3655/18 3669/5 3743/17
3730/14
deals [5] 3711/21 3681/10 3735/5 3692/23 discussing [2] 3634/7
Cross-Examination [4]
3711/23 3712/23 defined [1] 3670/13 devised [1] 3686/5 3638/14
3633/4 3633/10
3712/24 3713/24 definitely [12] 3675/18 Diane [5] 3707/8 discussion [4] 3708/9
3634/3 3730/14
dealt [2] 3699/23 3678/10 3681/17 3707/9 3707/10 3708/13 3709/12
crossed [1] 3735/10
3699/23 3681/18 3699/6 3707/17 3718/15 3723/8
crossover [1] 3644/16
debated [1] 3659/1 3714/16 3729/3 dieting [2] 3725/6 discussions [3] 3664/1
CTO [2] 3637/1 3637/4
decade [4] 3671/9 3729/14 3731/21 3725/7 3691/21 3692/6
Cue [1] 3744/12
3705/13 3718/17 3732/14 3737/14 difference [3] 3686/4 dislike [1] 3673/4
culminated [1] 3668/23
3727/16 3741/14 3704/16 3704/18 dismay [1] 3686/3
culmination [2]
December [1] 3653/7 definition [3] 3677/18 different [36] 3635/23 display [1] 3668/8
3674/11 3719/6
December 2018 [1] 3678/5 3690/2 3645/7 3649/14 displayed [1] 3641/16
curious [1] 3670/11
3653/7 definitive [1] 3692/7 3651/9 3651/13 dissimilar [1] 3699/8
current [2] 3645/19
decent [1] 3743/10 degree [2] 3645/1 3651/18 3651/18 distinction [1] 3729/15
3689/18
Dechert [1] 3632/11 3681/12 3652/1 3652/15 distinctions [1]
customer [1] 3676/9
decide [2] 3661/5 deliver [1] 3659/22 3663/10 3670/16 3677/23
customers [6] 3671/13
3742/5 Dell [2] 3713/2 3677/20 3686/23 distressing [1] 3722/17
3675/4 3676/7
decided [3] 3669/21 3713/10 3690/17 3696/13 distribute [2] 3689/16
3676/15 3692/5
distribute... - enroll 3754

D dozens [1] 3676/21 3740/18 3742/7 3742/8 3688/1 3689/18

Dr. [2] 3730/16 3742/17 3745/20 eight [2] 3680/24 3689/23 3690/10
distribute... [1] 3691/19
3740/17 3746/3 3746/5 3692/18 3690/24 3691/1
distribution [7] 3654/18
Dr. Ramaswamy [2] e-mail's [1] 3702/24 either [5] 3673/14 3691/8 3692/15
3658/16 3658/20
3730/16 3740/17 earlier [20] 3636/10 3694/9 3731/24 3693/4 3695/14
3690/5 3693/20
draft [2] 3715/14 3641/2 3644/7 3733/13 3745/25 3695/16 3695/17
3694/8 3730/1
3745/20 3654/14 3661/22 either/or [1] 3694/9 3696/1 3696/10
DISTRICT [4] 3631/1
drafted [2] 3746/1 3670/2 3671/4 elect [1] 3711/14 3696/19 3700/12
3631/1 3631/11
3746/7 3674/17 3675/1 election [1] 3687/6 3708/22 3710/5
3632/24
draw [3] 3677/23 3675/3 3675/9 3677/6 element [1] 3698/23 3710/22 3712/8
distrust [1] 3679/7
3715/13 3724/7 3700/25 3717/20 elevated [1] 3635/3 3714/7 3716/21
dived [1] 3708/4
drive [2] 3722/1 3720/2 3723/8 3726/8 eliminate [2] 3696/15 3720/9 3721/11
division [1] 3668/20
3737/4 3732/8 3734/6 3697/21 3723/17 3723/25
Docs [1] 3733/4
driven [7] 3693/10 3735/21 eliminates [1] 3696/17 3725/11 3725/16
document [22]
3698/11 3711/24 early [13] 3663/16 else [3] 3722/10 3725/21 3725/23
3637/18 3640/21
3713/22 3717/7 3668/17 3669/13 3745/13 3746/14 3726/2 3728/5 3733/6
3648/4 3648/8
3718/6 3721/6 3670/5 3671/21 elsewhere [1] 3716/20 3734/18 3735/19
3648/12 3652/22
driver [1] 3647/12 3672/5 3672/8 embarrassment [2] 3737/12 3737/16
3656/15 3657/11
driving [1] 3720/14 3676/24 3677/2 3660/14 3660/21 3737/23 3738/1
3682/6 3682/15
Dropbox [1] 3733/3 3686/17 3692/5 embedded [1] 3682/15 3738/3 3738/10
3682/22 3688/6
dropdown [1] 3655/1 3705/7 3733/17 embodiment [1] 3738/11 3738/12
3689/1 3702/7
droves [1] 3659/15 easier [8] 3678/2 3683/24 3740/6
3705/17 3705/18
dry [1] 3717/5 3683/16 3683/17 emergence [2] engineer [2] 3667/7
3708/13 3715/7
DuckDuckGo [8] 3690/23 3691/1 3648/18 3648/18 3667/12
3719/17 3730/22
3691/5 3737/20 3694/2 3696/18 employed [1] 3667/3 engineering [2]
3740/17 3740/24
3737/22 3738/7 3700/24 employees [1] 3699/25 3702/18 3743/4
documents [8] 3654/2
3738/15 3738/17 easily [3] 3655/5 enables [1] 3697/5 engineers [4] 3692/18
3664/3 3681/25
3738/22 3739/10 3698/18 3726/10 encompassed [1] 3700/14 3707/11
3689/4 3697/24
DuckDuckGo's [2] easy [11] 3655/7 3668/7 3721/13
3706/10 3706/22
3738/13 3739/6 3686/18 3689/21 encourage [1] 3734/9 engines [45] 3645/9
3730/19
dug [1] 3640/19 3693/11 3710/1 encyclopedias [1] 3646/4 3646/4 3646/9
DOJ [3] 3631/13
during [6] 3655/24 3711/10 3724/18 3684/7 3646/14 3646/18
3665/10 3746/15
3668/16 3669/7 3725/7 3725/8 3730/8 end [12] 3641/13 3649/2 3649/7
dollar [4] 3644/13
3689/9 3726/15 3735/8 3664/9 3674/1 3674/2 3649/16 3649/17
3685/9 3719/11
3734/13 Economically [1] 3692/6 3717/6 3717/8 3649/18 3650/4
3742/14
DX [1] 3648/7 3648/24 3718/2 3733/9 3650/5 3650/7
dollars [8] 3644/18
DX501 [4] 3633/18 economics [2] 3640/14 3744/25 3745/7 3650/25 3651/3
3644/21 3644/23
3652/16 3652/20 3646/6 3745/10 3652/5 3661/13
3664/6 3677/6 3700/8
3653/4 economies [1] 3698/25 ended [2] 3676/15 3661/18 3661/19
3713/18 3726/21
DX502 [1] 3656/14 economist [3] 3684/22 3692/24 3670/19 3671/14
domain [2] 3648/16
DX524 [3] 3633/16 3735/14 3737/11 ending [1] 3730/25 3673/23 3678/11
3650/11
3637/12 3637/15 Ecosia [1] 3691/5 engage [1] 3655/18 3681/18 3685/17
domains [3] 3646/7
DX673 [4] 3633/17 education [1] 3666/20 engaged [2] 3638/22 3688/20 3690/20
3646/8 3649/14
3647/13 3647/24 effect [2] 3735/4 3639/9 3690/23 3691/22
dominant [7] 3663/18
3648/2 3742/24 engaging [1] 3659/4 3691/25 3695/9
3663/24 3712/15
DX680 [5] 3633/15 effective [2] 3678/13 engine [89] 3636/15 3695/19 3695/21
3735/1 3735/23
3636/19 3636/21 3696/9 3645/4 3646/11 3696/23 3697/5
3741/8 3741/19
3637/7 3637/10 effectively [9] 3668/10 3646/15 3653/11 3697/7 3697/17
Dominic [1] 3702/9
dynamic [2] 3642/22 3691/15 3692/3 3654/25 3655/6 3705/9 3716/24
done [7] 3652/23
3700/6 3694/14 3711/7 3661/12 3667/9 3717/14 3719/2
3680/19 3688/20
3717/16 3718/3 3670/3 3670/7 3723/12 3728/10
3702/5 3721/16 E 3722/19 3735/24 3670/12 3670/13 3731/3
3722/24 3727/24
e-commerce [2] effects [1] 3715/19 3670/17 3670/25 engrained [1] 3689/22
DoubleClick [10]
3651/4 3651/10 efficacy [1] 3673/25 3671/5 3671/7 enjoyed [2] 3642/13
3668/24 3727/18
e-mail [31] 3636/21 efficient [3] 3689/16 3671/14 3671/18 3642/17
3727/21 3727/22
3636/22 3637/22 3689/20 3712/1 3672/3 3672/4 enormous [3] 3685/9
3727/24 3727/25
3637/25 3638/3 effort [1] 3740/1 3672/22 3672/23 3711/7 3712/9
3728/2 3728/15
3638/19 3640/8 efforts [6] 3638/8 3673/1 3673/3 3673/6 enormously [1]
3729/6 3729/19
3640/12 3647/14 3639/8 3642/19 3675/4 3675/6 3675/7 3697/19
down [8] 3638/18
3647/18 3653/6 3718/25 3719/2 3675/10 3675/14 enough [13] 3634/20
3675/3 3675/9
3673/22 3685/7 3719/6 3675/18 3676/6 3671/17 3674/24
3675/15 3686/17
3700/16 3701/21 egregious [8] 3740/10 3676/17 3676/19 3683/4 3684/5 3698/8
3692/16 3696/18
3702/11 3702/13 3740/15 3741/6 3678/13 3680/13 3699/24 3699/25
3716/3
3702/22 3703/15 3741/11 3741/17 3681/7 3685/10 3701/1 3701/2
download [1] 3635/17
3707/6 3707/20 3742/2 3742/3 3686/2 3686/5 3712/11 3721/13
downloads [4] 3635/9
3707/23 3708/7 3742/16 3686/17 3686/24 3738/6
3635/14 3635/21
3710/8 3714/13 egregiousness [1] 3687/11 3687/17 enroll [1] 3719/14
3636/13
ensure - final 3755

E 3741/17 3742/22 execs [1] 3692/8 exploit [1] 3699/14 3698/3

evening [1] 3745/14 executives [2] 3653/10 exploratory [1] 3708/5 fairness [1] 3729/4
ensure [2] 3726/3
eventually [1] 3668/11 3656/2 Explorer [2] 3711/4 fall [1] 3640/24
3726/11
everybody [9] 3679/2 exercise [1] 3636/9 3713/10 familiar [16] 3645/25
enter [3] 3642/14
3689/7 3690/15 exert [1] 3676/10 exploring [1] 3669/25 3672/14 3677/13
3712/18 3712/18
3713/3 3713/23 exhibit [16] 3633/14 exposing [1] 3736/2 3686/20 3688/22
enterprise [10] 3675/5
3716/20 3725/3 3633/15 3633/16 express [1] 3681/12 3691/17 3694/22
3675/13 3732/21
3725/6 3746/19 3633/17 3633/18 expressed [3] 3681/20 3695/4 3714/3 3715/7
3732/24 3733/1
everywhere [2] 3678/4 3633/19 3633/20 3718/6 3718/9 3722/12 3723/16
3733/4 3733/6
3722/24 3637/10 3637/15 expressing [1] 3639/7 3725/16 3727/11
3733/13 3733/20
evidence [25] 3633/15 3648/2 3653/4 3682/1 extended [1] 3729/17 3737/20 3739/17
3734/1
3633/16 3633/17 3683/11 3710/7 extension [5] 3686/18 famously [1] 3718/20
entire [3] 3640/12
3633/18 3633/19 3745/16 3746/11 3690/11 3690/16 far [2] 3691/21 3719/5
3664/4 3722/23
3633/20 3637/7 Exhibit 715 [1] 3700/3 3723/6 fast [4] 3674/24
entitled [1] 3747/5
3637/10 3637/12 3745/16 extensions [1] 3700/6 3699/24 3701/1
entitlement [1] 3694/13
3637/15 3647/23 Exhibit UPX93 [1] extent [2] 3690/1 3701/2
entity [1] 3678/8
3648/2 3653/4 3710/7 3694/16 faster [3] 3720/21
entry [2] 3647/1
3683/11 3705/18 exist [3] 3663/3 extract [1] 3688/14 3721/2 3721/4
3655/22
3705/19 3705/23 3697/10 3721/6 FastTap [8] 3693/13
environment [1]
3706/2 3706/16 existed [1] 3699/10 F 3719/22 3719/25
3674/21
3706/23 3709/2 existence [6] 3671/12 face [2] 3699/20 3720/5 3720/10
environments [1]
3715/6 3719/19 3676/17 3694/21 3716/24 3720/15 3720/18
3718/3
3745/17 3746/11 3716/2 3734/13 Facebook [6] 3647/8 3721/8
envisioned [1] 3672/22
evolution [1] 3712/13 3736/6 3651/8 3651/24 fears [1] 3679/3
equal [2] 3673/12
evolved [2] 3663/14 existing [1] 3705/9 3652/10 3654/3 feasible [1] 3673/7
3737/12
3727/18 exists [3] 3675/6 3654/10 feature [8] 3661/25
equally [2] 3672/9
EVP [1] 3742/12 3675/18 3698/18 faced [1] 3705/20 3720/16 3721/13
3724/11
exacerbated [2] expect [1] 3665/23 faces [2] 3638/15 3721/15 3721/18
equilibrium [1] 3736/16
3717/13 3717/15 expectation [4] 3653/20 3724/3 3733/18
ergo [1] 3742/2
exact [4] 3644/16 3665/22 3677/22 facility [1] 3733/2 3737/9
Erik [1] 3647/18
3647/7 3654/20 3678/7 3734/18 facing [1] 3707/18 features [9] 3662/4
escape [1] 3699/5
3741/25 expectations [3] fact [19] 3643/7 3673/17 3674/1
especially [10]
exactly [2] 3645/20 3701/3 3717/4 3721/1 3646/24 3647/2 3679/15 3687/23
3678/14 3693/25
3706/7 experience [27] 3650/5 3658/19 3708/25 3730/5
3700/22 3701/3
examination [12] 3641/12 3671/11 3658/24 3679/4 3730/5 3737/14
3717/6 3718/18
3633/4 3633/5 3633/8 3671/23 3672/25 3679/10 3681/19 feedback [3] 3685/25
3720/6 3721/10
3633/9 3633/10 3674/12 3678/25 3683/4 3684/3 3692/7 3737/14
3734/14 3743/3
3634/3 3661/9 3688/18 3693/1 3695/12 3703/2 feel [3] 3634/22
essentially [10] 3641/9
3662/13 3666/12 3693/13 3700/11 3703/2 3703/9 3645/20 3691/18
3667/21 3705/11
3714/24 3730/13 3700/23 3700/24 3717/13 3729/10 feeling [2] 3687/18
3720/5 3720/7
3730/14 3709/1 3715/18 3735/15 3739/23 3719/3
3720/10 3723/4
examining [1] 3745/16 3719/4 3719/6 3719/8 facto [1] 3712/6 felt [4] 3659/11
3729/16 3739/25
example [23] 3646/4 3719/16 3720/4 factor [3] 3663/15 3671/10 3679/16
3740/5
3651/18 3653/23 3720/14 3725/1 3709/3 3735/23 3739/24
establish [1] 3706/6
3673/3 3676/2 3728/11 3733/14 factors [5] 3689/23 few [8] 3634/8 3676/21
established [1] 3705/7
3680/15 3680/16 3737/7 3738/14 3705/15 3708/23 3693/5 3697/23
et [2] 3631/3 3682/16
3681/11 3687/7 3739/7 3741/4 3735/1 3739/25 3715/3 3740/12
etc [1] 3709/1
3688/5 3688/15 experiences [11] facts [1] 3636/3 3743/7 3744/22
Europe [1] 3703/8
3688/16 3690/8 3679/5 3698/20 fail [1] 3741/11 field [3] 3673/13
evaluation [2] 3652/23
3690/10 3690/18 3717/9 3718/1 3718/7 failed [2] 3637/11 3673/14 3695/3
3664/5
3693/25 3697/4 3718/10 3718/12 3712/10 fifth [2] 3631/17
even [36] 3634/21
3698/13 3711/22 3718/23 3719/13 fails [6] 3740/10 3639/2
3643/7 3646/19
3721/23 3726/6 3741/8 3741/17 3740/15 3741/6 Figma [1] 3736/25
3647/8 3648/24
3731/7 3735/5 expertise [1] 3701/10 3742/2 3742/3 figure [9] 3683/7
3655/17 3659/1
examples [2] 3688/11 explain [5] 3684/21 3742/16 3696/4 3696/6
3662/19 3664/10
3737/8 3717/2 3721/7 fair [16] 3635/3 3696/16 3696/25
3664/22 3671/9
excellence [2] 3741/9 3722/15 3732/24 3637/23 3638/14 3697/11 3698/22
3673/10 3673/13
3741/10 explaining [2] 3639/20 3638/19 3639/9 3700/20 3745/3
3675/10 3676/12
Except [1] 3711/12 3742/11 3639/15 3641/8 figured [1] 3698/9
3685/8 3686/7
exceptionally [1] explains [2] 3704/16 3642/8 3655/13 figures [2] 3700/21
3692/10 3693/15
3724/19 3704/17 3662/16 3669/10 3704/17
3693/24 3700/4
excerpt [1] 3684/22 explanation [2] 3681/14 3683/4 figuring [4] 3696/2
3700/14 3701/5
exchange [5] 3636/21 3664/11 3699/18 3692/17 3706/15 3697/18 3698/23
3701/9 3711/2
3636/22 3637/22 explicit [2] 3679/22 3738/6 3707/13
3719/12 3723/13
3703/7 3740/18 3680/5 fairly [6] 3653/17 file [1] 3744/16
3724/19 3725/5
exchanged [1] 3744/14 explicitly [2] 3694/14 3686/13 3686/14 final [4] 3674/5
3725/14 3725/22
exec [1] 3691/23 3731/21 3690/18 3697/13 3698/24 3718/13
3726/5 3736/9 3737/3
final... - history 3756

