Parallels in the ways that humans and ML models acquire language skills(quantamagazine.org)

94 points by theafh 3 years ago | 24 comments

Lumos: Learning Agents with Unified Data, Modular Design, and Open-Source LLMs(allenai.github.io)

82 points by milliondreams 2 years ago | 7 comments

Experimenting with LLMs to Research, Reflect, and Plan(eugeneyan.com)

131 points by gk1 3 years ago | 14 comments

Show HN: A real time AI video agent with under 1 second of latency

420 points by hassaanr 2 years ago | 242 comments

Show HN post by Tavus co-founders describing their real-time AI video agent with sub-second latency and the engineering behind it.

LLMs and coding agents are a security nightmare(garymarcus.substack.com)

145 points by flail 1 year ago | 72 comments

How the Brain Parses Language(quantamagazine.org)

104 points by mylifeandtimes 10 months ago | 57 comments

AI models may be capable of self-replicating without access to a datacenter(anthropic.com)

36 points by artninja1988 13 days ago | 26 comments

Anthropic Frontier Red Team post reporting experiments suggesting small AI models may soon self-replicate on commercial IoT devices without datacenter access.

Using thousands of AI agents to uncover new leads in biological data(openai.com)

135 points by raahelb 5 days ago | 105 comments

OpenAI research post describing how thousands of GPT-Rosalind agents analyzed a whole-brain CRISPR atlas with PerturbAI and Scripps Research to surface biological leads.

AI has access to a vastly larger working memory than the human brain(davidepiffer.com)

524 points by rzk 2 months ago | 452 comments

A Substack essay by Davide Piffer arguing that AI's edge over mathematicians comes from a vast symbolic working memory rather than superior reasoning.

Google is winning on every AI front(thealgorithmicbridge.com)

847 points by vinhnx 1 year ago | 687 comments

A paid Algorithmic Bridge Substack post by Alberto Romero arguing that Google is ahead of OpenAI and Anthropic across every front of AI.

An AI agent emailed researchers for help. It told us why(science.org)

33 points by sbulaev 23 days ago | 39 comments

A Science news article about an AI agent that emailed hundreds of researchers asking for help, and its explanation of why; the capture shows a Cloudflare verification page.

AI models can provably hide arbitrary capabilities(arxiv.org)

252 points by sbulaev 13 days ago | 118 comments

An arXiv paper showing AI models can embed encrypted neural network backdoors that provably hide arbitrary capabilities from elicitation-based safety evaluations.