@khoomeik
descending cross-entropy to ascend entropy @PeriodicLabs || prev research @OpenAI @CarnegieMellon '23
AI
Arora shared defense slides titled “Causal Interpretability for Analysing and Improving Language Model Architectures.”
AI
A user joked on October 13 that readers might not know the unnamed event was happening.
AI
A ProximalHQ team member says they dynamically adjust the inference-trainer ratio to maintain training throughput and reduce staleness.
AI
David Pfau objected to the definition, while Dylan Hadfield-Menell called the evaluation metric optimization.
AI
OpenAI says it drew on an independent mathematics advisory group’s advice about how to release the results.
AI
The person inviting applications says the work would involve collaborating closely with them.
AI
Google announces a 1 million-token output limit, up from 64,000, for its new model. Initial access goes to trusted cyber defenders through the Fairwind Program.
AI
A user praises a team and congratulates Calvin Chen.
AI
A user says standard semiconductor equipment can fabricate the robots, which they claim can move, think and communicate. They attribute the robots’ “walking” to a platinum-and-titanium materials breakthrough.
AI
A user describes the reported lab as a logical next step for AI drug discovery, arguing that computer simulations alone aren't enough.
Science
Neon's builders describe a loop between their Menlo Park materials labs and AI: experiments generate fresh data, models learn from it, and then help choose what to try next.
AI
Posts say 25 Fields Medal winners signed a declaration warning of a mismatch between AI companies and mathematics. Supporters stress understanding and scientific communities; critics say AI proofs can still yield insights.
AI
The writer pictures capital as a possibly “misaligned” intelligence, waiting in the shadows and accumulating for decades.
AI
The author points to fruit fly simulations modeled on real brains as systems that can be highly intelligent without necessarily being conscious.
AI
Benchmark tests models on extended autonomous software engineering tasks lasting up to 20 hours.
AI
AI researcher Rohan Pandey describes limits on public comments after brief frontier lab roles.
AI
Taylor Lorenz shares her article on AI anthropomorphizing discourse and its header.
AI
Stephen Balaban notes their use of Lambda's mission statement as prompt.
AI
Hiring targets data curation, evaluations, and training tasks to advance AI models for physical sciences.