F forget [3] 3647/7 funny [1] 3687/1 GM [1] 3668/1 half [3] 3636/12
3654/20 3724/16 further [10] 3640/25 GM-like [1] 3668/1 3708/8 3720/11
final... [1] 3744/16
form [1] 3663/15 3665/1 3676/25 goes [4] 3639/19 hall [1] 3666/14
finally [2] 3669/4
forma [1] 3727/9 3689/1 3699/13 3640/8 3651/20 hallucinations [3]
3691/3
formidable [1] 3695/18 3705/17 3712/4 3657/19 3661/23 3662/3
financial [1] 3640/9
forms [1] 3646/10 3714/21 3717/12 good [18] 3646/8 3662/4
financials [2] 3702/2
forth [2] 3652/10 3730/12 3650/5 3681/21 hand [7] 3652/19
3703/1
3705/10 futility [1] 3719/4 3699/7 3700/14 3666/4 3688/15
find [1] 3673/20
forward [2] 3680/12 future [1] 3740/9 3715/1 3715/18 3692/1 3694/3
finding [1] 3733/19
3743/22 3723/24 3728/13 3724/18 3744/3
fine [2] 3664/18 G
found [2] 3635/24 3729/5 3729/10 handed [1] 3719/18
3717/7
3686/3 gain [1] 3646/7 3729/13 3729/19 handle [1] 3697/10
fingertips [1] 3684/5
foundation [1] 3705/23 gaining [2] 3634/19 3729/21 3730/16 hands [2] 3689/20
finish [2] 3652/7
founder [3] 3671/7 3646/6 3731/17 3741/9 3692/22
3656/17
3674/22 3699/23 game [3] 3635/23 3742/8 hang [2] 3659/17
finished [1] 3740/23
four [4] 3667/18 3674/7 3712/19 GOOGLE [140] 3659/17
Firefox [3] 3690/9
3684/14 3699/21 Gates [1] 3636/23 Google's [17] 3638/7 happen [1] 3743/2
3713/3 3713/21
3700/10 gave [4] 3680/9 3638/21 3667/16 happened [4] 3660/9
firm [3] 3644/22
fourth [2] 3639/2 3719/8 3720/5 3668/19 3668/22 3699/8 3720/4
3672/10 3672/12
3703/24 3720/17 3669/8 3703/3 3704/7 3721/25
first [26] 3638/1
fraction [3] 3681/19 gears [1] 3689/15 3704/13 3705/12 happens [5] 3651/6
3644/11 3667/6
3686/13 3686/15 general [30] 3645/4 3709/7 3709/13 3677/25 3678/12
3667/15 3668/11
fragmented [1] 3645/6 3645/8 3727/11 3730/13 3741/10 3741/13
3670/14 3671/12
3646/25 3645/12 3646/3 3731/15 3739/22 happy [5] 3657/2
3672/8 3675/10
framed [1] 3685/4 3646/11 3646/15 3742/19 3658/11 3676/23
3680/19 3685/19
framework [1] 3711/6 3647/11 3649/18 Google.com [8] 3696/13 3745/5
3696/19 3696/23
free [28] 3650/8 3650/25 3651/3 3677/24 3707/14 hard [5] 3673/10
3697/8 3702/4
3657/19 3658/4 3651/6 3651/11 3711/8 3711/14 3681/15 3701/6
3703/24 3706/23
3658/14 3671/14 3652/5 3652/6 3711/14 3711/21 3730/9 3731/19
3706/25 3707/22
3672/2 3674/3 3661/12 3661/16 3731/18 3731/22 head [1] 3702/16
3710/20 3715/13
3676/10 3676/12 3670/6 3670/12 gradually [1] 3667/17 headaches [1] 3679/3
3716/1 3716/1 3721/5
3677/2 3679/1 3670/17 3670/19 great [10] 3643/6 headed [1] 3745/5
3729/9 3746/5
3679/13 3684/19 3670/25 3676/1 3643/7 3658/22 heading [5] 3683/18
firsthand [1] 3664/18
3685/1 3685/3 3685/6 3678/6 3683/8 3664/13 3669/25 3684/16 3685/15
five [4] 3669/14 3691/6
3685/8 3688/1 3719/1 3689/17 3708/13 3683/18 3720/14 3703/25 3721/5
3742/18 3743/10
3725/11 3734/9 3710/5 3716/17 3726/3 3726/11 heard [1] 3685/25
flexibility [1] 3659/6
3734/20 3735/16 3719/2 3728/25 hearsay [1] 3683/5
Floor [2] 3632/4
3735/23 3735/24 generally [9] 3670/9 grew [3] 3666/21 heart [1] 3667/21
3632/12
3736/3 3736/10 3673/22 3676/18 3667/14 3729/14 heavily [1] 3646/3
flows [2] 3678/19
3737/3 3678/23 3686/9 Greylock [4] 3669/17 heed [1] 3743/16
3678/19
freedom [1] 3676/8 3687/8 3697/7 3672/9 3672/11 help [6] 3648/6
focus [13] 3637/25
freely [1] 3690/8 3740/15 3741/17 3732/9 3656/17 3696/25
3640/4 3645/10
frequently [1] 3655/24 generation [1] 3639/11 ground [1] 3670/5 3697/8 3699/17
3672/22 3673/1
Friday [3] 3745/9 generative [4] 3671/22 group [2] 3653/10 3744/4
3674/3 3676/24
3745/10 3745/11 3674/6 3719/8 3708/1 helped [1] 3716/13
3677/2 3678/9
front [6] 3659/8 3719/13 groups [2] 3638/7 helpful [5] 3688/4
3678/10 3688/24
3681/24 3688/8 gentleman [2] 3672/14 3638/11 3697/14 3710/16
3721/15 3734/1
3730/17 3730/25 3740/18 grow [3] 3701/1 3701/2 3738/18 3745/6
focused [5] 3673/14
3740/17 genuinely [1] 3700/23 3710/1 Herculean [1] 3700/16
3676/9 3687/25
fulfilling [1] 3724/10 geographical [1] growing [2] 3663/17 hey [3] 3659/24
3688/1 3688/12
full [5] 3648/5 3669/21 3703/7 3677/7 3679/9 3679/23
focusing [1] 3686/18
3693/12 3706/3 geography [1] 3705/2 grown [2] 3634/12 hide [2] 3659/9
folks [5] 3648/21
3736/3 Germany [1] 3694/1 3654/7 3659/12
3669/7 3691/24
full-time [1] 3669/21 gets [2] 3689/24 growth [4] 3651/6 Higgins [1] 3744/24
3702/10 3746/14
function [1] 3733/14 3712/7 3674/20 3674/25 high [7] 3637/22
follow [2] 3690/18
fundamental [4] Giannandrea [1] 3717/4 3638/20 3642/10
3710/17
3640/12 3650/9 3744/12 guarantees [1] 3726/5 3663/23 3664/22
following [3] 3635/8
3708/20 3736/4 giant [1] 3687/19 guess [4] 3670/12 3692/23 3734/17
3685/16 3723/22
fundamentally [2] given [6] 3641/2 3682/11 3682/11 higher [2] 3696/9
fool [1] 3659/13
3634/21 3646/18 3667/25 3678/8 3724/22 3743/6
For Plaintiffs [1]
funded [3] 3638/12 3678/8 3697/18 guys [1] 3708/7 highly [4] 3638/25
3631/22
3674/18 3724/18 3697/19 3639/1 3645/17
foray [1] 3668/22
funding [6] 3672/6 gives [3] 3691/8 H 3657/11
force [1] 3694/4
3672/7 3672/8 3696/2 3712/13 habit [2] 3689/25 histories [1] 3705/12
forced [2] 3692/20
3672/11 3674/21 giving [1] 3688/12 3699/22 history [10] 3679/10
3724/15
3676/9 global [2] 3702/16 habits [1] 3690/4 3679/18 3679/23
foregoing [1] 3747/4
funds [1] 3732/13 3703/1 habitual [1] 3690/13 3680/11 3680/12
foremost [1] 3729/9
history... - JR 3757

H 3711/25 index [3] 3641/11 3690/10 3694/10 3706/1 3706/16

hunch [1] 3698/5 3700/13 3733/22 3700/3 3708/4 3711/2
history... [5] 3680/14
hundred [10] 3634/14 India [2] 3666/21 installing [1] 3690/16 3711/15 3712/5
3684/9 3697/14
3634/16 3634/16 3666/23 installs [2] 3634/24 3716/8 3719/14
3698/14 3705/4
3651/1 3658/16 indicate [4] 3685/8 3694/3 3720/7 3721/20
hit [1] 3720/20
3658/19 3659/2 3731/14 3734/24 instance [1] 3687/21 3723/6 3731/18
hold [4] 3643/5
3659/20 3659/21 3738/11 instead [4] 3659/14 3731/22 3735/25
3658/25 3659/1
3722/19 indicated [1] 3635/5 3671/24 3720/10 3745/17 3746/11
3664/3
hypothetical [2] indicates [1] 3740/18 3745/8 intricate [1] 3690/19
home [28] 3635/15
3742/21 3742/22 indirectly [1] 3739/8 institutions [1] 3732/9 introduce [1] 3665/12
3665/6 3707/23
industry [1] 3663/10 instruction [1] 3743/17 introduction [5]
3708/17 3709/4 I infamous [1] 3741/22 instruments [1] 3694/1 3634/10 3634/18
3709/5 3709/9
IBM [1] 3713/2 infer [1] 3680/4 intelligence [1] 3634/25 3662/9
3710/10 3710/11
ID [1] 3734/16 inferior [2] 3660/15 3696/12 3715/10
3710/15 3711/8
IDE [1] 3666/22 3742/19 intent [10] 3686/20 intuitions [1] 3698/12
3711/22 3711/24
idea [10] 3642/2 influence [1] 3676/10 3686/22 3686/24 invented [1] 3684/1
3712/2 3712/5
3658/22 3659/7 influenced [1] 3641/17 3687/1 3687/2 3687/2 invest [1] 3645/11
3712/14 3712/15
3674/13 3688/11 influences [1] 3708/22 3687/3 3687/8 invested [2] 3636/5
3712/23 3713/7
3720/24 3721/19 information [24] 3687/10 3687/14 3732/15
3713/11 3713/14
3721/21 3723/7 3670/9 3670/18 interact [2] 3704/23 investing [2] 3641/24
3713/19 3713/24
3733/5 3670/22 3678/19 3704/24 3669/20
3714/5 3731/4 3731/9
ideas [1] 3708/6 3681/3 3688/3 interaction [4] 3677/22 investment [5] 3637/23
3731/16 3732/4
identical [1] 3706/10 3688/13 3688/14 3678/6 3678/8 3639/21 3640/3
honestly [1] 3697/16
identified [2] 3706/6 3694/23 3694/25 3680/13 3662/24 3692/19
Honor [46] 3634/5
3730/22 3695/21 3695/25 interest [6] 3662/7 investments [2]
3637/6 3637/11
identify [1] 3680/25 3696/25 3697/20 3674/21 3722/3 3662/13 3663/4
3647/23 3652/20
IL [1] 3631/20 3716/3 3716/5 3722/3 3722/6 3728/24 investor [1] 3636/4
3656/16 3661/6
impact [3] 3680/15 3725/23 3725/25 3729/25 investors [2] 3684/25
3661/11 3665/1
3738/14 3738/23 3726/1 3726/10 interested [4] 3645/12 3694/20
3665/3 3665/10
implications [1] 3648/9 3727/2 3727/4 3651/19 3678/17 involved [2] 3653/11
3665/24 3682/13
implicit [2] 3679/22 3737/13 3688/18 3671/6
3682/14 3682/21
3680/2 informed [3] 3718/11 interesting [3] 3635/15 iOS [8] 3636/10
3682/25 3683/13
importance [3] 3718/22 3737/3 3639/5 3658/10 3692/15 3692/18
3689/12 3689/15
3696/15 3709/13 informing [1] 3687/16 interests [4] 3693/10 3692/24 3693/24
3701/15 3705/17
3714/10 infrastructure [9] 3724/21 3729/2 3694/10 3705/11
3706/3 3706/7 3706/9
important [4] 3639/21 3639/19 3642/3 3729/9 3720/3
3706/15 3706/18
3678/12 3710/21 3667/8 3667/13 interface [1] 3673/22 IP [4] 3680/16 3680/24
3706/23 3714/18
3741/15 3667/16 3668/14 interfaces [1] 3667/18 3726/9 3726/10
3714/21 3715/6
imposes [1] 3673/2 3671/17 3672/5 internal [1] 3729/5 iPads [1] 3655/11
3719/19 3730/12
impossible [1] 3692/14 3707/18 internet [6] 3684/18 iPhones [1] 3655/11
3738/6 3738/24
improve [2] 3662/5 inherent [1] 3640/16 3711/4 3712/16 ish [2] 3671/12
3743/24 3744/8
3696/12 initial [9] 3635/1 3713/9 3721/25 3676/21
3744/14 3745/2
improved [1] 3737/6 3638/20 3639/7 3722/17 Island [1] 3666/24
3745/9 3745/10
inability [3] 3699/18 3639/13 3639/13 interrupt [5] 3656/16 isolation [1] 3717/7
3745/15 3745/17
3701/4 3724/2 3662/7 3676/8 3675/20 3689/2 issue [1] 3646/22
3746/13 3746/15
Inaudible [1] 3656/19 3734/18 3734/19 3697/25 3723/19 issued [1] 3660/11
3746/17 3746/18
inbox [1] 3700/18 initially [5] 3635/2 interruption [1] 3666/3 iterations [1] 3737/18
HONORABLE [1]
incentive [1] 3720/8 3635/7 3636/22 intervals [1] 3686/10
3631/10
incident [2] 3660/9 3729/7 3734/14 intimately [1] 3691/17 J
HOOK [3] 3632/23
3660/12 innovate [3] 3674/4 into [51] 3633/15 jargon [1] 3720/1
3747/3 3747/10
include [1] 3703/5 3688/1 3739/24 3633/16 3633/17 Jed [2] 3647/14
hopefully [4] 3648/6
includes [1] 3707/6 innovation [3] 3640/15 3633/18 3633/19 3647/15
3674/12 3691/8
including [8] 3653/10 3660/25 3721/6 3633/20 3637/6 JEFF [3] 3632/23
3712/13
3695/17 3708/24 innovations [1] 3693/2 3637/10 3637/12 3747/3 3747/10
hopes [3] 3646/25
3718/22 3720/4 Inovia [1] 3672/12 3637/15 3640/19 job [11] 3646/8
3648/14 3648/15
3735/3 3741/6 inquiries [1] 3714/17 3642/14 3643/2 3649/24 3650/5
hour [1] 3708/9
3741/22 inroads [1] 3643/2 3644/25 3647/23 3667/6 3668/4 3668/7
household [1] 3726/11
incorporated [1] inside [3] 3651/17 3648/2 3653/4 3671/15 3687/24
housekeeping [1]
3670/4 3661/14 3661/15 3659/13 3663/17 3695/16 3718/15
3744/10
incorporates [1] insight [2] 3671/11 3668/1 3668/22 3729/13
HP [2] 3709/23 3710/9
3634/11 3708/6 3675/17 3683/11 JOHN [1] 3632/6
huge [1] 3660/14
increase [1] 3662/9 insights [4] 3648/9 3689/20 3689/22 joined [2] 3667/7
huh [6] 3703/21
increased [1] 3718/14 3648/10 3696/2 3692/21 3693/3 3667/12
3719/21 3731/6
incredibly [1] 3691/7 3698/7 3693/4 3695/14 JONATHAN [1] 3632/2
3731/10 3731/13
independent [3] Instagram [3] 3651/10 3696/2 3696/14 Jorgensen [2] 3647/18
3740/25
3659/25 3702/19 3652/11 3654/3 3696/19 3699/22 3647/20
humanity [1] 3684/4
3712/8 install [4] 3635/21 3705/19 3705/23 JR [1] 3631/23
humans [2] 3695/11
Juda - maps 3758

