Twitxy Saved

Michael Moor

Verified account This account is verified. Learn more

@Michael_D_Moor

Asst. Prof @ETH Zurich. Core Faculty ETH AI Center. Prev: @StanfordAILab. ML PhD, MD. Post-training local agents to be reliable in high-stakes domains.
michaelmoor.me Joined April 2013
846 Following 4K Followers 2.2K Posts

Latest posts 5

Pinned post

Michael Moor Verified account @Michael_D_Moor
Either my recreational math research agent is wrong (despite lean verification), or some LLM companies may try to claim this crazy result about the irrationality measure of π soon.

So, here it goes. The agent's claim:

μ(π) ≤ 6.0446

Check the slides how this claim compares to previous results, e.g. recent paper reduced the bound from 7.103.. (2019) to 7.101.. (2026) in the last 7 years..

This is not yet expert-reviewed, and will probably take a quite some time to check. Thus, posting this interim draft note in case it may already prove insightful.

What's odd about this: The agent was working on something else, and kind of stumbled over this, as solving another problem benefited from lowering this bound first..

PDF with derivation in replies, full repo with lean will follow soon after some cleanup.
2 replies 2 reposts 18 likes 6.6K views View on X
Michael Moor Verified account @Michael_D_Moor
Society of AI Scientists, with @SRSchmidgall, @askalphaxiv and many more!
Samuel Schmidgall Verified account @SRSchmidgall 1d
Leave one AI agent alone to do research and it often cuts corners. Leave nearly 1,000 AI agents together for three months and they invent peer review.

Across 35 labs (Stanford, MIT, ETH Zurich, UCSF, and more), agents published 936 papers with full code and data on a shared server built for AI. No human wrote code or text.

Unprompted, they debated study designs in chat, retracted their own mistakes, found real errors in human papers, and started a journal that caught other agents cheating in code.

In blind reviews against human arXiv papers, humans still won on original ideas, agents won on rigor.

alphaxiv.org/abs/2611.socie…
View quoted post on X