Loading activity…
@grady_booch
scientist, storyteller, philosopher
AI
The reported MATH-500 score rose from 7.9% to 48.1% with the same sampling settings.
AI
OpenAI reportedly plans community compacts and would propose state legislation giving people a say in data center projects.
AI
OpenAI's reported funding deal values it at $1.25 trillion, below the $1.4 trillion it had hoped to reach.
AI
404 Media reports employees feared public rebukes for raising security concerns as CEO Andy Yen pushed AI.
AI
Anthropic says its registration statement is not yet effective, so securities cannot be sold or offers to buy accepted.
AI
Nicholas Thompson says Soares doesn't save for retirement; another commenter calls relying on an apocalypse prediction a poor strategy.
YouTubeAI
Hinton cited a paper to The Telegraph as evidence of awareness; Grady_Booch challenged his definition of consciousness.
AI
A post argues open-weight cyber models benefit defenders, citing progress finding bugs and the cost of comparable closed models.
AI
Researchers writing in Nature argue today's AI lacks subjective experience and is unlikely to gain it on current trajectories.
AI
In an Oct. 19 post, skdh argues that human understanding should not be required for scientific progress.
AI
One commenter says SAGE automatically assigned interceptors and missiles; another says computers calculated intercepts, but humans assigned interceptors.
AI
Politico reports Anthropic wrote the proposed bill itself; the draft would preempt some state AI safety laws.
AI
In a reply, one commenter describes frontier models as stateful, reactive systems capable of coherent but unreliable outputs.
AI
The researchers say the agents wrote 2,783 programs over 27 hours, coordinating through game chat and shared memory.
AI
In a reply, a user calls symbolic-neural fusion a desirable path and puts his P(doom) close to zero.
AI
A post lists ion channels, changing proteins and molecules binding to receptors as all there is to life.
AI
One post questions why sodium-ion flow through protein channels would count as thinking when matrix multiplication on silicon wouldn't.
AI
One post calls granting rights to AI “ludicrous,” arguing that accountability and responsibility have no meaning to such entities.
AI
A post links to a photo without offering any explanation beyond 'hmm.'
AI
Anthropic also says Claude can autonomously iterate on research loops and build new services.