J last [12] 3634/23 limitation [1] 3682/18 3699/11 3706/21 3710/8 3714/13
3640/18 3650/10 limited [1] 3677/23 3708/11 3716/11 3740/18 3742/7
Juda [2] 3744/6
3650/17 3672/1 line [2] 3639/2 3648/15 3722/9 3735/18 3742/17 3745/20
3744/8
3680/24 3685/16 lines [3] 3669/10 3742/6 3744/18 3746/3 3746/5
JUDGE [1] 3631/11
3700/7 3733/8 3736/5 3685/4 3737/2 looked [5] 3646/21 mail's [1] 3702/24
judgment [1] 3663/3
3742/6 3744/10 LinkedIn [1] 3646/20 3687/15 3718/15 mails [1] 3700/18
July [2] 3701/23
late [6] 3669/13 links [6] 3671/24 3722/22 3734/21 main [3] 3658/24
3702/25
3670/4 3734/3 3740/2 3674/8 3674/9 3693/7 looking [14] 3670/14 3663/16 3728/9
July 2018 [1] 3701/23
3740/3 3740/4 3697/6 3722/2 3686/10 3687/22 Maine [1] 3632/8
jumps [1] 3687/19
latency [1] 3721/1 list [9] 3679/11 3688/5 3688/12 mainstream [1]
June [1] 3734/6
later [1] 3640/8 3691/22 3691/25 3688/17 3696/16 3719/12
Justice [4] 3631/14
launch [2] 3719/8 3693/5 3695/11 3708/12 3715/3 maintain [1] 3739/11
3631/17 3631/19
3721/15 3703/15 3706/11 3718/19 3721/20 maintaining [1]
3666/15
launched [10] 3635/7 3706/12 3706/16 3721/22 3744/5 3693/18
justified [1] 3729/11
3651/14 3668/21 literally [2] 3679/5 3745/2 major [4] 3647/10
justifying [1] 3639/20
3670/2 3674/17 3710/23 looks [1] 3652/23 3694/12 3701/5
Justin [1] 3647/14
3693/13 3721/18 little [15] 3645/3 loose [1] 3687/6 3739/25
K 3723/3 3734/5 3645/19 3646/25 lose [1] 3742/23 majority [6] 3668/18
keep [4] 3653/20 3739/14 3648/23 3659/18 lot [23] 3648/20 3670/18 3676/15
3656/12 3705/5 launches [1] 3718/14 3670/21 3675/13 3673/14 3673/24 3681/17 3731/24
3711/6 law [3] 3632/3 3644/22 3681/10 3687/6 3678/19 3685/9 3732/4
KENNETH [3] 3631/13 3645/1 3688/16 3702/10 3688/2 3688/18 makers [1] 3694/3
3632/7 3666/14 lead [4] 3658/21 3702/12 3704/19 3694/16 3697/12 makes [7] 3638/3
kept [5] 3659/4 3672/3 3668/6 3668/11 3722/3 3728/18 3697/13 3698/21 3662/24 3671/19
3672/18 3733/17 3672/16 live [2] 3681/13 3699/4 3704/25 3691/1 3694/2
3733/21 leadership [1] 3667/15 3724/17 3709/16 3711/21 3696/17 3718/1
Kevin [2] 3636/23 leads [2] 3678/23 lives [3] 3673/18 3711/22 3712/2 making [12] 3634/23
3636/25 3722/21 3724/14 3724/16 3712/7 3724/15 3639/10 3644/10
key [3] 3676/19 learn [1] 3670/12 living [1] 3669/21 3729/14 3733/24 3658/24 3659/3
3707/11 3709/20 least [13] 3635/2 LLC [1] 3631/6 3736/4 3736/15 3692/16 3692/24
kind [3] 3663/21 3637/20 3639/7 LLP [3] 3631/23 3717/6 3723/14
lots [7] 3635/17 3663/4
3673/15 3723/11 3640/12 3641/17 3632/7 3632/11 3670/23 3689/24 3726/15 3739/9
kinds [4] 3659/6 3647/1 3647/10 load [1] 3718/14 3694/3 3705/8 3729/43742/6
3674/13 3688/22 3650/3 3657/15 local [11] 3647/21 manage [1] 3700/19
loud [3] 3702/5 3704/5
3718/15 3658/1 3681/10 3649/19 3649/20 3704/13 managed [1] 3636/1
knew [1] 3639/23 3726/10 3740/11 3650/21 3680/15 louder [1] 3702/12 management [2]
knowing [2] 3678/4 leave [2] 3669/12 3680/23 3709/1 lousy [1] 3649/23 3663/3 3668/14
3678/17 3669/24 3725/25 3726/7 love [2] 3647/2 manager [1] 3725/10
knowledge [1] 3664/19 leaving [1] 3742/13 3727/2 3727/7 3664/24 Manber [1] 3672/14
known [2] 3697/9 led [2] 3648/18 located [1] 3681/2 loved [2] 3669/22 maniacal [1] 3674/3
3713/16 3686/17 location [9] 3680/11 3700/23 manifest [1] 3678/21
knows [3] 3678/13 left [5] 3669/13 3680/14 3680/16 low [3] 3648/25 manners [1] 3679/3
3679/2 3739/9 3669/19 3704/2 3680/17 3680/19 3655/13 3663/25 manufacturer [2]
Kolotouros [1] 3744/24 3731/8 3743/10 3680/20 3680/21 Lowcock [1] 3744/6 3711/23 3713/17
legal [1] 3673/13 3680/25 3681/2 lower [1] 3644/5 manufacturers [1]
L leisurely [1] 3687/21 locations [1] 3688/17 lowered [1] 3736/14 3713/13
labeled [1] 3641/14 length [1] 3640/9 log [3] 3680/21 3681/4 lunch [1] 3634/6 many [35] 3635/9
labeling [1] 3641/9 less [2] 3676/24 3711/2 3646/12 3649/15
3677/6 logged [1] 3683/4 M 3650/7 3662/22
lack [5] 3678/3
3681/18 3717/25 letter [3] 3660/11 logs [1] 3709/5 machines [1] 3713/23 3671/25 3673/2
3717/25 3724/9 3707/1 3707/3 long [4] 3647/7 macro [1] 3674/20 3674/10 3676/3
laid [1] 3705/22 letters [1] 3693/5 3668/20 3686/25 Macs [1] 3655/10 3677/20 3677/20
language [2] 3657/15 level [9] 3637/22 3735/7 Madras [1] 3666/22 3681/16 3684/4
3697/8 3638/20 3642/10 long-studied [1] magical [1] 3684/7 3684/18 3686/3
large [11] 3669/5 3679/24 3679/25 3686/25 magnitude [1] 3721/2 3690/4 3690/21
3673/5 3686/13 3742/8 3742/8 longer [2] 3675/6 mail [31] 3636/21 3692/2 3693/25
3686/14 3694/17 3742/18 3742/22 3675/18 3636/22 3637/22 3696/13 3700/3
3695/3 3697/8 libraries [1] 3684/6 look [27] 3640/13 3637/25 3638/3 3700/7 3703/10
3698/12 3713/12 life [3] 3716/14 3644/3 3647/13 3638/19 3640/8 3712/15 3713/16
3715/19 3716/13 3726/15 3726/15 3648/7 3650/10 3640/12 3647/14 3715/19 3718/14
larger [1] 3733/23 light [1] 3692/17 3650/14 3652/14 3647/18 3653/6 3718/16 3718/21
largest [3] 3654/7 liked [2] 3695/24 3652/16 3659/24 3673/22 3685/7 3719/15 3720/1
3688/19 3698/18 3737/14 3660/18 3662/15 3700/16 3701/21 3733/9 3734/24
LARRABEE [1] likely [5] 3670/16 3662/19 3679/5 3702/11 3702/13 3735/24 3742/1
3632/10 3676/22 3741/15 3683/6 3684/6 3702/22 3703/15 maps [8] 3647/22
Larry [1] 3668/3 3742/1 3742/4 3684/23 3687/7 3707/6 3707/20 3659/7 3660/1 3660/5
LaSalle [1] 3631/20 likes [1] 3725/6 3698/13 3698/20 3707/23 3708/7 3660/9 3660/15
maps... - Mr. Nadella 3759

