Search Hacker News by keywords, user, score, etc.
Subscribe to new search results by RSS or JSON

so, because science as whole is not pursuing the idea that people with different genetics as a population are inferior in some ways to others with sufficient vigor, that we should expect a justifiable general distrust of science including completely unrelated results like global warming. I don't see how this is prescriptive in any way, except maybe to ... I guess find scientists that are will to accepting funding for ideas that are popular with some people? do you think that would help if they found those ideas to be meritless? or even if they didn't?

I would be very surprised if they can scale hiring contractors to reliably renovate buildings.

I'd rather work for an AI than some of the managers I've had in the past.

> For instance, from AirBnb's terms of service: "Do not use bots, crawlers, scrapers, or other automated means to access or collect data or other content from or otherwise interact with the Airbnb Platform."

> There is no similar prohibition against using screen readers.

A screen reader uses automated means to access or collect data or other content from or otherwise interact with a platform.

This thread is very confusing. Everyone is saying diametrically opposed things. But I think this may be a clue: AWS bedrock means api billing, no? I’m guessing those complaining about the recently lowered quality of Claude are on subscriptions. And those who are still loving Claude are on work accounts.

So are we still going to be free to be creative while AI does the menial jobs?

2001 Space Odyssey presents a different scenario

Perplexity said:

> If replacement cargoes are coming from the U.S. East Coast, typical sea-freight transit to North Europe is about 15 days and to South Europe about 18 days. For longer-haul routes from East Asia toward Europe, a typical voyage is about 30 days, and some general Europe-bound ocean freight can take 30 to 45 days depending on route and congestion.

yes. this is current issue. there are two solutions:

1. the reason it's slow as you select more series over longer periods of time is that the series has to be pulled for each time bucket in the range, and then the samples have to be pulled for each bucket. By compacting older buckets and merging samples together, historical queries should be pretty comparable to 'more recent' cold queries. 2. We don't pre-cache all the metadata today. If we did that, then we could parallelize sample loads much more efficiently, lowering latency. 3. There is a lot of room to do better batching and tune the parallelism of cold reads.

We've only been at this for a couple of months. THe techniques to improve latency on object storage are well known, we just have to implement them.

Another benefit is this: all the data is on S3, so spinning up more optimized readers to transform older data to do more detailed analysis is also an option with this architecture.

Not sure why you're being downvoted, I guess it's because how your reply is worded. Anyway, Qwen3.7 35B-A3B should have intelligence on par with a 10.25B parameter model so yes Qwen3.5 27B is going to outperform it still in terms of quality of output, especially for long horizon tasks.

Oh boy, the gap between the average it professional and ai pros here is already staggering, let alone the rest of the world. I feel like an alien, no matter where.

It's been down intermittently all day. In some ways, that's worse than being completely down.

Seems to be fine if you're using another appview like blacksky's one, or directly querying PDSs like red dwarf does. Social media isn't very useful if most other people can't use it though.

a great assessment thanks I hadn't seen that

Not really, offside is Abseits.

In this case Aus means out like in spit it out or out with it, "raus damit".

Not exactly. I got (and renewed) the permit with zero knowledge of any official language. However, my wife had to present the basic certificate or my promise that she would learn the language.

There's a systematic marketing campaign from oai on reddit and HN - there's a huge uptick of "codex is better than claude code" comments and posts this last week which is perfectly timed with the claude code increased limits

Okay let me elaborate how I envision that attack to work:

1. attacker wants to use your yubikey-backed ssh key, let's say for running ssh-copy-id once with their own key so they can gain access to your server

2. thus they need to trick you into touching the key when they run that command

3. the best way to trick you is to wait until you do something where you'd normally need to touch that key yourself

4. so they alias ssh to a script that detects when you're trying to connect to this server yourself, and invoke ssh-copy-id instead and spit out a reasonable looking error (something that makes you think "bloody DNS, it's always DNS, innit")

5. then they undo the alias so you succeed on the next try and suspect nothing

I am curious to see more tests on the reading path. The article mentions matching 500 series of 6h window with 1m step - and it takes 2s for warmed caches. That doesn't sound good at all.

Especially nowadays, when metrics from k8s ramping up churn rate to hundreds of thousands and millions series.

Yeah, that's a fair assesment. We're working on improvements to cold query latencies but there's a lot of potential to optimize historical queries with heavy parallelism on read replicas. Something similar to mini-spark jobs (since all the files are on S3 anyway).

maybe UN resolution?

>background computer use

How does that even work technically? macOS doesn't support multiple cursors. On native Cocoa apps you can pass input to a window without raising via command+click so possibly they synthesized those events, but fewer and fewer apps support that these days. And AppleScript is basically dead, so they can't be using that either.

I also read they acquired the Sky team (who I think were former Apple employees). No wonder they were able to pull of something so slick.

> Also, black Americans don't call themselves "African-American" unless they were raised in a white environment. Never have.

I'm guessing you either don't remember or weren't alive in the 1990's. It was a whole grassroots movement and pretending it didn't exist is extremely insensitive, to put it mildly.

https://en.wikipedia.org/wiki/African_Americans#Terminology

That's helpful data, thank you. Sounds like it may depend on the service. (I'm genuinely shocked to see that many hotmail addresses, and can't help but wonder if there are correlations with other factors.)

Has anyone figured out how to stop the Codex app from draining my M5 Pro's battery in like 2 hours? I can literally just have it open and my lap turns into a heater. I've tried adjusting all sorts of settings and haven't been able to make a dent. I'm assuming its the garbage renderer.

show us the benchmarks with "adaptive thinking" turned on

Early benchmark results on our private complex reasoning suite: https://gertlabs.com/?mode=agentic_coding

Opus 4.7 is more strategic, more intelligent, and has a higher intelligence floor than 4.6 or 4.5. It's roughly tied with GPT 5.4 as the frontier model for one-shot coding reasoning, and in agentic sessions with tools, it IS the best, as advertised (slightly edging out Opus 4.5, not a typo).

We're still running more evals, and it will take a few days to get enough decision making (non-coding) simulations to finalize leaderboard positions, but I don't expect much movement on the coding sections of the leaderboard at this point.

Even Anthropic's own model card shows context handling regressions -- we're still working on adding a context-specific visualization and benchmark to the suite to give you the objective numbers there.