William Isaac
williamis.bsky.social
William Isaac
@williamis.bsky.social
Research Scientist @DeepMind | Previously @OSFellows & @hrdag. RT != endorsements. Opinions Mine. Pronouns: he/him

Reposted by William Isaac
A ton of attention over the years goes to plots comparing open to closed models.
The real trend that matters for AI impacts on society is the gap between closed frontier models and local consumer models.
Local models passing major milestones will have major repercussions.
buff.ly/ccMJydQ
October 4, 2025 at 6:40 PM
Reposted by William Isaac
Claude, "We all know among Sauron's many evils was that he ran Mordor using an Excel spreadsheet with multiple tabs. Show me the spreadsheet"

It made 12 tabs "so bureaucratically complex that even the Eye of Sauron would need reading glasses to review it." Some very funny stuff. Creative, even.
September 15, 2025 at 3:48 AM
Reposted by William Isaac
Gemini Live is giving me the confidence to speak my second language

The anxiety I experience when trying to speak my second language disappears when I use Gemini Live.

#ai #gemini #geminiai
Gemini Live is giving me the confidence to speak my second language
The anxiety I experience when trying to speak my second language disappears when I use Gemini Live.
www.androidauthority.com
September 14, 2025 at 1:55 PM
Reposted by William Isaac
This is a cool paper that suggests that AI agents can indeed be used for social science experiments, but that just using a chatbot isn't good enough, instead prompts developed based on social & game theory makes AI agent actions predictive of real human outcomes. benjaminmanning.io/files/optimi...
September 4, 2025 at 2:13 AM
Reposted by William Isaac
Duolingo’s recent earnings shows the gap between online vocal AI critics and businesses real-world value.

Despite the backlash, Duolingo’s revenue and daily users jumped 40%.

The AI hype is wild, but so is the value people are getting. This isn’t blockchain, it’s more like mobile or the internet
August 8, 2025 at 1:35 PM
Reposted by William Isaac
"In collaboration with the @archive.org TV News Archive we examined 14,749 evening news broadcasts spanning the big three networks from July 2010 to present, asking Gemini 2.5 Flash Thinking to make an index of the stories covered in each broadcast, geography, sentiment, frame, narrative structure.
July 31, 2025 at 11:43 PM
Reposted by William Isaac
We may have the chance to hire an outstanding researcher 3+ years post PhD to join Tarleton Gillespie, Mary Gray and me in Cambridge MA bringing critical sociotechnical perspectives to bear on new technologies.

jobs.careers.microsoft.com/global/en/jo...
Search Jobs | Microsoft Careers
https://jobs.careers.microsoft.com/global/en/job/1849026/Principal-Researcher-–-Sociotechnical-Systems-–-Microsoft-Research
July 28, 2025 at 5:26 PM
🤯
Google is processing 980 trillion+ monthly tokens across our products and APIs (up from 480T in May) 🤯

wen googolplex? 🤩❤️‍🔥
July 25, 2025 at 4:22 AM
Reposted by William Isaac
🥇 I am so happy to see the Google DeepMind Gemini team **OFFICIALLY** achieved gold-medal performance.

Not only did we get 5 out of 6 questions correct. What's more amazing is the official IMO graders find those answers to be clear, precise and easy to follow! 1/

deepmind.google/discover/blo...
Advanced version of Gemini with Deep Think officially achieves gold-medal standard at the International Mathematical Olympiad
Our advanced model officially achieved a gold-medal level performance on problems from the International Mathematical Olympiad (IMO), the world’s most prestigious competition for young...
deepmind.google
July 21, 2025 at 10:24 PM
Reposted by William Isaac
"[video game] as a community theater production" may be one of the most delightful Veo 3 Fast prompts

Please enjoy, in order: GTA, Pokemon, Mario Kart, The Witcher 3, Stardew Valley, Tetris, Mortal Kombat, The Sims, & Death Stranding.