M 3728/18 3737/2 3733/5 3740/19 modestly [1] 3734/17 3700/14 3707/21

3738/3 3738/18 3742/13 modicum [1] 3655/21 3708/8 3708/20
maps... [2] 3660/15
meaning [8] 3679/11 Microsoft's [7] 3637/23 moment [2] 3647/16 3710/21 3712/1
3680/22
3686/14 3690/23 3640/10 3641/3 3698/1 3716/7 3717/14
March [2] 3675/2
3692/17 3699/21 3644/8 3651/21 monetization [4] 3720/6 3722/1
3707/21
3713/17 3734/15 3651/23 3662/15 3667/9 3667/17 3724/15 3724/20
March 22nd [1]
3736/6 mid [1] 3702/24 3717/25 3720/13 3731/14 3732/12
3707/21
meaningful [3] 3693/19 middle [1] 3697/17 monetize [2] 3720/10 3734/24 3735/1
marginal [1] 3644/22
3693/21 3694/5 might [10] 3662/24 3722/1 mostly [4] 3646/17
marginally [1] 3644/20
means [6] 3677/20 3673/12 3673/12 money [13] 3636/6 3652/13 3685/8
marked [2] 3648/5
3694/7 3717/16 3673/14 3696/21 3664/12 3668/17 3725/4
3657/11
3731/17 3732/24 3700/8 3722/4 3674/19 3675/4 move [13] 3637/6
market [23] 3643/3
3743/13 3722/10 3724/4 3687/4 3692/4 3700/1 3637/11 3637/12
3643/12 3643/18
meant [12] 3638/25 3742/12 3703/8 3716/16 3647/23 3652/20
3643/20 3645/6
3685/8 3688/4 million [6] 3634/14 3727/7 3732/15 3661/5 3699/9 3704/9
3647/12 3651/18
3690/12 3703/10 3634/16 3634/16 3732/18 3705/19 3705/22
3652/23 3653/13
3721/7 3728/5 3728/8 3636/12 3677/6 monitor [2] 3636/6 3705/23 3706/1
3662/10 3663/21
3729/2 3738/11 3735/10 3636/8 3743/22
3663/22 3663/23
3745/22 3745/25 millions [1] 3699/6 month [4] 3676/14 moved [1] 3643/1
3663/24 3699/19
measure [1] 3742/2 milliseconds [1] 3726/14 3726/21 moving [2] 3652/10
3700/2 3709/7
measured [2] 3731/24 3726/16 3735/10 3652/13
3709/13 3710/2
3732/5 mind [5] 3670/10 monthly [1] 3734/12 Mr Dintzer [2] 3665/21
3710/4 3713/22
mechanism [2] 3693/9 3699/9 3711/6 months [1] 3734/20 3712/22
3724/3 3724/21
3693/21 3731/15 3724/11 mop [1] 3634/7 Mr. [45] 3633/8 3634/2
market-driven [1]
media [1] 3647/6 minority [2] 3645/22 more [56] 3634/12 3634/6 3636/23
3713/22
MEHTA [1] 3631/10 3645/24 3634/15 3639/14 3637/17 3637/25
marketing [4] 3691/24
memory [1] 3643/11 minus [1] 3643/17 3643/14 3646/25 3638/14 3638/19
3714/17 3714/19
Menlo [1] 3669/17 minutes [4] 3710/17 3652/12 3656/20 3639/19 3639/23
3728/5
mental [1] 3711/6 3742/18 3743/7 3656/21 3660/4 3640/8 3640/23
marketplace [1] 3699/1
mention [2] 3702/11 3743/11 3660/6 3661/15 3647/15 3647/20
masters [1] 3666/23
3702/13 misjudgments [1] 3663/19 3664/1 3648/4 3653/6
match [1] 3739/22
mentioned [4] 3696/11 3662/21 3664/3 3668/1 3672/3 3656/15 3656/18
material [1] 3743/10
3727/3 3732/8 missing [1] 3710/9 3673/10 3673/12 3661/7 3661/11
math [1] 3736/9
3736/23 misspellings [1] 3673/18 3673/18 3665/4 3665/11
matter [3] 3663/6
menus [1] 3690/18 3696/20 3674/15 3677/1 3665/15 3666/10
3745/15 3747/5
merger [1] 3641/6 mistakes [1] 3662/20 3678/22 3686/14 3675/22 3682/19
matters [1] 3744/10
merits [1] 3712/8 mobile [31] 3634/19 3687/21 3688/2 3689/2 3689/8
MATTHEW [1]
message [1] 3729/23 3634/20 3640/5 3688/18 3688/18 3697/25 3698/3
3632/10
met [1] 3666/14 3647/21 3663/7 3692/25 3696/6 3714/23 3723/19
maximize [1] 3676/25
metasearch [1] 3663/8 3663/10 3696/7 3696/8 3696/8 3743/7 3743/12
may [8] 3654/17
3648/23 3663/14 3663/19 3697/12 3699/1 3743/21 3744/6
3666/1 3675/1 3675/3
methodology [3] 3663/23 3663/24 3704/19 3709/6 3744/6 3744/24
3689/12 3689/13
3682/16 3684/25 3669/4 3669/5 3669/9 3711/11 3713/20 3744/24 3744/25
3699/13 3736/23
3735/9 3690/17 3690/17 3713/21 3717/8 3745/16 3745/16
maybe [8] 3643/15
metric [2] 3645/18 3692/2 3692/2 3719/5 3720/3 3746/3 3746/7
3644/15 3645/21
3731/23 3692/16 3692/23 3723/14 3723/14 3746/16
3645/21 3664/25
metrics [2] 3703/22 3693/23 3700/6 3723/15 3724/5 Mr. Cavanaugh [3]
3710/9 3745/23
3708/2 3704/10 3704/14 3727/17 3728/18 3714/23 3723/19
3745/24
Microsoft [40] 3632/11 3704/21 3705/10 3733/8 3733/8 3746/16
MEAGAN [1] 3631/16
3638/15 3639/21 3705/14 3716/8 3734/21 3734/21 Mr. Dintzer [4] 3633/8
meager [1] 3649/1
3640/24 3641/7 3720/3 3721/19 3735/2 3735/2 3682/19 3689/2
mean [38] 3635/12
3642/3 3644/14 3723/7 3737/13 3697/25
3635/14 3635/17
3644/18 3646/12 modalities [1] 3651/9 morning [4] 3743/13 Mr. Dischler [1]
3636/4 3639/10
3646/21 3647/5 model [15] 3674/2 3743/18 3743/22 3744/25
3643/4 3645/6
3650/2 3652/24 3674/14 3676/5 3746/20 Mr. Gates [1] 3636/23
3645/10 3646/5
3653/10 3653/16 3677/1 3679/14 most [39] 3635/5 Mr. Higgins [1]
3646/5 3648/21
3653/19 3653/20 3697/6 3697/13 3640/15 3644/1 3744/24
3648/25 3650/1
3653/24 3654/2 3698/6 3698/24 3645/21 3649/1 Mr. Jed [1] 3647/15
3651/9 3653/19
3654/15 3654/17 3699/10 3716/18 3649/15 3664/10 Mr. Jorgensen [1]
3655/7 3658/7 3660/5
3654/23 3654/24 3721/7 3735/11 3667/24 3668/17 3647/20
3663/2 3664/15
3655/18 3656/8 3740/23 3742/11 3668/19 3671/15 Mr. Juda [1] 3744/6
3664/20 3670/8
3658/5 3662/17 modeling [1] 3740/9 3675/18 3676/11 Mr. Kolotouros [1]
3672/17 3685/24
3662/20 3662/23 models [4] 3671/22 3678/12 3683/25 3744/24
3686/22 3695/6
3663/19 3664/6 3697/9 3697/20 3684/8 3688/4 Mr. Lowcock [1]
3695/7 3706/5
3671/19 3698/17 3741/21 3689/16 3689/20 3744/6
3712/17 3716/5
3713/8 3713/12 modest [2] 3677/5 3697/9 3697/18 Mr. Nadella [7] 3634/6
3716/11 3717/3
3728/9 3728/13 3699/5 3699/6 3700/14 3637/17 3648/4
3718/11 3727/6
Mr. Nadella... - one 3760

M 3637/17 3648/4 3688/24 3690/5 3642/25 OEM [1] 3713/6

3653/6 3656/18 3690/9 3690/12 nowhere [1] 3648/24 OEMs [7] 3657/19
Mr. Nadella... [4]
3661/9 3661/11 3690/14 3690/19 nuances [1] 3690/7 3658/4 3658/7 3658/8
3653/6 3656/18
3665/4 3690/25 3691/7 number [33] 3634/15 3658/14 3713/4
3661/11 3665/4
name [8] 3665/17 3691/18 3692/10 3635/21 3635/22 3713/5
Mr. Ramaswamy [8]
3666/14 3666/16 3692/13 3692/14 3644/5 3644/16 of whom [1] 3690/21
3665/11 3665/15
3666/17 3715/1 3692/23 3693/1 3648/7 3648/8 off [8] 3636/22 3670/4
3666/10 3675/22
3715/14 3720/5 3693/20 3694/10 3652/21 3654/10 3679/16 3680/8
3689/8 3698/3
3720/15 3694/15 3695/13 3655/19 3676/25 3717/5 3717/23
3743/12 3743/21
named [1] 3672/14 3696/11 3699/4 3681/14 3682/1 3740/11 3744/16
Mr. Sallet [1] 3745/16
Nana [2] 3665/13 3701/7 3712/10 3685/12 3688/7 offer [4] 3682/13
Mr. Schmidtlein [2]
3665/19 3712/12 3715/10 3690/20 3692/17 3692/10 3693/2
3634/2 3661/7
narrow [2] 3640/4 3715/11 3717/21 3694/18 3703/4 3694/12
Mr. Scott [6] 3638/14
3649/4 3718/22 3718/25 3704/5 3709/17 offered [7] 3681/22
3638/19 3639/19
natively [1] 3693/1 3719/19 3720/17 3713/18 3721/10 3683/8 3726/5
3639/23 3640/8
natural [1] 3673/2 3720/23 3721/16 3725/24 3726/21 3733/18 3736/21
3640/23
nature [2] 3692/8 3721/18 3721/23 3728/14 3731/21 3736/24 3745/17
Mr. Scott's [1] 3637/25
3706/10 3721/23 3722/16 3731/23 3734/12 offering [2] 3675/17
Mr. Severt [1] 3656/15
navigate [1] 3648/6 3724/5 3724/19 3734/20 3741/21 3738/8
Mr. Smurzynski [1]
Nebraska [2] 3631/23 3725/19 3726/4 3742/2 3742/15 offers [2] 3697/2
3743/7
3715/2 3727/5 3732/8 numbered [3] 3683/14 3738/14
Mr. Tinter [2] 3745/16
necessarily [1] 3732/15 3732/20 3703/18 3703/18 Office [2] 3733/5
3746/7
3720/14 3732/21 3733/12 numbers [6] 3677/4 3737/4
Mr. Tinter's [1] 3746/3
necessary [1] 3723/1 3733/25 3734/5 3703/12 3704/17 officer [1] 3637/2
Ms. [2] 3707/21
necessity [1] 3673/24 3734/8 3734/13 3705/16 3740/14 Official [2] 3632/23
3708/7
need [19] 3657/2 3735/11 3735/18 3741/25 3747/3
Ms. Tang [2] 3707/21
3661/17 3670/9 3736/18 3736/22 numerous [1] 3644/14 offline [1] 3652/13
3708/7
3671/10 3685/6 3737/4 3737/16 NW [3] 3631/14 offs [4] 3693/24
MSN [4] 3731/3
3694/11 3697/21 3739/14 3739/20 3631/17 3632/24 3694/6 3718/15
3731/8 3731/9 3732/3
3697/21 3699/2 3740/9 3740/11 NY [2] 3631/25 3730/8
much [34] 3641/10
3699/2 3700/18 3741/14 3742/3 3632/12 often [12] 3673/20
3646/22 3647/4
3711/6 3712/4 3717/8 3742/25 3743/1 3678/21 3692/20
3663/12 3664/11
3718/18 3718/19 Neeva's [11] 3676/15 O 3693/6 3698/7
3665/4 3665/7
3722/10 3734/20 3676/17 3694/16 object [1] 3682/17 3698/19 3711/3
3671/23 3671/23
3745/13 3694/21 3699/18 objection [17] 3637/8 3722/25 3723/1
3672/25 3674/15
needed [10] 3671/17 3700/10 3724/2 3637/13 3647/25 3730/4 3733/22
3676/2 3676/17
3677/8 3688/13 3724/7 3736/5 3740/9 3652/21 3659/17 3742/3
3677/2 3681/13
3717/22 3718/24 3741/3 3664/14 3664/17 old [2] 3684/5 3697/17
3681/22 3685/5
3725/23 3726/2 Netflix [1] 3699/8 3682/14 3682/18 older [1] 3664/3
3685/5 3688/12
3729/8 3729/10 network [1] 3646/19 3682/22 3683/1 omnibar [1] 3712/18
3688/24 3691/13
3734/15 new [16] 3631/25 3683/5 3706/8 once [5] 3665/8
3695/25 3698/11
needing [2] 3720/12 3632/12 3634/7 3706/11 3706/14 3676/20 3722/22
3698/11 3698/22
3734/15 3634/10 3634/18 3738/24 3746/5 3726/14 3726/14
3699/3 3700/11
needle [1] 3643/1 3634/19 3635/8 objections [1] 3705/20 one [78] 3635/20
3709/16 3712/2
needs [4] 3670/18 3670/3 3679/23 objectives [1] 3652/15 3635/22 3637/13
3717/24 3720/1
3670/22 3725/13 3690/12 3696/5 obligation [2] 3730/11 3639/5 3639/18
3724/4 3739/24
3733/10 3698/7 3724/9 3736/7 3640/6 3640/13
3743/19
Neeva [104] 3665/15 3733/24 3734/17 obligations [1] 3729/16 3640/14 3644/11
Mueller [2] 3665/12
3670/2 3670/24 3737/11 observable [1] 3723/9 3646/10 3646/16
3665/19
3670/24 3671/3 news [1] 3737/10 observation [1] 3646/24 3648/10
multi [1] 3686/13
3671/4 3671/9 next [12] 3639/17 3711/25 3648/13 3650/12
multi-choice [1]
3671/13 3671/20 3639/18 3650/10 observed [1] 3696/8 3653/23 3653/24
3686/13
3672/1 3672/5 3672/6 3650/17 3665/8 observing [1] 3678/14 3655/8 3655/24
multibillion [2] 3644/13
3672/7 3672/19 3667/17 3684/14 obvious [2] 3711/10 3657/1 3657/2 3657/4
3742/14
3672/22 3672/24 3692/14 3708/12 3711/13 3660/4 3660/6
multibillion-dollar [2]
3674/16 3674/19 3708/16 3717/11 obviously [9] 3640/16 3663/23 3663/24
3644/13 3742/14
3675/6 3675/23 3744/23 3645/3 3682/10 3669/7 3670/21
multiple [4] 3634/23
3676/3 3676/6 nine [1] 3663/5 3695/11 3704/17 3671/15 3675/22
3669/10 3677/10
3676/14 3676/19 Nitin [4] 3707/23 3704/22 3705/3 3677/20 3677/24
3706/1
3676/20 3676/22 3707/25 3708/1 3705/22 3744/18 3678/1 3678/12
multiplied [1] 3714/9
3676/22 3677/3 3708/8 occasions [1] 3655/19 3678/22 3681/9
mutually [2] 3727/5
3678/9 3678/10 Nokia [1] 3698/17 occurrence [1] 3682/9 3683/25
3727/7
3679/8 3679/9 Non [1] 3632/10 3661/25 3683/25 3684/8
myself [1] 3672/10
3679/11 3679/12 Non-Party [1] 3632/10 October [3] 3631/5 3685/19 3685/21
N 3679/17 3683/22 none [1] 3648/22 3667/5 3669/13 3686/12 3686/14
3686/19 3687/9 Notion [1] 3736/25 October 2018 [2] 3691/22 3692/21
NADELLA [10] 3633/3
3687/21 3688/10 notwithstanding [1] 3667/5 3669/13 3694/19 3695/19
3634/3 3634/6
one... - percent 3761

