Anne Jackley

AI will aid and empower authors and editors, not replace them. This author captures some of the nuances of what we are trying to build at #Spines well

6

Such an insightful conversation highlighting some of the challenges and opportunities for our industry; breaking down boundaries and evolving to balance the needs of best practice and author experience. Looking forward to the next conversations in this series!

6

Today I mailed the latest issue of the Scalene newsletter to subscribers, examining the changing landscape of peer review in the age of AI. This week we’re looking at the Black Spatula Project, reviewing articles via voice, and moving the review process upstream to the point of authoring. Take a look here (link in comments) and sign up for next week’s update.

16

2 Comments

Three Critical Mistakes with User Stories https://rb.gy/6x1rk2 Ever found yourself scratching your head over a user story that seemed clear at first but left your development team confused? Picture this: you're on an agile team working on a new feature for an ecommerce platform's product recommendation engine. The user story reads, "As a user, I want personalized product recommendations." Sounds straightforward, right? This is where blindly relying on user stories can fall short. Let's look at three critical mistakes often made when using user stories in software development.

1

1 Comment

Fellow writers, what's your writing process? For me, I indulge in the pre-writing stage: I need time—hours! Days! Weeks if you'll let me!—to talk through my ideas. These discussions take place in my own head (usually whilst in the shower or driving or trying to fall asleep at night) or with colleagues (or anyone who will listen!). Now I use #AI to bounce off ideas and to help me untangle my narrative knots... LISTEN. I'm fiercely protective of my voice and how I express my ideas, so ultimately I'm always in control. But I'd be lyin' if I claimed AI doesn't make me a stronger writer, producer, and, frankly, idea generator. Tell me how you're using AI to enhance your #creativeprocess!

13

1 Comment

“The dramatic increase in the amount of accessibility testing performed in automation is not correlating to a proportional increase in the accessibility of the digital world.” ~ Noah Mashni, CPACC & Mark Steadman #TrueStory, unfortunately. Automating the discovery of bugs that never get solved (essentially infinite TTR*) results in an ever-growing mountain of accessibility bugs and little benefit for end users. This is why Assistiv Labs takes the approach with our end-to-end accessibility testing service of understanding our customers’ digital delivery processes and working closely with them to properly assign and prioritize our bugs to make them more likely to get resolved. For example, we meet regularly with the accessibility team at Asana to ensure our bug reporting fits into their existing processes so issues don’t land in product teams’ backlogs in a way that’s disruptive to the ways they usually work. The fact these bug reports are automated rather than on the cadence of manual audits means the responsible teams find out about them before attention decay** sets in. Our E2E service delivers audit-quality results at the speed of automation. That’s a bold claim, but one we stand by. *Time To Resolution **Our term for the tendency of developers to deprioritize work where they’re no longer in that context. https://lnkd.in/gTJN5ybs

32

1 Comment

Interesting article!

5

3 Comments

Mike Harris, Digital Camera World's How To Editor, writes,"AI-generated imagery poses a threat that isn't being talked about enough – and I can't find an answer." How will AI affect our perception of imagery in the long term? https://ow.ly/azkb50UFNuk

6

"'Theoretically,' they said, 'theoretically!' That's what they said!" - Carlotta Shetana RVSM - Persona-In-Chief for Public Relations at BAD TRIP RECORDS. In Re: Rights in dataComet source code: No one anywhere in the world has a legitimate interest in preventing me from immigrating to Solothurn CH to serve as a contributing employee of DATABEAST AG, inasmuch as I have but three potential bars to sharing my code, and in particular the code in dataComet? All such covenants were made under my birthname, Kevin Eric Saunders: 1. Source code incorporated into dataComet under Cornell's purchase of the MIT PC/IP code from its author for $50,000 in 1984, comprising routines for providing doubly-linked lists used for managing timers in COMET, easily remedied if felt to be an issue? - EASY - [1a. Cornell's enduring rights to source code in my work for Cornell, COMET, as provided them on May 1 1994, have been superseded under the terms in Cornell Research Foundation D-1716 as "modifications to this code by DATABEAST INC."] - *STRICTLY* THEORETICAL! - 2. A Non-Disclosure Agreement covering the source code for AT&T UNIX System 5, provided as part of a seminar paid for by management at Cornell ca. 1988? - NADA - runrun(); // OH YEAH BABY, SPIN IT ON! 3. A Noncompete Agreement signed in 1994 as part of my employment contract as a Senior Software Engineer working on the Michigan Access Project for Millennium Computer Corporation? - NADA - All other source code from third parties incorporated in the codebase for dataComet and dataComet-Secure was made publicly available _gratis_ under BSD "Free-To-Use-With-Credit" licensing statements duly included in dataComet's "About dataComet" text! - Bonze Annette Rose Blayk, President -- DATABEAST INC / BAD TRIP RECORDS __________ HERE FOLLOWS INSANITY MADE MANIFEST, FOR UNDER THIS RULE NO RATIONAL INVESTOR WILL EVER AGAIN INVEST IN R&D IN THE UNITED STATES! FILE UNDER: "The Madness of Smokin' Joe Biden - 'He's from Scranton!'" "Theoretically, noncompetes are meant to stop primarily high-level employees from jumping ship to other companies, bringing proprietary information and other knowledge with them."