The whole prompt was the one above. And the glitches add to it
July 19, 2025 at 3:08 AM
Reposted by William Isaac
If you're at @icmlconf.bsky.social this week, come check out our poster on "Position: Evaluating Generative AI Systems Is a Social Science Measurement Challenge" presented by the amazing @afedercooper.bsky.social from 11:30am--1:30pm PDT on Weds!!! icml.cc/virtual/2025...
ICML Poster Position: Evaluating Generative AI Systems Is a Social Science Measurement ChallengeICML 2025
icml.cc
July 15, 2025 at 6:35 PM
Reposted by William Isaac
The #FAccT and #ICML tags are good places to find some brilliant folks
July 13, 2025 at 5:31 PM
Reposted by William Isaac
The best article ive read so far on social impacts: henryfarrell.net/large-ai-mod...
Large AI models are cultural and social technologies – Henry Farrell
henryfarrell.net
July 13, 2025 at 11:36 PM
Reposted by William Isaac
Did a starter pack a while back too of scholars and practitioners who cover this angle in case that’s useful go.bsky.app/5sFqVNS
July 14, 2025 at 7:18 AM
Reposted by William Isaac
If it's helpful, I followed the recommendations in this post from @nsaphra.bsky.social and as a result have a much more interesting and ML-focused feed: nsaphra.net/post/bsky/
The AI Researcher's Guide to a Non-Boring Bluesky Feed | Naomi Saphra
How to migrate to bsky without a boring feed.
nsaphra.net
July 14, 2025 at 8:20 AM
Reposted by William Isaac
These are the ML feeds I've pinned that now get mixed into my main feed:

- ML Feed: Trending bsky.app/profile/smcg...
- MLSky: bsky.app/profile/alex...
- GenAI Bluesky Network: bsky.app/profile/hell...
- Nuanced AI Commentary: bsky.app/profile/dame...
July 14, 2025 at 8:20 AM
Reposted by William Isaac
I can recommend following: @simonwillison.net, @metr.org, @abeba.bsky.social, @mmitchell.bsky.social , @wang.social, @sashamtl.bsky.social, @justinhendrix.bsky.social, @bvlsingler.bsky.social

Some of the people I follow to keep up with that side of things. :)
July 13, 2025 at 9:44 AM
I would love to see more discussion on ML/AI on bsky. Particularly more nuanced discussions of the societal impacts. If you tag in the relevant post, I'll be sure to repost it!
July 13, 2025 at 9:17 AM
Reposted by William Isaac
Still not a lot of ML talk on bsky (at least in my feed), hence paper Sunday: my two most interesting recent reads
- H Nets arxiv.org/abs/2507.07955
- Energy Based Transformers arxiv.org/abs/2507.02092
Dynamic Chunking for End-to-End Hierarchical Sequence Modeling
Despite incredible progress in language models (LMs) in recent years, largely resulting from moving away from specialized models designed for specific tasks to general models based on powerful archite...
arxiv.org
July 13, 2025 at 6:14 AM
Reposted by William Isaac
I’ve loved everything about the Red Lobster turnaround:

• A young Nigerian immigrant CEO who previously ran PF Chang (which I love)
• Listened to customers and made swift changes
• Stars in ads, bringing real authenticity to the “we’re fixing it” vibe
From a popular CEO to viral seafood boils: The story behind Red Lobster’s biz comeback
New offerings and marketing initiatives spark the ‘polish casual’ restaurant’s re-emergence after years of sinking fortunes.
www.campaignlive.com
July 10, 2025 at 6:28 PM
Reposted by William Isaac
Reasons you should donate to EFF this month:
1. We're turning 35! Wish us a happy birthday 🥳
2. We've got cool new swag--our gift to you 👕
3. We're advocating for your rights to privacy and free speech online 👩‍💻
35 Years for Privacy, Free Speech, and a Brighter Future
Through July 10, new monthly or annual Sustaining Donors get an EFF35 Challenge Coin! With your help, EFF is here to stay.
eff.org
July 3, 2025 at 7:00 PM
Reposted by William Isaac
Anyone at #facct2025 want to go to the acropolis this afternoon?
June 26, 2025 at 8:20 AM
Reposted by William Isaac
Claude Code, meet your rival: Gemini CLI 🔥🚀

Google just unleashed Gemini 2.5 Pro at the command line—free, open-source (Apache 2.0) , 1-million-token brain baked right into your terminal. Chat, code, search, script—all with zero tab fatigue.
June 26, 2025 at 1:21 AM
Reposted by William Isaac
The real Bear Grylls
June 26, 2025 at 1:44 AM
Reposted by William Isaac
The two most common questions I get asked about AI are “which AI should I use” and “how do I start using AI?”

I wrote a short guide attempting to answer both questions. www.oneusefulthing.org/p/using-ai-r...
Using AI Right Now: A Quick Guide
Which AIs to use, and how to use them
www.oneusefulthing.org
June 23, 2025 at 4:27 PM