O opportunity [2] 3718/2 3718/18 painful [1] 3681/15 3699/6 3700/5

3664/24 3742/14 3720/3 3720/11 pains [1] 3726/3 3735/23 3735/25
one... [30] 3695/25
opposed [6] 3688/23 3724/12 3728/7 pandemic [1] 3692/22 3737/1
3698/15 3699/8
3722/8 3731/16 3733/2 3733/8 3742/5 paper [1] 3685/2 payload [1] 3659/22
3701/25 3706/25
3731/25 3734/1 overall [3] 3694/6 paradoxical [1] payment [1] 3694/1
3707/11 3707/21
3738/4 3712/20 3717/9 3695/22 Payments [1] 3668/10
3708/20 3709/3
optimization [1] 3650/6 overcome [1] 3735/20 paragraph [9] 3638/18 pays [1] 3664/11
3712/4 3716/1
option [4] 3655/5 overcoming [1] 3639/3 3639/4 PC [2] 3711/23
3717/13 3717/21
3655/10 3692/10 3735/11 3640/18 3648/16 3713/13
3722/17 3722/21
3692/11 overlapping [1] 3733/9 3650/11 3650/14 PCs [2] 3711/22
3724/1 3724/12
options [3] 3673/7 overly [1] 3657/16 3717/11 3740/22 3713/10
3727/3 3727/9 3728/6
3690/19 3691/6 overnight [1] 3743/18 Park [1] 3669/17 peak [1] 3734/12
3728/8 3728/16
oracle [1] 3724/12 own [9] 3641/12 parse [1] 3720/24 peaked [1] 3735/9
3730/19 3732/7
order [5] 3685/6 3668/22 3700/13 parsing [1] 3720/23 peers [3] 3672/15
3732/21 3736/15
3690/9 3716/10 3700/13 3716/14 part [25] 3641/21 3672/17 3702/18
3741/8 3744/22
3721/2 3744/24 3722/3 3732/15 3642/19 3667/25 penetration [1] 3643/7
3745/15 3746/6
ordinary [1] 3690/3 3732/18 3740/6 3668/2 3669/18 people [80] 3635/17
one-stop [3] 3670/21
organic [1] 3650/3 owned [3] 3638/11 3671/6 3672/19 3642/14 3647/5
3728/6 3728/8
organization [1] 3668/1 3643/22 3657/23 3677/8 3681/19 3649/19 3649/20
ones [11] 3663/5
organized [1] 3681/25 3684/24 3685/25 3649/24 3650/4
3681/21 3692/3
oriented [3] 3669/9 P 3686/12 3690/22 3651/2 3651/4 3651/4
3695/17 3695/23
3687/5 3688/2 p.m [4] 3631/6 3690/25 3691/14 3651/7 3651/12
3710/22 3713/20
original [2] 3727/18 3689/10 3689/11 3693/1 3698/9 3653/10 3653/16
3721/10 3735/23
3727/24 3746/21 3700/25 3701/11 3667/14 3673/3
3735/25 3737/1
others [3] 3646/8 pack [5] 3703/2 3703/2 3708/1 3718/17 3674/8 3674/10
ongoing [1] 3713/20
3646/17 3654/3 3703/9 3706/17 3723/6 3729/3 3676/11 3676/19
online [10] 3652/14
otherwise [1] 3687/10 3706/19 3729/12 3741/19 3676/20 3676/22
3673/18 3673/25
ourselves [2] 3681/4 package [1] 3706/12 part-time [2] 3669/18 3677/1 3677/20
3677/21 3678/6
3740/1 packs [1] 3706/1 3672/19 3678/23 3678/24
3681/18 3716/2
out [44] 3642/2 3643/5 page [53] 3633/2 participating [2] 3652/5 3680/7 3681/10
3722/20 3726/22
3660/18 3662/2 3633/14 3638/1 3652/9 3681/17 3685/3
3733/20
3662/3 3663/5 3669/4 3648/7 3677/18 particular [6] 3640/3 3686/4 3686/6
only [16] 3645/8
3680/25 3681/22 3683/14 3683/18 3680/3 3683/10 3686/13 3686/15
3646/16 3659/4
3683/7 3684/4 3690/3 3684/4 3684/14 3684/25 3685/5 3686/19 3687/4
3674/8 3682/22
3691/11 3692/7 3688/6 3693/15 3695/23 3687/23 3688/16
3686/10 3691/4
3692/18 3693/5 3703/17 3703/18 particularly [1] 3723/1 3690/13 3691/1
3691/6 3694/17
3696/3 3696/4 3696/6 3708/17 3709/4 parties [3] 3656/17 3692/18 3692/22
3696/1 3717/13
3696/16 3696/25 3709/5 3709/9 3671/20 3744/11 3693/6 3693/8
3717/19 3722/2
3697/11 3697/18 3709/17 3710/10 partly [2] 3692/14 3695/23 3700/3
3732/18 3735/18
3698/9 3698/22 3710/11 3710/15 3693/10 3700/23 3703/15
3746/2
3698/23 3700/1 3711/8 3711/22 partner [2] 3669/16 3705/8 3708/2 3708/4
open [7] 3657/17
3700/20 3700/21 3711/24 3712/2 3669/18 3711/13 3711/14
3658/25 3659/3
3702/5 3704/5 3712/5 3712/14 partners [3] 3669/17 3712/1 3712/16
3659/20 3665/23
3704/13 3704/20 3712/15 3712/23 3672/9 3680/22 3712/18 3712/18
3694/11 3694/11
3707/13 3719/12 3713/7 3713/11 partnership [1] 3713/21 3714/6
OpenAI [5] 3636/6
3723/4 3727/18 3713/14 3713/19 3669/16 3716/11 3717/24
3637/23 3638/3
3729/17 3736/9 3713/24 3714/5 party [3] 3632/10 3719/7 3721/24
3639/21 3640/10
3738/21 3741/20 3715/13 3720/9 3652/22 3668/25 3724/3 3724/13
OpenAI's [1] 3634/11
3743/2 3743/25 3720/13 3720/22 pass [1] 3714/22 3724/16 3725/9
opened [2] 3690/12
3745/3 3721/5 3721/20 passport [1] 3673/19 3725/13 3726/13
3711/9
out-of-the-box [2] 3721/22 3722/7 password [2] 3725/10 3727/2 3727/3 3729/4
opening [1] 3710/23
3690/3 3691/11 3722/9 3722/11 3734/16 3731/18 3731/21
openly [1] 3718/13
outset [2] 3732/20 3723/2 3726/25 past [4] 3648/17 3735/15 3735/18
opens [1] 3711/2
3739/14 3727/1 3730/24 3674/22 3695/20 3735/25 3736/4
operate [5] 3673/3
over [34] 3634/23 3731/4 3731/9 3743/7 3737/9 3741/17
3681/2 3723/15
3636/12 3641/10 3731/16 3741/12 path [1] 3686/17 per [1] 3745/6
3724/23 3726/2
3646/24 3647/2 Page's [1] 3668/3 Patterson [1] 3631/23 percent [29] 3643/15
operated [3] 3643/22
3654/15 3667/17 pages [12] 3697/3 pay [14] 3664/16 3643/15 3643/17
3669/24 3729/24
3667/18 3668/5 3697/10 3697/18 3664/21 3664/23 3645/21 3651/1
operates [3] 3667/21
3673/17 3680/19 3700/21 3707/23 3664/25 3676/11 3657/22 3658/2
3723/2 3729/25
3681/3 3686/19 3716/7 3720/15 3679/15 3681/20 3658/15 3658/16
operating [7] 3657/23
3693/13 3696/7 3721/10 3721/24 3686/2 3686/16 3658/20 3658/20
3669/5 3694/3
3696/8 3698/5 3721/25 3722/25 3687/4 3716/16 3659/2 3659/2
3694/12 3698/16
3699/10 3704/9 3732/4 3734/20 3735/19 3659/20 3659/21
3698/19 3701/6
3705/20 3706/5 paid [6] 3675/4 3685/6 3736/7 3672/19 3676/15
opine [1] 3664/17
3714/5 3714/9 3713/18 3726/20 pay if [1] 3664/16 3693/22 3695/14
opportunities [1]
3716/18 3717/5 3736/10 3736/15 paying [6] 3664/23 3698/14 3709/6
3698/8
percent... - promising 3762

P piece [1] 3725/22 positioned [2] 3671/13 presenting [1] 3697/16 3746/21 3747/5
pieces [2] 3701/8 3737/16 preset [1] 3713/5 process [2] 3683/22
percent... [8] 3731/8
3722/24 positioning [2] 3734/19 president [1] 3667/23 3696/20
3731/11 3731/12
pitch [1] 3684/24 3737/19 Press [1] 3635/5 processes [1] 3729/21
3731/20 3731/20
pivotal [1] 3718/16 positions [4] 3667/11 pressures [1] 3716/24 product [54] 3641/13
3736/10 3740/12
place [12] 3663/14 3667/15 3737/23 pretend [1] 3730/8 3643/6 3644/8
3742/10
3670/14 3670/16 3738/7 pretty [14] 3674/22 3658/25 3660/15
percentage [8]
3670/17 3677/5 positive [1] 3717/9 3677/5 3680/17 3662/23 3668/21
3645/13 3645/24
3685/9 3698/13 positives [1] 3692/25 3686/10 3688/22 3669/10 3669/23
3655/13 3676/23
3700/18 3711/16 possible [5] 3712/1 3692/23 3695/3 3669/25 3671/4
3704/6 3704/12
3729/21 3733/11 3712/3 3716/22 3700/12 3713/9 3671/6 3672/18
3731/4 3736/8
3743/9 3719/16 3724/12 3725/21 3726/10 3673/16 3674/14
percentages [1]
placement [1] 3661/20 post [2] 3692/22 3735/5 3735/21 3674/25 3675/15
3681/17
places [5] 3651/2 3719/19 3743/5 3675/19 3676/7
Perception [1] 3684/15
3665/9 3679/7 potential [4] 3671/13 preventer [2] 3722/13 3677/2 3683/21
perceptions [1] 3686/6
3693/25 3714/6 3675/2 3693/23 3722/15 3683/24 3687/17
perfectly [3] 3676/23
Plaintiff [1] 3632/2 3696/24 prevention [2] 3723/4 3687/20 3687/22
3686/4 3711/10
plaintiffs [5] 3631/4 potentially [2] 3695/12 3723/15 3688/2 3688/12
performing [1] 3636/7
3631/13 3631/22 3724/8 previously [1] 3697/13 3688/13 3688/21
perhaps [4] 3643/15
3665/10 3746/15 power [6] 3671/18 priced [1] 3676/14 3689/16 3691/19
3643/23 3648/23
planet [1] 3698/19 3671/22 3712/9 prices [1] 3688/4 3692/21 3694/18
3742/5
platform [11] 3658/25 3735/12 3735/15 pricing [2] 3676/13 3696/1 3699/3 3699/7
period [8] 3635/1
3659/3 3659/20 3735/20 3677/9 3699/24 3702/18
3635/22 3644/19
3663/18 3694/5 powered [6] 3667/13 primarily [2] 3669/9 3717/22 3721/22
3654/19 3663/8
3699/3 3705/7 3667/16 3672/4 3707/13 3721/22 3721/25
3695/15 3706/6
3727/22 3728/2 3674/17 3697/4 primary [5] 3676/19 3721/25 3729/1
3713/19
3728/5 3729/1 3697/13 3716/19 3716/20 3729/13 3729/17
periods [1] 3706/14
platforms [6] 3704/22 powerful [4] 3676/10 3723/25 3724/1 3729/21 3729/25
permanent [1] 3734/23
3705/3 3705/4 3696/2 3709/24 principles [1] 3718/20 3730/7 3733/1
permission [1] 3694/14
3728/11 3729/3 3711/25 priorities [2] 3644/11 3734/19 3735/2
perpetually [1] 3711/1
3730/1 powers [1] 3722/20 3694/20 3735/6 3736/6
person [3] 3678/3
play [1] 3713/10 practical [1] 3730/4 prioritize [1] 3640/14 product's [1] 3675/17
3692/19 3693/12
player [2] 3741/8 practice [1] 3700/22 prioritizing [1] 3716/25 product-oriented [1]
personal [4] 3673/17
3741/19 precautions [1] priority [1] 3692/23 3688/2
3679/2 3718/23
players [1] 3741/14 3681/14 privacy [30] 3661/1 products [25] 3651/7
3733/20
please [7] 3666/4 precise [1] 3728/18 3668/15 3677/14 3651/12 3651/13
personalization [7]
3666/11 3666/15 precisely [1] 3735/6 3677/19 3677/25 3662/16 3662/20
3673/15 3679/21
3675/21 3697/22 predicated [1] 3699/4 3678/3 3678/9 3662/21 3668/8
3680/2 3680/5 3680/6
3709/22 3710/19 predictable [4] 3680/13 3680/23 3681/8 3668/9 3668/25
3737/5 3737/6
plus [3] 3643/17 3680/17 3699/21 3681/9 3681/18 3669/2 3669/8
personalize [3]
3671/16 3718/2 3700/17 3681/23 3719/3 3670/12 3673/5
3673/10 3679/13
point [18] 3647/21 predominant [2] 3723/8 3723/25 3676/10 3684/1
3679/18
3650/9 3658/21 3645/10 3647/11 3724/3 3724/14 3685/1 3687/7
personalized [1]
3659/3 3663/2 predominantly [2] 3724/15 3724/16 3687/15 3688/5
3719/5
3663/23 3674/5 3646/20 3652/12 3725/2 3725/4 3725/5 3688/19 3698/7
perspective [1]
3698/6 3698/24 preempting [1] 3725/9 3725/13 3703/5 3722/2
3643/17
3713/23 3714/3 3699/13 3726/11 3736/1 3722/23 3736/25
persuade [3] 3655/19
3716/15 3716/18 prefer [3] 3673/12 3738/8 3738/12 profiles [4] 3722/19
3656/9 3658/21
3733/7 3733/25 3679/24 3737/11 3738/13 3738/22 3723/5 3723/10
pertain [1] 3673/14
3742/6 3742/11 preferences [1] private [8] 3671/14 3739/11
petering [1] 3692/6
3742/21 3673/11 3672/2 3737/16 profitability [1] 3644/23
Petitioner [1] 3632/11
point is [1] 3742/21 preferred [1] 3737/9 3737/23 3737/25 profitable [6] 3644/8
PhD [1] 3666/23
pointed [1] 3643/5 premier [1] 3728/10 3738/2 3738/3 3738/9 3644/10 3644/13
Phillipp [1] 3702/16
pointing [1] 3741/20 premise [1] 3664/2 pro [1] 3727/9 3644/20 3668/20
phone [6] 3663/14
points [6] 3647/1 premium [4] 3677/1 probably [3] 3646/17 3684/9
3680/18 3694/1
3655/22 3676/23 3679/15 3736/6 3653/2 3718/16 profits [1] 3644/18
3694/15 3700/17
3714/5 3714/8 3736/9 3737/1 problem [10] 3657/9 program [1] 3713/17
3716/8
politely [1] 3742/17 prepared [3] 3682/12 3658/13 3697/8 programmatic [1]
phones [4] 3692/3
Political [1] 3687/4 3703/12 3744/16 3697/17 3700/16 3671/19
3692/4 3692/12
popular [2] 3635/6 presence [1] 3709/1 3718/19 3733/19 progress [1] 3634/22
3704/24
3733/5 present [2] 3643/1 3736/4 3740/12 progressively [1]
phrase [3] 3723/17
popularity [1] 3635/2 3709/2 3745/20 3735/2
3725/16 3735/13
popularized [1] presentation [3] problems [5] 3671/16 projects [1] 3708/5
pick [7] 3663/5 3673/8
3735/13 3653/7 3682/9 3696/18 3699/20 promise [1] 3680/9
3686/14 3691/2
portion [1] 3740/22 3709/12 3717/12 3730/4 promises [1] 3726/3
3693/6 3722/22
portrayed [1] 3687/13 presented [3] 3706/12 proceed [1] 3689/12 promising [2] 3639/11
3737/9
position [1] 3663/17 3708/6 3726/23 proceedings [2] 3692/6
pictures [1] 3688/3
promoted - request 3763

