Who's Who of AI

Thomas Dietterich

Safe and robust AI/ML, former AAAI president

641 trust researcher @tdietterich.bsky.social · 8,078 followers
AI research
Why they matter

Safe and robust AI/ML, former AAAI president with public evidence across AI research.

AI signals
5
past 30d
Sources
5
distinct domains
Discussões
27
past 30d
Latest signal
11d ago
View every signal from Thomas Dietterich →
Safe and robust AI/ML, computational sustainability. Former President AAAI and IMLS. Distinguished Professor Emeritus, Oregon State University. https://web.engr.oregonstate.edu/~tgd/
What they're sharing

Articles & links

↻ Thomas Dietterich reposted
Clément Canonne @ccanonne.github.io

The report from the AI + TCS Working Group held at the @simonsinstitute.bsky.social is out, including "twelve concrete near-term recommended actions." Currently reading it: while I have some reservations, but there is some good in it! Notably items 7, 8, and 11. simons.berkele…

AI + TCS Working Group simons.berkeley.edu
AI Weekly's analysis →
  • The Simons Institute published 'AI and TCS: The Next Six Months' on September 28, 2026, following a two-day working group at Calvin Lab.
  • The report lists twelve recommended actions, from a standardized AI methodology section in papers to grant coverage of reasonable LLM expenses.
  • Discussions were informed by 134 survey responses; organizers were Tony Feng, Zeph Landau, Amit Sahai, and Nikhil Srivastava.
Read full analysis →
 View on Bluesky →

That said, assessment of the significance of the work or the depth of insight are things I suspect LLMs will not do well. See also openreview.net/pdf?id=xz1gO... "Position: AI Should Verify, Not Judge, Scientific Work" by Prabhant Singh, et al. and arxiv.org/abs/2511.21843

Verifying your browser | OpenReview openreview.net
View on Bluesky · ♥ 2 ↻ 0 ↩ 0 · 101d ago

That said, assessment of the significance of the work or the depth of insight are things I suspect LLMs will not do well. See also openreview.net/pdf?id=xz1gO... "Position: AI Should Verify, Not Judge, Scientific Work" by Prabhant Singh, et al. and arxiv.org/abs/2511.21843

FLAWS: A Benchmark for Error Identification and Localization in Scientific Papers arxiv.org
AI Weekly's analysis →
  • FLAWS is a new benchmark of 713 paper-error pairs built by using LLMs to insert claim-invalidating errors into peer-reviewed papers.
  • GPT-5 led five frontier models with 39.1% identification accuracy at k=10, meaning it missed the planted flaw more often than not.
  • Claude Sonnet 4.5, DeepSeek Reasoner v3.1, Gemini 2.5 Pro and Grok 4 were also evaluated, all below GPT-5's top score.
Read full analysis →
View on Bluesky · ♥ 2 ↻ 0 ↩ 0 · 101d ago
Their own posts

Recent commentary

At @arxiv.bsky.social, we are receiving a new type of paper that I call an "I did this experiment" paper. These papers typically report some experiment with an LLM or LLM "agentic" workflow. They are the kind of experiments an "insider" engineer would run to optimize a system. 1/

View on Bluesky · ♥ 55 ↻ 11 ↩ 9 · 148d ago

I suggest that the right attitude to take toward the OpenAI math release is to imagine we have discovered a notebook full of results from a deceased mathematician. Our job is to make sense of the notebook.

View on Bluesky · ♥ 25 ↻ 0 ↩ 1 · 10d ago

Have people noticed that the word "evolution" is suddenly popular in machine learning papers that have nothing to do with modeling evolution or evolutionary algorithms? It seems to be used simply as a synonym for "training". I assume some LLM has started using it...

View on Bluesky · ♥ 11 ↻ 2 ↩ 1 · 28d ago

"readout" seems to be the latest LLM word of the day, judging from the arXiv papers I'm moderating today

View on Bluesky · ♥ 7 ↻ 0 ↩ 1 · 55d ago

It seems that all of the organizations training AI models are training for broad, general intelligence. This includes knowledge about dangerous topics such as bioterrorism and cybersecurity. I don't think such models should be available to the general public.

View on Bluesky · ♥ 6 ↻ 0 ↩ 1 · 17d ago
Their network

In Thomas Dietterich's orbit

Center = Thomas Dietterich. Left = members they follow (green edges). Right = members who follow them (blue edges). Top = mutual follows (orange edges, slightly larger). Drag any node to reposition; click to open that profile.

Are you Thomas Dietterich? Show it.

Add the Who’s Who of AI badge to your site or bio. It links back to this profile.

Listed in AI Weekly's Who's Who of AI

Markdown: [![Listed in AI Weekly's Who's Who of AI](https://aiweekly.co/modules/custom/aiweekly_whoswho/images/whoswho-badge.svg)](https://aiweekly.co/whos-who/person/thomas-dietterich)