Can language models perform end-to-end scientific discovery? In our NeurIPS Spotlight paper, we show: very rarely. Our best model found <20% of discoveries, our best PhDs found nearly all. Paper: Code/Web: @allen_ai @MSFTResearch

RT @KexinHuang5: 🤝Excited to announce @ProjectBiomni × @AnthropicAI! . AI agents are set to transform how biologists do everyday research.….

RT @Thom_Wolf: We are so excited to announce a new open-source challenge in collaboration with @proximafusion : unlocking fusion with AI. I….

Table 5. Average number of proposals funded decreased from current 26% to 7%. Instead of 1 in 4 submitted proposals being funded (already low), it could decrease to 1 in 14.

A reduction of almost 75% (330k -> 90k) of NSF-funded scientific staff. For every ~10 PhD students funded in 2024, only ~3 would be funded in 2026. Source (page 22, summary tables 6):

RT @james_y_zou: Introducing Fractional Reasoning: a mechanistic method to quantitatively control how much thinking a LLM performs. tldr:….

RT @allen_ai: 🚨 We're hiring a #ResearchScientist in #AI for Scientific Discovery at Ai2!. Are you passionate about intelligent agents, dat….

RT @ChengleiSi: Are AI scientists already better than human researchers?. We recruited 43 PhD students to spend 3 months executing research….

RT @allen_ai: Introducing SciArena, a platform for benchmarking models across scientific literature tasks. Inspired by Chatbot Arena, SciAr….

RT @AnthropicAI: Anthropic staff realized they could ask Claude to buy things that weren’t just food & drink. After someone randomly deci….

RT @allen_ai: Today we’re releasing a prototype of Genesys, an autonomous multi-agent LLM discovery system that aims to discover new types….

RT @MishaTeplitskiy: Verrrrry intriguing-looking and labor-intensive test of whether LLMs can come up with good scientific ideas. After imp….

RT @PHDcomics: Meeting with your PI

RT @LucasPCaccia: RAG and in-context learning are the go-to approaches for integrating new knowledge into LLMs, making inference very ineff….

RT @akshay_pachaar: A RAG engine for deep document understanding!. RAGFlow lets you build enterprise-grade RAG workflows on complex docs wi….

RT @StasBekman: My first project at @Snowflake AI Research is complete! . I present to you Arctic Long Sequence Training (ALST) . Paper: ht….

RT @josephimperial_: NeurIPS D&B track in a nutshell:. (1) An LLM-generated benchmark dataset.(2) used to test performance of LLMs.(3) eval….

RT @emnlpmeeting: Update on reviews for EMNLP, which were due on June 18th: as of today, only about 80% of reviews have been completed 😧. A….

25 years ago, when he was a week into his professorship and I was a week into my undergraduate, I asked some very naive questions in one of his seminars on quantum computing that he very patiently and encouragingly answered. I remember that kindness to this day.

RT @j_g_allen: Wow. Read this sentence:.—> Robert F. Kennedy Jr.’s new group of vaccine panel appointees “fails to include experts on any d….

RT @__paleologo: I know more than one people who can guess betas and correlations visually. Once I met a person (not joking) who guessed th….