P 3707/12 3707/17 Rajesh [1] 3742/12 reasonably [1] 3739/10 relationships [2]
3708/24 3708/25 RAMASWAMY [15] reasoning [1] 3675/11 3725/24 3726/12
promoted [1] 3667/23
3717/4 3737/6 3633/7 3665/11 reasons [2] 3640/13 relative [2] 3634/20
prompt [1] 3741/22
3738/14 3738/23 3665/15 3666/10 3717/21 3635/20
pronounced [1] 3699/1
3739/18 3739/22 3666/12 3666/17 recall [9] 3643/14 relatively [1] 3700/17
properties [1] 3668/23
3740/12 3742/9 3675/22 3689/8 3647/16 3654/17 released [2] 3636/10
prophecy [1] 3724/10
3742/18 3742/19 3698/3 3714/24 3661/23 3662/13 3662/8
proposals [2] 3663/4
3742/22 3743/6 3730/14 3730/16 3662/14 3663/8 relentless [1] 3672/25
3744/25
quarter [2] 3702/24 3740/17 3743/12 3667/23 3740/14 relevance [2] 3682/23
proposed [3] 3660/20
3702/25 3743/21 recap [1] 3702/2 3683/1
3744/14 3745/7
queries [42] 3641/7 ran [6] 3647/21 received [4] 3635/9 relevancy [1] 3738/24
proposing [1] 3744/11
3642/14 3645/14 3672/16 3700/12 3636/12 3687/9 relevant [6] 3682/24
proposition [3] 3664/7
3645/22 3645/24 3702/17 3707/17 3732/4 3697/18 3713/22
3683/9 3699/4
3645/25 3646/13 3727/15 receives [1] 3645/14 3716/13 3716/21
propositions [1]
3649/4 3649/5 range [1] 3645/13 recent [1] 3664/1 3738/25
3675/23
3649/24 3664/10 ranking [1] 3646/10 Recess [1] 3689/10 reliable [1] 3694/17
protects [1] 3738/12
3671/25 3676/21 rate [1] 3740/11 recipients [2] 3701/25 reliance [1] 3716/18
provide [6] 3666/5
3676/21 3686/25 rates [2] 3674/21 3702/11 relied [2] 3681/19
3670/19 3671/24
3687/1 3687/2 3687/4 3736/8 recognition [1] 3657/15 3695/20
3695/16 3733/6
3687/5 3687/6 3688/3 rather [4] 3646/14 recognize [2] 3676/18 reluctant [3] 3673/25
3739/15
3688/19 3688/23 3694/11 3701/17 3682/6 3674/23 3676/11
provided [3] 3641/19
3693/4 3695/11 3705/10 recognized [3] 3646/12 reluctantly [1] 3674/8
3654/24 3676/8
3695/11 3695/14 ratio [1] 3741/17 3647/5 3657/16 rely [1] 3679/22
Providence [1]
3696/3 3696/4 3696/7 reaction [1] 3724/1 recollection [4] relying [4] 3731/16
3666/24
3696/20 3712/18 reactions [2] 3639/8 3740/16 3741/1 3732/3 3741/5 3742/9
provider [4] 3701/5
3720/6 3720/6 3639/13 3745/23 3745/24 remain [1] 3666/25
3723/18 3724/20
3720/11 3727/3 read [9] 3640/11 record [3] 3681/4 remained [1] 3714/10
3725/17
3731/20 3731/23 3640/20 3702/4 3689/11 3747/5 remains [1] 3647/11
providers [4] 3649/20
3732/5 3742/1 3743/4 3704/5 3704/12 redacted [1] 3703/19 remarkably [1]
3717/20 3726/6
3743/5 3707/20 3714/15 redactions [2] 3744/11 3677/19
3737/9
query [32] 3642/21 3740/21 3740/22 3744/15 remember [12]
provides [1] 3655/4
3646/6 3646/8 readily [2] 3655/5 Reddit [1] 3721/24 3643/10 3643/12
providing [4] 3641/11
3673/14 3687/9 3724/6 redirect [4] 3633/5 3644/15 3657/17
3671/24 3725/20
3687/18 3688/15 reading [2] 3702/6 3661/8 3661/9 3665/2 3684/5 3691/5
3733/1
3694/22 3694/25 3740/23 reduces [1] 3721/10 3693/22 3694/11
PSX1023 [1] 3719/18
3695/17 3695/21 reads [4] 3640/19 refer [3] 3645/25 3708/11 3708/13
public [6] 3653/2
3695/23 3695/25 3640/21 3702/7 3725/19 3727/13 3709/12 3741/24
3660/11 3674/16
3696/5 3696/19 3740/24 reference [4] 3638/3 remind [2] 3636/25
3675/7 3734/5 3738/4
3696/21 3696/22 ready [2] 3634/2 3684/1 3707/10 3707/15
published [1] 3685/2
3696/24 3697/12 3665/8 3727/14 remove [1] 3680/24
publishers [3] 3668/9
3697/14 3697/18 real [2] 3674/6 3700/1 references [1] 3719/9 renders [1] 3716/9
3669/1 3703/6
3697/19 3703/25 reality [2] 3648/21 referred [3] 3650/20 renewed [1] 3654/21
purchase [3] 3676/2
3704/7 3704/10 3724/17 3717/12 3728/4 reorganization [1]
3694/4 3734/9
3704/14 3711/16 realization [1] 3671/21 referring [2] 3684/11 3668/1
purchasing [1] 3676/16
3725/20 3727/4 realize [1] 3718/17 3740/22 repeat [3] 3638/23
purpose [5] 3645/12
3741/12 3741/18 realized [1] 3671/22 refers [7] 3683/23 3694/24 3718/8
3689/17 3710/5
3742/4 really [26] 3643/1 3683/24 3684/3 replacement [1]
3716/19 3716/21
query-related [1] 3661/21 3668/21 3684/25 3686/24 3685/3
pursue [2] 3674/14
3696/21 3669/21 3670/4 3695/24 3735/15 report [1] 3648/9
3701/12
querying [1] 3695/2 3671/23 3672/20 reformulate [1] reported [4] 3650/24
push [3] 3714/17
quickly [3] 3705/10 3673/6 3677/3 3680/1 3741/18 3702/20 3703/13
3714/19 3717/24
3716/22 3721/1 3680/4 3686/2 3686/7 refresh [2] 3740/16 3707/17
pushes [1] 3716/3
quite [8] 3673/10 3686/16 3692/10 3741/1 Reporter [3] 3632/23
put [8] 3637/18 3649/2
3674/2 3677/7 3693/11 3697/6 refunded [1] 3675/4 3632/23 3747/3
3659/8 3671/25
3689/19 3711/3 3711/19 3713/15 regarding [2] 3638/15 reporting [1] 3653/2
3681/24 3714/17
3729/14 3733/16 3719/5 3724/3 3724/4 3715/24 reports [1] 3635/12
3714/18 3732/7
3736/15 3727/6 3730/8 regards [1] 3731/2 represent [4] 3715/2
PX1022 [1] 3715/4
3734/23 3735/23 region [1] 3709/8 3729/2 3729/9 3730/6
PX1023 [1] 3721/5 R realtime [1] 3719/10 regions [1] 3709/9 representation [1]
Q R-A-M-A-S-W-A-M-Y reason [4] 3659/4 regular [2] 3653/17 3706/13
[1] 3666/18 3693/6 3701/11 3721/3 represented [3]
Q2 [4] 3702/2 3702/23
race [1] 3699/24 3728/14 regularly [1] 3712/12 3728/10 3730/10
3704/7 3704/14
raise [1] 3666/4 reasonable [7] reject [1] 3663/5 3746/8
quadrupled [1]
raised [4] 3656/2 3677/22 3678/3 related [4] 3689/25 reputation [2] 3741/8
3634/24
3656/5 3674/18 3678/7 3679/16 3696/20 3696/21 3741/10
quality [21] 3655/25
3732/8 3725/6 3743/9 3697/11 request [2] 3681/3
3656/3 3667/20
raising [1] 3674/25 3744/22 relationship [1] 3717/4 3726/15
3687/14 3696/10
requests - set 3764

R 3668/19 3669/6 rolling [1] 3745/12 saying [6] 3659/20 Section [1] 3632/3
3677/5 3677/10 rough [3] 3645/13 3675/9 3717/3 3724/9 security [1] 3725/9
requests [1] 3726/8
3701/4 3703/6 3703/6 3695/1 3705/13 3742/17 3743/3 seeing [1] 3688/10
require [1] 3637/21
3713/20 3717/1 roughly [10] 3634/14 scale [15] 3642/17 seems [3] 3709/2
requirements [1]
3717/5 3717/8 3643/12 3643/17 3643/8 3643/16 3709/8 3722/23
3726/22
3742/14 3658/1 3672/19 3663/19 3669/6 select [1] 3720/17
reserved [1] 3737/1
revenues [2] 3668/16 3675/11 3676/14 3695/4 3695/6 selected [1] 3654/25
resolve [1] 3705/21
3668/24 3684/13 3702/24 3695/24 3698/12 self [1] 3724/10
respect [4] 3657/16
reviews [1] 3688/2 3714/6 3698/25 3700/12 self-fulfilling [1]
3665/22 3722/6
revolutionary [1] round [4] 3672/8 3733/21 3733/22 3724/10
3726/4
3684/1 3672/8 3672/11 3736/14 3740/1 sell [3] 3657/20 3658/4
respectful [1] 3657/12
Rhode [1] 3666/24 3696/23 scale as [1] 3642/17 3658/14
respectively [1]
right [95] 3634/13 rounds [1] 3672/7 scan [2] 3716/7 selling [3] 3658/5
3719/14
3634/19 3635/6 route [1] 3692/16 3716/10 3658/7 3658/21
respects [1] 3639/20
3635/10 3635/17 row [1] 3704/2 scans [1] 3680/18 sells [1] 3675/12
respond [1] 3723/22
3635/19 3636/2 rules [1] 3729/5 schedule [1] 3744/3 SEM [1] 3728/4
response [1] 3637/21
3636/5 3636/10 run [9] 3641/12 3668/5 scheme [1] 3720/11 semantically [3]
responses [1] 3685/7
3636/13 3637/3 3675/7 3675/14 Schindler [1] 3702/16 3696/4 3696/20
responsibilities [2]
3638/11 3638/12 3700/20 3720/25 SCHMIDTLEIN [4] 3697/11
3667/11 3729/16
3640/8 3640/10 3721/3 3722/25 3632/6 3633/4 3634/2 semantically-related [2]
responsibility [1]
3641/9 3641/22 3736/16 3661/7 3696/20 3697/11
3667/25
3642/4 3642/11 running [9] 3667/18 school [1] 3645/1 semi [1] 3734/23
responsible [1]
3642/15 3642/21 3667/24 3702/19 science [1] 3666/22 semi-permanent [1]
3707/13
3643/19 3643/21 3704/25 3707/18 scientists [1] 3708/3 3734/23
rest [1] 3708/9
3643/24 3643/25 3711/1 3733/21 scope [4] 3721/17 send [7] 3641/6
restricted [1] 3729/20
3644/6 3644/14 3733/22 3742/12 3721/18 3721/23 3642/4 3680/18
rests [1] 3708/21
3646/15 3648/19 runs [3] 3667/20 3738/21 3693/14 3702/20
result [9] 3643/16
3649/5 3650/8 3651/1 3700/1 3702/24 scopes [2] 3661/14 3714/13 3726/9
3693/7 3721/3
3653/14 3653/17 3661/16 sender [2] 3702/22
3721/10 3727/7
3653/21 3653/25 S Scott [8] 3636/23 3745/22
3741/12 3742/3
3654/8 3654/9 S-R-I-D-H-A-R [1] 3636/25 3638/14 sense [6] 3646/25
3742/25 3743/1
3654/12 3654/13 3666/17 3638/19 3639/19 3663/4 3675/13
resulted [1] 3662/9
3655/25 3656/1 SA360 [5] 3727/14 3639/23 3640/8 3676/1 3678/23
results [44] 3641/8
3656/3 3656/6 3727/17 3728/17 3640/23 3694/9
3641/11 3641/16
3657/23 3657/24 3728/22 3729/24 Scott's [1] 3637/25 sensitive [3] 3637/19
3641/19 3642/4
3658/3 3658/6 3658/8 Safari [13] 3655/1 scratch [2] 3670/3 3637/21 3657/16
3642/7 3650/3
3658/17 3659/15 3655/20 3656/9 3700/12 sent [4] 3702/13
3671/20 3673/4
3660/12 3666/4 3658/22 3664/2 screen [6] 3637/19 3743/5 3746/1 3746/8
3673/9 3673/13
3670/25 3679/19 3664/10 3690/24 3683/16 3683/17 sentence [8] 3638/24
3679/12 3679/13
3683/19 3683/22 3691/3 3691/6 3691/8 3685/12 3701/16 3639/17 3640/18
3680/3 3680/23
3684/20 3685/7 3691/23 3692/15 3703/19 3650/10 3650/17
3686/9 3687/12
3685/19 3687/5 3694/12 seal [1] 3744/12 3702/4 3719/10
3687/16 3693/2
3687/20 3688/15 safe [2] 3665/5 seams [3] 3698/20 3742/7
3693/11 3695/23
3690/2 3691/4 3725/11 3699/11 3699/14 SEO [5] 3646/10
3696/6 3696/12
3696/12 3699/9 safeguards [1] search [355] 3646/19 3646/20
3696/24 3697/1
3701/21 3701/25 3680/23 searched [2] 3679/25 3650/5 3650/9
3708/24 3716/3
3704/3 3706/7 3710/6 sales [1] 3729/17 3733/11 September [1] 3669/13
3716/9 3716/22
3710/15 3713/8 SALLET [2] 3632/2 searches [10] 3646/14 sequence [1] 3730/6
3720/13 3720/18
3714/13 3714/19 3745/16 3650/3 3651/17 Sequoia [3] 3672/9
3720/20 3721/4
3714/20 3716/2 same [27] 3634/14 3679/6 3679/11 3672/12 3732/9
3725/20 3726/6
3730/9 3730/23 3641/20 3649/10 3680/19 3709/6 series [1] 3717/6
3726/25 3736/2
3731/5 3731/8 3731/9 3651/11 3652/4 3715/24 3731/4 seriously [1] 3711/20
3737/12 3739/15
3731/12 3732/10 3657/3 3657/4 3657/5 3731/8 serve [2] 3673/23
3739/18 3739/21
3732/19 3732/22 3660/7 3672/18 searching [10] 3649/25 3725/23
3740/10 3741/5
3734/6 3734/7 3673/22 3677/17 3651/7 3651/8 serves [1] 3724/10
3741/13
3734/10 3735/12 3706/13 3718/19 3651/12 3673/21 service [5] 3675/24
resume [1] 3689/6
3736/19 3739/15 3720/19 3720/25 3673/21 3675/16 3679/8 3679/14
retrieval [2] 3696/24
3745/7 3746/9 3721/3 3722/5 3716/4 3716/6 3744/1 3681/2 3681/21
3697/4
rightly [1] 3643/5 3733/19 3741/17 seat [1] 3666/11 services [3] 3684/18
retrospect [1] 3711/9
rights [1] 3636/8 3742/4 3742/7 3742/8 Seattle [1] 3729/8 3685/3 3687/7
return [4] 3641/7
rise [1] 3669/4 3742/18 3742/22 second [12] 3638/18 serving [3] 3668/22
3697/6 3720/20
rival [1] 3698/17 3743/16 3744/2 3648/16 3660/24 3701/9 3736/14
3743/13
robustly [1] 3677/7 satisfy [1] 3687/13 3672/11 3684/8 session [2] 3631/7
return-a-set-of-links [1]
role [5] 3669/18 SATYA [3] 3633/3 3685/21 3706/22 3665/22
3697/6
3669/19 3705/12 3634/3 3661/9 3708/12 3709/23 set [26] 3671/24
returning [1] 3716/25
3707/15 3713/9 savvy [1] 3686/4 3716/23 3719/17 3676/5 3690/9
returns [1] 3673/19
rolled [1] 3719/12 saw [2] 3664/3 3734/5 3746/6 3690/19 3690/19
revenue [13] 3642/6
set... - states 3765