2

4 Comments

If trad publishers want to survive and thrive going forward, they have to learn how to not only co-exist with but also BENEFIT from AI and all related information technologies. https://lnkd.in/e8BDYqam

1

I agree 100%. ChatGPT + Claude is a winning combo. After an intensive AI project this week, I realized that Claude 3.5 Sonnet V2 provides high-quality text output (more nuance, more depth, less robotic). However, it runs out of tokens really fast. Even on a paid account, I kept receiving warnings to start a new chat because continuing with the same chat for too long can make you reach the usage limit faster. This is ironic because Claude does not store the information you provide in its memory—you have to provide the context each time you start a new chat. This is where ChatGPT comes in. I've never managed to hit the limits on a paid ChatGPT account. Moreover, it retains every crucial piece of information from all the projects I am working on and the preferences I share in its memory. It's essentially become an editorial assistant for me after consistent use over the past year. The best part is that it can review very large datasets and perform useful analyses for me. That said, the data visualizations I can create through Claude Artifacts are undoubtedly superior. ChatGPT cannot provide the visualizations directly; it can only generate the code to run. So, at this point, it’s really a tie!

3

1 Comment

"While anthropomorphizing AI — the common practice — is better than having no human element at all, there is a more effective approach: Companies could see a meaningful impact in the acceptance of their AI tools if they proactively communicate the human input in their development." By Oguz A. Acar and Mirjam Tuk https://lnkd.in/e59MMRi7

21

3 Comments

I’m seeing a worrying trend of archival conferences slowly transitioning into in-person only. At first they were digital, then they were hybrid, and then digital accommodations for presenters under extenuating circumstances. This was problematic enough as digital audiences began to be eliminated. Many of whom finally felt they were able to participate for the first time. Presenters also needed to disclose why they needed digital accommodations. Something that may make presenters quite uncomfortable to disclose. Me 10 years ago certainly wouldn’t have been able to. There are certainly exceptions, but I’ve noticed that we are getting to a stage where some conferences are eliminating language that even suggests digital accommodations in their callouts. Conferences heavily focus on networking and hosting in-person only eliminates those who may benefit from that networking most. Travel can be difficult for many of us. For myself, it’s health related. I also think of those who need to care for small children or sick relatives, students/early career professionals/those in the global south who cannot afford to travel because few stipends exist, and those who work in very small non-profits that are not awarded the time to be away. This is particularly troubling when so many conferences are themed around underrepresentation. I understand the difficulties that go into organizing a conference. I’ve been responsible for organizing events, including conferences in the past. I was organizing guest lectures during the pandemic. There are contractual obligations to sponsors and host sites but we need to be doing more. So, think about why certain demographics are more prevalent in our field. By returning to in-person only, we’ll always be stuck at the stage of “we need to be a more inclusive field.” Those voices that are necessary for such inclusion initiatives will never get to participate in the “here’s how”.

34

7 Comments

#Specsmanship of #AI #Processors is what everyone is talking about today. It's important and other pubs and conferences are doing great job! The Ojo-Yoshida Report, however, approaches AI differently. If you are a system designer, #whatstheplan for your brand new AI products? How long will they last? Who will bear the burden of updating them? #lifecycle #endoflife Bryan Reimer and Philip Koopman dive into the pragramatic discussion on the social, economic and technical impact of #AI products -- be it an #edgeAI device or #SDV. #BOSS Register free to watch the video session on demand: https://lnkd.in/eqs3Dr4C

10

1 Comment

from the letter: "Publishers who wish to avoid having their content summarized in the form of AI-generated search results can only do so if they opt out of being indexed for search completely, which would result in a materially significant drop in referral traffic." (This will be one of the topics covered at the upcoming IETF "Workshop on AI-Control" -- can robots.txt and similar standards be extended to cover content permissions by purpose)

4

12 Comments

OpenAI just launched its new model, GPT-4o. It's faster and more real-time, but it also incorporates vision and audio. What's amazing is what it can potentially do to solve accessibility challenges at cultural institutions. The translation tools act more like a translator does, summarizing to encourage brevity. Descriptive audio tours will be a thing of the past; instead, just train an AI model on how to "see" your site.

30

1 Comment