34 stories tagged by Digg
AI
In posts tied to a new MIT Technology Review piece, former Google DeepMind researcher Thore Graepel argued that advanced AI needs auditable reasoning, then drew responses from Yann LeCun, David Duvenaud and Gary Marcus.
AI
Gary Marcus said on PBS NewsHour that self-regulation is not enough for AI safety. PBS says researchers and lawmakers are calling for stronger safeguards.
AI
A user describes a “split screen”: what they characterize as “trust us” messaging from Trump and Jensen alongside daily news they say gives people reason not to trust OpenAI.
AI
CNBC reports the accord calls for internal monitoring, outside assessments and board oversight; Trump called it “morally binding.”
AI
A post citing EY says the survey covered over 200 US senior AI decision-makers at publicly traded companies. It says 36% reported that their organizations had experienced an AI incident or failure with a materially negative impact.
AI
A user points to the cybersecurity cat-and-mouse game as part of where attention should be focused now and shares what they describe as a clip from their CNN appearance.
AI
A critic says he and others warned for weeks that OpenAI’s cybersecurity was an immediate problem, with little attention. He says Jensen is addressing it with new software, though he doesn’t yet know how well it works.
AI
The author says well-designed harnesses can make structurally similar tasks look nearly identical to individual model calls, even when the tasks come from different domains.
AI
A user says they agree with professor Gary Marcus that current state-of-the-art AI technology is neurosymbolic. Marcus reposted the comment.
AI
Quote from a Stanford AI Club talk draws agreement from Gary Marcus on AI limits.
AI
Stephen Casper and Gary Marcus react to recent OpenAI and METR reports on X.
AI
Payments company shared its view in a letter to investors covered by Axios.
AI
Physician-scientist supports AI in health while rejecting recent cure timelines as unrealistic.
AI
Cognitive scientist responds to Florian Tramèr over whether modern LLMs qualify.
AI
Conversation among AI researchers highlights unsolicited messages asserting knowledge of private experiments.
AI
He credits the o3 test-time compute breakthrough with changing his earlier assessment of large language models.
AI
Chollet posted that million-line codebases calling neural models at inference qualify as neurosymbolic.
AI
Jay Hack highlights debate over autoregressive models in programming systems.
AI
AI researchers debate whether public calls to slow frontier progress amount to illegal collusion.
AI
Cognitive scientist compares AI demo to historical math advances like calculus and algebra.