S 3679/6 3679/12 3677/25 3678/1 3736/18 3739/9 special [1] 3661/1

3680/6 3686/25 3678/4 3678/8 3739/11 3745/24 specialization [2]
set... [21] 3690/23
3687/1 3687/3 3678/16 3678/22 somebody's [1] 3648/17 3650/12
3691/1 3691/4 3691/7
3687/24 3688/14 3680/3 3711/17 3680/25 specialized [10]
3692/15 3697/6
3693/7 3693/11 3720/12 3722/3 somehow [1] 3651/5 3661/12 3669/2
3698/6 3703/9
3696/6 3697/23 3722/18 3723/8 someone [2] 3662/24 3670/13 3723/17
3705/19 3705/22
3700/18 3709/5 3723/9 3716/15 3723/17 3725/16
3705/23 3706/4
3716/10 3720/18 sites [19] 3646/22 sometimes [10] 3725/17 3726/6
3706/16 3709/6
3720/21 3721/4 3647/6 3648/11 3635/18 3645/25 3726/12 3737/1
3729/2 3729/7 3731/9
3721/23 3722/2 3649/11 3651/3 3662/3 3662/5 3678/2 specific [5] 3670/15
3734/24 3735/7
3722/6 3722/23 3651/5 3668/25 3681/15 3699/11 3670/19 3684/11
3744/23 3745/6
3739/7 3673/4 3678/14 3699/12 3726/13 3711/5 3740/14
sets [3] 3709/23
showed [5] 3684/24 3678/14 3678/22 3736/2 spell [2] 3666/16
3710/10 3710/11
3719/15 3726/5 3687/24 3693/7 somewhat [2] 3669/19 3697/10
setting [5] 3709/4
3742/7 3745/16 3693/14 3695/2 3686/3 spend [3] 3692/4
3712/14 3713/14
showing [5] 3668/25 3695/10 3695/17 son [1] 3720/16 3729/10 3730/1
3735/3 3735/3
3674/8 3693/2 3712/19 3722/2 soon [3] 3669/21 split [1] 3672/8
seven [1] 3685/16
3695/10 3739/6 situation [1] 3718/1 3687/17 3711/23 spoke [1] 3737/5
several [7] 3676/23
shown [5] 3667/14 situations [2] 3690/4 sophisticated [1] Spotlight [2] 3654/20
3691/23 3703/15
3680/3 3701/18 3716/16 3732/13 3655/23
3708/23 3722/18
3707/13 3739/12 six [1] 3636/13 sorry [24] 3638/23 spreadsheets [1]
3735/10 3736/8
shows [6] 3645/14 sizable [1] 3646/16 3652/18 3652/18 3743/6
SEVERT [3] 3631/16
3685/7 3693/5 3731/2 slate [1] 3698/10 3656/16 3664/17 sprouting [1] 3669/1
3633/5 3656/15
3731/7 3731/21 sleeker [1] 3700/24 3665/14 3675/20 squishy [2] 3677/19
shaped [1] 3718/10
shut [4] 3675/3 3675/9 slice [1] 3649/5 3679/17 3702/12 3681/10
share [34] 3642/6
3675/15 3711/5 slide [8] 3683/23 3705/25 3706/18 SRIDHAR [5] 3633/7
3643/3 3643/12
sic [1] 3634/12 3684/14 3684/25 3707/3 3710/9 3666/12 3666/17
3643/18 3643/20
side [7] 3643/4 3685/11 3704/9 3712/22 3717/18 3714/24 3730/14
3643/25 3644/1
3688/15 3715/19 3730/24 3731/2 3718/8 3727/13 stable [1] 3736/16
3644/3 3644/4 3646/6
3722/8 3729/21 3731/20 3734/14 3736/12 staff [3] 3668/3
3646/7 3648/25
3729/22 3735/4 small [9] 3635/21 3737/25 3740/3 3668/12 3691/23
3649/1 3657/22
sided [1] 3699/1 3663/22 3663/23 3743/24 3744/1 stake [1] 3640/5
3658/1 3658/15
sides [1] 3662/6 3677/7 3677/9 3693/3 3744/2 stand [4] 3665/11
3658/16 3658/20
sign [2] 3734/15 3697/13 3722/24 sort [39] 3635/14 3705/21 3743/2
3659/23 3662/10
3744/16 3745/15 3635/15 3635/22 3744/13
3663/23 3663/25
signals [2] 3679/19 smaller [2] 3721/2 3639/20 3642/2 standalone [1] 3644/8
3663/25 3664/22
3695/19 3741/14 3644/1 3645/19 standard [1] 3712/6
3703/25 3704/7
significance [1] smallmerchant.com [2] 3645/19 3647/8 start [11] 3649/24
3704/10 3704/14
3638/21 3678/17 3678/18 3648/20 3650/9 3665/14 3671/3
3709/7 3709/8
significant [4] 3639/15 smart [1] 3698/8 3660/6 3661/15 3671/9 3700/15
3709/13 3710/2
3639/24 3643/2 SMURZYNSKI [3] 3662/7 3663/16 3707/3 3715/3 3720/7
3710/4 3713/20
3697/8 3632/7 3633/10 3664/24 3671/21 3729/4 3730/13
shared [2] 3668/4
significantly [1] 3743/7 3679/5 3679/9 3743/15
3711/3
3641/17 Snowflake [5] 3675/5 3679/24 3680/10 started [9] 3636/22
shareholders [2]
Silicon [3] 3669/17 3675/8 3675/11 3686/10 3688/2 3670/3 3675/2
3664/25 3716/25
3672/10 3732/13 3675/14 3675/15 3688/9 3691/20 3696/11 3712/14
shares [2] 3706/6
silly [1] 3663/2 Snowflake's [1] 3695/22 3696/18 3717/21 3734/14
3708/16
similar [2] 3654/2 3675/19 3699/23 3700/19 3737/19 3740/1
Sharma [1] 3708/1
3741/16 soap [1] 3670/16 3701/4 3711/10 starting [5] 3646/22
Shearer [2] 3665/13
similarly [1] 3680/18 social [7] 3646/19 3712/10 3713/22 3650/25 3698/5
3665/20
simple [6] 3677/20 3647/6 3650/21 3718/12 3730/2 3711/20 3733/24
shift [1] 3644/1
3688/16 3698/3 3651/3 3651/4 3652/3 3734/18 3734/22 starts [3] 3639/5
shipped [3] 3690/22
3711/25 3725/8 3715/19 3734/23 3744/23 3650/15 3650/17
3723/5 3723/6
3725/12 software [9] 3640/6 sound [2] 3634/13 startup [1] 3699/23
shoe [1] 3722/22
simpler [1] 3699/3 3667/7 3667/12 3686/10 startups [3] 3692/20
shoestring [1] 3719/7
simplest [1] 3678/5 3675/13 3687/19 sounds [1] 3725/5 3698/23 3737/18
shop [3] 3670/21
simply [6] 3684/3 3701/9 3722/18 source [4] 3653/2 stat [3] 3706/1
3728/6 3728/8
3686/15 3712/5 3725/22 3743/4 3679/7 3680/17 3706/17 3706/19
shopping [6] 3650/21
3712/9 3712/10 sold [2] 3692/4 3725/22 state [11] 3631/22
3661/14 3668/10
3719/1 3711/22 sources [3] 3652/22 3631/23 3632/2
3669/3 3678/16
single [4] 3656/15 sole [1] 3668/6 3679/23 3737/10 3666/16 3702/21
3688/21
3697/16 3733/6 solemnly [1] 3666/5 South [1] 3631/20 3702/21 3702/23
short [3] 3677/18
3733/10 somebody [13] 3636/5 space [2] 3654/4 3703/11 3712/11
3705/5 3739/4
Siri [2] 3654/20 3691/7 3694/10 3654/7 3715/17 3718/13
shortly [1] 3689/7
3655/22 3702/22 3711/2 spare [1] 3657/1 states [9] 3631/1
shot [1] 3668/20
site [16] 3652/3 3711/9 3711/11 speaking [2] 3673/23 3631/3 3631/11
show [26] 3636/18
3677/23 3677/23 3724/2 3731/25 3686/9 3665/2 3703/8 3704/3
3673/25 3674/9
states... - timing 3766

S 3699/6 3735/25 surveys [5] 3681/11 tapping [2] 3689/22 3673/11 3677/4
subscription [8] 3681/16 3682/15 3720/23 3682/11 3700/7
states... [3] 3715/2
3675/24 3676/3 3685/15 3686/8 task [2] 3670/9 3711/5 3729/11 3729/12
3715/12 3730/17
3676/6 3676/16 Susan [2] 3668/4 tax [1] 3673/19 3729/20 3729/24
stay [2] 3724/2
3677/5 3679/14 3668/5 team [34] 3635/13 3736/16 3744/21
3734/22
3699/3 3734/9 susceptible [1] 3647/17 3667/8 Terrific [1] 3665/25
stayed [1] 3713/19
subscriptions [2] 3648/11 3667/15 3667/16 territory [1] 3660/7
staying [1] 3659/13
3676/4 3737/10 suspect [1] 3684/23 3667/18 3667/19 testified [4] 3644/7
stays [1] 3679/9
subsequent [1] sustain [1] 3659/17 3667/19 3667/20 3644/10 3654/14
steady [1] 3742/24
3667/10 sustainable [2] 3668/2 3668/6 3742/18
step [2] 3690/15
subsequently [1] 3736/11 3736/13 3668/14 3672/16 testimony [8] 3641/2
3712/4
3639/14 SVP [1] 3668/3 3675/1 3688/21 3651/20 3652/8
steps [6] 3671/2
substantial [6] 3636/5 SVPs [1] 3725/19 3688/22 3691/24 3665/5 3666/5 3689/8
3681/15 3690/8
3644/16 3662/9 SW [1] 3632/8 3692/19 3694/20 3732/8 3743/17
3690/25 3725/9
3672/7 3674/19 swear [1] 3666/5 3699/7 3702/19 tests [1] 3643/11
3725/12
3692/19 switch [6] 3656/9 3702/23 3703/13 thanks [1] 3643/8
stick [1] 3676/22
substantially [1] 3658/22 3674/8 3703/16 3707/12 that they [1] 3660/23
still [13] 3641/13
3674/22 3686/19 3741/22 3707/12 3707/17 then had [1] 3728/17
3646/18 3647/11
substitutes [1] 3651/3 3742/5 3708/2 3719/7 theories [1] 3648/13
3663/17 3667/1
success [7] 3646/16 switched [1] 3660/14 3724/19 3727/15 therefore [2] 3640/16
3674/16 3684/19
3647/3 3648/24 switching [3] 3659/14 3729/7 3729/12 3680/4
3688/7 3694/10
3655/22 3659/11 3686/23 3689/4 3729/17 thesis [5] 3648/13
3697/19 3719/11
3694/16 3735/1 Sydney [1] 3719/13 teams [4] 3663/3 3671/4 3671/6
3730/16 3742/12
successful [5] 3662/17 synonymous [1] 3668/15 3702/18 3672/24 3700/10
stock [1] 3725/25
3662/20 3699/15 3690/1 3707/19 third [9] 3635/5 3639/2
stop [4] 3670/21
3724/5 3739/10 system [9] 3667/22 tech [1] 3686/4 3639/4 3652/22
3728/6 3728/8 3743/9
suffer [1] 3660/21 3687/14 3687/19 tech-savvy [1] 3686/4 3654/7 3668/25
stopped [2] 3641/22
sufficient [1] 3669/24 3694/4 3698/16 technical [3] 3671/10 3671/20 3683/14
3641/24
suggest [3] 3662/24 3698/19 3700/12 3698/12 3743/4 3740/22
storage [1] 3673/20
3682/23 3709/3 3700/21 3701/9 technically [1] 3671/15 third-party [2] 3652/22
Store [2] 3635/6
Suite [3] 3631/20 systems [4] 3669/5 techniques [1] 3688/14 3668/25
3635/14
3631/24 3733/5 3680/22 3694/12 technological [1] Thomas [2] 3665/12
stored [3] 3679/24
summaries [1] 3702/20 3701/6 3638/21 3665/19
3679/25 3694/1
summarization [1] technologies [3] though [6] 3654/23
story [2] 3690/17
3697/15 T 3638/16 3641/24 3655/17 3664/22
3724/7
summary [1] 3703/9 tab [1] 3690/12 3733/15 3685/8 3693/24
straight [1] 3680/12
Sundar [1] 3668/10 table [1] 3685/7 technologist [1] 3723/13
strategic [3] 3659/6
supplant [1] 3697/21 tailored [1] 3687/16 3698/19 thought [11] 3658/19
3708/17 3709/24
supplier [1] 3660/24 takeaways [1] 3709/20 technologists [1] 3669/23 3686/7
strategy [2] 3660/18
support [3] 3673/21 taker [1] 3685/5 3699/11 3699/14 3717/20
3660/20
3674/25 3699/7 talented [1] 3724/19 technology [17] 3717/22 3718/20
streaming [1] 3699/10
supported [9] 3671/7 talk [13] 3644/22 3634/11 3637/2 3723/24 3733/10
Street [3] 3631/14
3673/2 3673/16 3645/2 3648/15 3638/11 3639/8 3745/24 3745/24
3631/17 3631/20
3678/11 3678/12 3662/12 3695/21 3639/11 3640/25 thousand [1] 3733/23
strength [1] 3708/25
3699/10 3716/24 3709/2 3712/10 3641/20 3659/11 thousands [1] 3723/10
strike [2] 3651/21
3723/12 3723/12 3716/1 3716/11 3659/24 3666/22 three [5] 3667/17
3712/23
supporting [1] 3640/9 3725/9 3738/18 3675/15 3675/16 3731/2 3745/1 3745/3
striking [1] 3712/24
sure [25] 3634/9 3740/14 3745/13 3701/8 3722/5 3723/3 3745/5
strong [3] 3643/6
3638/24 3640/24 talked [11] 3645/3 3733/12 3733/21 three-day [1] 3745/3
3671/10 3687/8
3653/18 3654/5 3645/3 3660/5 teens [1] 3643/20 throughout [2] 3652/22
struck [3] 3691/16
3655/12 3664/1 3670/24 3691/12 telling [1] 3722/9 3694/21
3711/23 3713/24
3666/17 3669/20 3714/4 3719/3 3719/4 tells [2] 3647/2 throw [1] 3741/23
structure [1] 3647/3
3677/17 3686/21 3720/2 3726/8 3687/20 Thursday [2] 3744/10
struggled [1] 3674/19
3687/15 3694/25 3744/10 temporal [1] 3706/5 3745/8
stuck [1] 3676/22
3704/20 3710/12 talking [10] 3648/17 tend [2] 3695/15 tide [1] 3723/13
studied [1] 3686/25
3714/16 3718/9 3652/1 3683/21 3698/25 tied [1] 3720/13
study [1] 3714/16
3720/17 3723/21 3688/9 3689/17 tends [3] 3669/18 tightly [1] 3729/20
stuff [3] 3716/10
3726/9 3730/9 3738/4 3706/9 3707/9 3677/19 3717/13 TikTok [6] 3651/12
3726/24 3735/16
3738/16 3739/9 3710/20 3727/8 tens [1] 3681/16 3651/14 3651/17
subject [4] 3653/13
3745/21 3744/1 term [10] 3646/1 3651/17 3651/19
3702/2 3707/22
surface [4] 3669/2 talks [5] 3648/9 3670/6 3686/20 3652/2
3707/22
3688/3 3700/17 3648/17 3648/19 3687/6 3694/22 TikTok's [1] 3651/16
submit [1] 3744/17
3721/21 3648/20 3695/2 3695/4 3714/3 Tim [1] 3660/11
subscribe [1] 3694/2
surfacing [1] 3722/3 Tang [4] 3702/9 3719/24 3722/12 times [5] 3635/19
subscribed [1] 3736/18
surprisingly [1] 3709/3 3707/8 3707/21 3735/7 3689/18 3726/14
subscriber [1] 3701/2
survey [2] 3685/5 3708/7 terms [13] 3643/1 3733/23 3737/11
subscribers [4]
3686/6 tap [1] 3674/9 3643/17 3644/23 timing [2] 3743/8
3676/24 3676/25
timing... - users 3767

