Terence Tao: At the Erdos problem website, AI assistance now becoming routine(mathstodon.xyz)

260 points by dwohnitmok 11 months ago | 56 comments

Quantitative AI progress needs accurate and transparent evaluation(mathstodon.xyz)

196 points by bertman 1 year ago | 96 comments

Terence Tao: The role of small organizations in society has shrunk significantly(mathstodon.xyz)

968 points by bertman 1 year ago | 490 comments

A Mastodon post by mathematician Terence Tao reflecting on how society operates at four scales and how the role of small organized groups has shrunk.

I doubt that anything resembling genuine AGI is within reach of current AI tools(mathstodon.xyz)

56 points by gmays 10 months ago | 33 comments

A human metaphor for evaluating AI capability(mathstodon.xyz)

127 points by bertman 1 year ago | 24 comments

Since Linux 6.9, LUKS suspend stopped wiping disk-encryption keys from memory(mathstodon.xyz)

516 points by IngoBlechschmid 3 months ago | 217 comments

Mastodon thread by Ingo Blechschmidt describing how since Linux 6.9 the LUKS suspend tool silently failed to wipe disk-encryption keys from memory.

Tao: Open math problems being non-renewably mined by AI(mathstodon.xyz)

418 points by _alternator_ 1 month ago | 359 comments

A Mastodon thread by Terence Tao arguing that good, fruitful open math problems are a scarce resource being mined non-renewably, using a drinking-water analogy.

“Math 2.0” will need to value mathematical progress more holistically(mathstodon.xyz)

579 points by ent101 18 days ago | 603 comments

A Mastodon post by Terence Tao arguing that "Math 2.0" should value exposition, community building and new directions over raw problem solving, with AI playing a role.

What mathematicians should know about the Lean Theorem Prover: reliability & AI(terrytao.wordpress.com)

178 points by matt_d 16 days ago | 50 comments

Terence Tao blog post on what mathematicians should know about the Lean theorem prover, covering reliability and AI; the capture shows only a browser-check interstitial.

AI agents: Less capability, more reliability, please(sergey.fyi)

393 points by serjester 1 year ago | 230 comments

Blog post by Sergey Filimonov arguing AI agents should prioritize reliability over expanded capability, critiquing flight-booking demos.

FrontierMath was funded by OpenAI(lesswrong.com)

429 points by wujerry2000 1 year ago | 156 comments

A LessWrong shortform post by user meemi, linked as reporting that the FrontierMath AI math benchmark was funded by OpenAI.

Did OpenAI solve the wrong Navier-Stokes problem?(scientificamerican.com)

92 points by tomjakubowski 1 month ago | 46 comments

Scientific American article on whether OpenAI's claimed Navier-Stokes proof qualifies for the $1-million Millennium Prize only through a controversial loophole.