T trade-off [3] 3679/16 3685/11 3697/22 understood [2] UPX715 [2] 3633/20
3680/8 3717/5 3701/15 3703/17 3639/23 3710/12 3746/11
timing... [1] 3744/21
trade-offs [4] 3693/24 3719/17 3723/16 unease [3] 3678/23 UPX93 [3] 3706/21
Tinter [2] 3745/16
3694/6 3718/15 3730/24 3744/23 3679/7 3681/13 3710/7 3714/12
3746/7
3730/8 turnaround [1] 3745/3 unending [1] 3722/1 UPX940 [3] 3633/19
Tinter's [1] 3746/3
traffic [9] 3646/19 turned [1] 3735/25 unexpected [1] 3679/7 3682/2 3683/11
tire [1] 3699/25
3646/20 3650/8 turning [1] 3723/13 unfair [1] 3643/11 URL [5] 3711/15
title [4] 3667/6 3708/16
3693/22 3720/15 TV [1] 3699/10 unfortunately [1] 3711/15 3712/5
3708/18 3719/22
3731/15 3731/18 two [21] 3638/7 3648/22 3712/17 3731/22
today [22] 3641/2
3731/24 3732/4 3646/10 3654/10 unidentified [1] usage [6] 3634/15
3642/13 3643/20
trails [1] 3654/12 3656/20 3656/21 3745/22 3689/24 3694/6
3644/5 3644/7 3645/5
trained [1] 3697/9 3670/1 3671/12 unintended [1] 3712/7 3712/9 3735/5
3648/16 3650/11
transcript [2] 3631/10 3671/16 3672/7 3715/19 use [27] 3635/18
3650/15 3654/14
3747/4 3672/20 3677/9 unique [4] 3650/1 3676/19 3677/14
3654/23 3663/10
transcripts [2] 3744/12 3699/1 3704/9 3661/25 3695/15 3678/24 3679/6
3663/20 3671/16
3745/12 3704/16 3704/18 3734/12 3680/22 3683/17
3697/7 3708/3
transferring [1] 3706/22 3710/17 unit [1] 3703/12 3685/6 3685/6 3686/1
3711/17 3715/3
3641/10 3717/13 3717/19 UNITED [7] 3631/1 3686/8 3686/15
3719/24 3720/6
transfers [1] 3661/17 3726/14 3742/15 3631/3 3631/11 3689/18 3689/25
3722/17 3739/7
transformed [1] 3672/2 two-ish [1] 3671/12 3703/8 3704/3 3692/23 3694/18
today's [1] 3644/5
transitional [1] 3669/19 two-sided [1] 3699/1 3715/12 3730/17 3696/5 3696/24
together [1] 3649/2
transitioning [1] 3669/8 Tyler [3] 3631/23 University [1] 3666/23 3711/14 3712/2
told [2] 3637/11
travel [8] 3648/23 3665/13 3665/19 unlaunchable [1] 3720/25 3725/10
3727/13
3661/15 3669/3 type [12] 3646/15 3721/14 3725/10 3725/11
tomorrow [3] 3743/16
3688/15 3688/17 3653/23 3672/22 unless [2] 3681/14 3726/4 3733/7 3735/6
3743/21 3744/9
3688/19 3688/19 3687/17 3693/4 3691/3 used [18] 3657/2
tone [1] 3729/7
3688/21 3695/10 3696/19 unlike [2] 3698/24 3680/20 3682/18
tonight [1] 3744/16
travels [1] 3665/6 3711/17 3712/5 3700/16 3683/8 3685/17
took [7] 3639/18
treading [1] 3660/7 3716/8 3723/25 unlock [1] 3676/20 3686/6 3688/13
3667/18 3671/2
TRIAL [1] 3631/10 3731/21 unpleasant [1] 3716/17 3689/19 3690/2
3671/16 3687/21
trials [1] 3719/14 type-in [1] 3695/10 unseal [1] 3745/12 3690/3 3697/13
3726/3 3726/11
tricky [4] 3680/8 typed [1] 3695/14 unsurprisingly [1] 3705/8 3711/16
tools [1] 3733/9
3691/20 3699/22 types [3] 3653/16 3735/24 3712/12 3719/24
top [13] 3671/18
3735/5 3654/2 3736/22 unthinkable [1] 3733/2 3733/13
3672/4 3701/10
tried [8] 3647/7 typically [2] 3703/3 3698/17 3739/14
3703/22 3704/2
3655/18 3676/20 3708/5 unusual [1] 3661/22 useful [6] 3671/25
3709/20 3714/12
3688/24 3700/23 typing [8] 3693/3 up [42] 3634/7 3678/5 3701/11
3716/2 3716/7 3716/9
3718/14 3720/1 3693/5 3693/8 3642/13 3643/4 3722/4 3723/1
3716/9 3739/24
3730/8 3693/12 3711/14 3643/9 3643/22 3737/13
3745/21
trouble [1] 3720/22 3711/16 3720/7 3647/10 3652/19 user [29] 3641/13
topic [6] 3655/24
true [9] 3678/20 3731/18 3653/20 3664/9 3655/5 3674/20
3686/23 3686/25
3683/1 3695/12 3666/21 3674/1 3677/21 3679/19
3723/16 3723/20
3695/13 3700/7 U 3674/24 3676/5 3680/5 3681/1 3681/4
3727/12
3700/8 3724/12 U.S [7] 3631/14 3676/16 3684/23 3685/15 3687/25
topics [3] 3681/9
3724/13 3747/4 3631/17 3631/19 3689/19 3692/6 3694/14 3696/16
3689/4 3708/4
truly [1] 3674/11 3632/24 3693/25 3692/24 3695/12 3700/5 3700/8 3701/1
toto [1] 3682/15
trumps [1] 3709/3 3704/7 3704/13 3700/15 3702/20 3701/2 3703/22
touch [1] 3672/18
trustees [1] 3667/1 Udi [4] 3672/14 3703/13 3705/1 3708/25 3713/18
touched [1] 3672/21
truth [3] 3666/6 3666/6 3672/15 3672/16 3710/17 3711/2 3722/4 3733/20
towards [1] 3687/16
3666/7 3672/20 3711/3 3711/4 3711/8 3734/16 3734/17
trace [1] 3667/10
try [17] 3635/18 Uh [6] 3703/21 3716/17 3717/6 3735/6 3735/6
traced [1] 3726/10
3637/20 3640/19 3719/21 3731/6 3717/8 3718/2 3737/13 3741/14
tracing [1] 3712/16
3648/5 3659/12 3731/10 3731/13 3720/10 3720/16 3742/4 3742/24
track [2] 3673/24
3677/2 3687/12 3740/25 3720/18 3721/13 user's [2] 3655/6
3743/22
3734/8 3734/17 UK [2] 3672/13 3733/9 3734/15 3680/10
tracked [2] 3678/25
3735/19 3736/4 3693/25 3735/3 3735/3 user-focused [1]
3681/13
3741/12 3741/15 ultimate [1] 3700/21 3741/23 3744/3 3687/25
tracking [6] 3634/15
3741/18 3742/4 ultimately [4] 3676/3 updating [1] 3702/23 users [72] 3634/12
3722/12 3722/15
3742/25 3743/2 3691/24 3699/15 upon [1] 3653/2 3634/19 3634/20
3723/3 3723/15
trying [9] 3639/1 3699/18 upside [1] 3640/9 3635/24 3636/2
3725/11
3649/11 3653/19 unbounded [1] upward [1] 3707/20 3646/13 3655/14
tracking-free [1]
3656/12 3659/8 3695/12 UPX [1] 3688/7 3659/13 3660/11
3725/11
3662/5 3685/1 under [4] 3655/4 UPX123 [5] 3706/21 3664/3 3671/13
trackpads [1] 3710/25
3696/16 3736/16 3683/25 3744/5 3708/13 3708/18 3671/25 3673/1
trade [7] 3679/16
turn [13] 3636/18 3744/12 3709/15 3730/20 3673/7 3673/16
3680/8 3693/24
3670/8 3670/15 underlying [3] 3641/20 UPX475 [3] 3697/24 3673/24 3674/2
3694/6 3717/5
3682/1 3685/2 3700/2 3720/19 3701/16 3705/18 3674/3 3674/12
3718/15 3730/8
users... - zero 3768

U versus [4] 3646/22 3700/20 3708/20 3720/20 3726/16 3676/14 3677/6

3704/21 3705/14 Webb [1] 3631/23 3728/11 3679/16 3700/7
users... [53] 3674/15
3730/5 week [2] 3635/8 work [10] 3669/22 3700/9 3700/22
3676/3 3677/3 3680/9
vertical [27] 3646/4 3685/16 3673/15 3673/24 3736/5
3681/7 3685/9 3686/1
3646/9 3646/14 week's [3] 3744/25 3674/1 3698/7 yearly [1] 3676/16
3688/23 3688/25
3646/18 3646/22 3745/7 3745/10 3715/15 3726/19 years [20] 3634/23
3689/18 3689/21
3648/11 3648/16 weeks [3] 3702/25 3733/17 3736/9 3644/14 3644/17
3690/7 3693/25
3648/19 3648/20 3726/14 3740/12 3739/17 3646/12 3646/24
3694/5 3694/17
3648/21 3649/2 weighing [1] 3700/25 worked [13] 3660/18 3650/7 3667/18
3694/18 3695/2
3649/7 3649/16 Weinberg [1] 3744/11 3667/4 3667/7 3669/11 3669/14
3696/7 3699/2
3649/17 3649/24 well-funded [1] 3667/13 3669/6 3671/12 3671/16
3699/21 3699/25
3650/2 3650/4 3650/5 3724/18 3670/1 3671/8 3671/8 3672/5 3672/20
3704/23 3709/5
3650/7 3650/11 what's [9] 3650/24 3672/1 3672/17 3691/13 3695/20
3712/11 3717/10
3661/12 3661/14 3665/21 3667/20 3672/19 3691/12 3699/21 3700/10
3725/1 3726/4 3726/9
3661/16 3661/17 3675/12 3684/1 3702/23 3710/20 3711/12
3726/12 3731/20
3661/19 3723/18 3703/2 3707/11 working [5] 3638/7 3733/9
3731/23 3732/5
3725/17 3743/8 3744/22 3672/4 3677/16 Yelp [5] 3726/17
3733/3 3734/8
verticals [1] 3688/20 whenever [4] 3688/16 3677/18 3724/21 3726/18 3726/25
3734/12 3734/22
via [1] 3698/14 3711/9 3726/8 works [3] 3667/20 3727/3 3727/5
3735/10 3735/22
viable [2] 3701/7 3741/10 3745/9 3745/10 yep [10] 3636/11
3735/23 3735/24
3717/14 whereas [2] 3663/17 world [8] 3677/8 3646/2 3647/19
3736/15 3736/21
vice [2] 3667/23 3722/8 3681/14 3701/3 3657/6 3657/21
3736/24 3737/2
3728/13 whereby [1] 3641/6 3705/10 3705/11 3658/18 3707/2
3737/3 3737/7
view [2] 3738/15 white [2] 3641/9 3712/6 3722/23 3707/5 3715/25
3737/14 3738/8
3738/23 3730/17 3736/3 3719/23
3739/11 3740/11
views [4] 3638/20 who's [1] 3646/16 worries [1] 3694/20 yesterday [1] 3680/1
3742/1 3742/15
3645/2 3718/6 3718/9 whole [4] 3665/23 worry [2] 3679/1 York [3] 3631/25
3742/23
virtue [2] 3687/25 3666/6 3700/20 3679/13 3632/12 3737/11
using [10] 3655/14
3742/15 3705/1 worry-free [2] 3679/1 YouTube [2] 3668/5
3670/6 3696/11
vision [1] 3674/11 Wilberforce [2] 3679/13 3703/6
3701/10 3709/4
visual [1] 3688/18 3665/13 3665/19 worse [2] 3718/2
3714/7 3720/1 3722/5
visualize [1] 3698/10 WILLIAM [1] 3631/23 3718/2 Z
3733/9 3739/20
Vivek [2] 3671/8 Williams [1] 3632/7 write [2] 3635/15 zero [4] 3663/25
usual [1] 3726/24
3674/23 willing [5] 3661/1 3715/9 3735/12 3735/15
utilize [1] 3725/19
3661/3 3664/6 3687/4 writes [2] 3707/21 3735/20
V W 3716/15 3708/7
waiting [1] 3744/15 willingness [1] 3681/20 written [2] 3684/5
vague [1] 3678/23
walk [1] 3671/2 WilmerHale [2] 3724/8
Valley [3] 3669/18
wants [1] 3734/17 3665/13 3665/18 wrong [4] 3664/9
3672/10 3732/13
Washington [5] 3631/5 Windows [7] 3643/20 3666/2 3716/12
valuable [1] 3684/19
3631/15 3631/18 3657/17 3658/1 3745/23
valuation [1] 3701/4
3632/8 3632/25 3658/16 3658/20 wrote [2] 3715/8
value [7] 3685/1
watched [1] 3678/24 3659/15 3664/21 3719/10
3685/1 3685/3 3685/9
wave [1] 3698/9 winners [1] 3712/19
3708/17 3711/8 X
way [25] 3643/11 wish [1] 3644/22
3716/25
3648/25 3659/4 within [6] 3644/1 X744 [1] 3703/19
value-free [1] 3685/1
3663/21 3669/23 3653/16 3662/23 xyz.com [1] 3673/9
varied [1] 3720/20
3671/12 3671/19 3703/12 3726/16
variety [4] 3648/19
3677/20 3677/24 3739/7 Y
3688/13 3713/16
3680/1 3687/13 without [8] 3646/8 Yahoo [26] 3641/3
3726/1
3688/4 3689/16 3693/15 3720/12 3641/6 3641/8
various [5] 3662/15
3689/20 3690/13 3721/12 3735/3 3641/10 3641/12
3676/12 3698/12
3691/14 3692/21 3735/3 3738/18 3641/14 3641/21
3703/5 3736/22
3694/17 3701/17 3739/8 3641/21 3641/24
variously [1] 3708/3
3710/1 3712/1 witness [8] 3633/2 3642/14 3642/21
vast [1] 3670/18
3712/15 3712/16 3640/21 3665/8 3643/9 3643/13
vastly [1] 3705/13
3731/24 3743/2 3702/7 3705/21 3643/23 3643/24
velocity [1] 3699/5
ways [9] 3674/10 3714/22 3740/24 3643/25 3644/2
venture [8] 3669/16
3678/13 3696/13 3745/21 3644/4 3644/5 3645/4
3669/18 3672/10
3697/16 3704/23 witness' [2] 3682/20 3645/5 3691/5 3705/9
3672/12 3674/18
3718/19 3719/15 3683/7 3731/3 3731/11
3674/21 3676/8
3725/7 3725/8 Wojcicki [1] 3668/4 3732/3
3732/13
weapon [1] 3709/24 word [3] 3636/5 Yandex [1] 3728/8
Ventures [2] 3672/12
weather [2] 3670/10 3677/14 3738/10 year [16] 3635/23
3732/9
3726/1 words [8] 3644/3 3636/10 3647/8
versa [1] 3728/13
web [5] 3641/22 3678/15 3678/24 3668/5 3672/1
version [2] 3677/2
3661/20 3684/4 3697/5 3713/5 3674/17 3674/22
3703/19
3675/1 3675/3