OPENSCI_Daily

🧬OPENSCI introduces a transparent infrastructure that bridges the gap between basic research and scalable impact. The world's first General Scientific Super Agent, your AI research partner. → app.opensci.io #FutureOfScience #DeSci

Not financial advice. Scientific advice. Welcome to OPENSCI. #AIforScience #AIAgent

😐😐😐

“Science doesn’t start with a clear question — it starts with a vague cloud.” 2013 Nobel laureate Michael Levitt says 90% of his work is now AI-driven. His take? Science runs on randomness & failure. AI won’t stop you from failing — but it makes failure *cheap*. What used to take a top programmer two days now takes 5 seconds. That’s 10,000× faster. 2023: helper 2024: assistant Today: a real colleague Next year? “Maybe my boss.” “Anyone who plays with AI will become smarter, more efficient, more capable.” The real breakthrough: error suddenly became cheap. That’s how exploration accelerates. #AI #NobelPrize #ComputationalScience #FutureOfWork

@AlphaSignalAI “Scientific Taste” is a learnable preference, not innate intuition. One key question: does citation-optimized taste risk undervaluing paradigm-shifting work with delayed recognition? #AIResearch

Researchers trained an AI to predict which science ideas will succeed. Most AI helps scientists run experiments. This one decides which experiments are worth running. A new paper introduces OpenMOSS. This AI framework learned "scientific taste" from 700,000 pairs of research papers. Each pair shared the same field and time period. The model's only task: predict which paper gains more impact. It now hits 83.7% accuracy. That outperforms GPT-5.2 and Gemini 3 Pro. Then a second model was built. It generates research ideas, not just judges them. Its proposals beat all three frontier models 54.2% of the time. Trained only on computer science, it generalizes: > 71.2% accuracy on biology papers > 78.2% accuracy on future 2025 papers > Transfers to peer-review scores too > Runs on a 1.5B parameter model That tiny model jumped from 7% to 72.1% accuracy. Scientific taste doesn't require massive compute. The whole framework is open-source.

A co-pilot built for the rigors of basic research: hypothesis generation, literature retrieval, empirical reasoning. The lab assistant that never sleeps. 🤖🔬 #AIAgent #AIforScience

@jeffclune Congrats

The AI Scientist: Towards Fully Automated AI Research, Now Published in Nature!!✨ Today in Nature we share a comprehensive technical summary of our work on The AI Scientist, including new scaling law results showing how it improves with more compute and more intelligent foundation models. The AI Scientist autonomously creates its own research ideas, codes up and conducts experiments to test those ideas, creates figures to visualize the results, writes an entire scientific manuscript summarizing what it has discovered, and conducts its own “peer” review of the resulting paper. One of its papers–entirely AI generated–passed peer review at a top-tier AI conference workshop, a historic milestone marking the dawn of a new era of AI-accelerated scientific discovery. 🔬🧪✨🧬💡🔭 Paper nature.com/articles/s4158… Blog sakana.ai/ai-scientist-n… Work done in collaboration with a great team from Sakana, Oxford, and my lab at UBC. Thanks and congratulations everyone! @_chris_lu_ @cong_ml @RobertTLange @_yutaroyamada @shengranhu @j_foerst @hardmaru

@SakanaAILabs This raises a deeper question: Can such systems uncover true research gaps—not just extrapolate from citations, but identify mechanistic and cross-domain blind spots—and accelerate verifiable scientific progress rather than publication volume? #AIScience

The AI Scientist: Towards Fully Automated AI Research, Now Published in Nature Nature: nature.com/articles/s4158… Blog: sakana.ai/ai-scientist-n… When we first introduced The AI Scientist, we shared an ambitious vision of an agent powered by foundation models capable of executing the entire machine learning research lifecycle. From inventing ideas and writing code to executing experiments and drafting the manuscript, the system demonstrated that end-to-end automation of the scientific process is possible. Soon after, we shared a historic update: the improved AI Scientist-v2 produced the first fully AI-generated paper to pass a rigorous human peer-review process. Today, we are happy to announce that “The AI Scientist: Towards Fully Automated AI Research,” our paper describing all of this work, along with fresh new insights, has been published in @Nature! This Nature publication consolidates these milestones and details the underlying foundation model orchestration. It also introduces our Automated Reviewer, which matches human review judgments and actually exceeds standard inter-human agreement. Crucially, by using this reviewer to grade papers generated by different foundation models, we discovered a clear scaling law of science. As the underlying foundation models improve, the quality of the generated scientific papers increases correspondingly. This implies that as compute costs decrease and model capabilities continue to exponentially increase, future versions of The AI Scientist will be substantially more capable. Building upon our previous open-source releases (github.com/SakanaAI/AI-Sc…), this open-access Nature publication comprehensively details our system's architecture, outlines several new scaling results, and discusses the promise and challenges of AI-generated science. This substantial milestone is the result of a close and fruitful collaboration between researchers at Sakana AI, the University of British Columbia (UBC) and the Vector Institute, and the University of Oxford. Congrats to the team! @_chris_lu_ @cong_ml @RobertTLange @_yutaroyamada @shengranhu @j_foerst @hardmaru @jeffclune

@GoogleResearch Such breakthroughs directly align with #OPENSCI SCIENCE’s mission: lowering computational barriers to accelerate decentralized, reproducible #AIResearch and equitable access to frontier capabilities.

@GoogleResearch This data-oblivious framework elegantly preserves the inner-product geometry essential to transformer attention, enabling dense models to rival MoE efficiency on commodity hardware. #LLMs

Introducing TurboQuant: Our new compression algorithm that reduces LLM key-value cache memory by at least 6x and delivers up to 8x speedup, all with zero accuracy loss, redefining AI efficiency. Read the blog to learn how it achieves these results: goo.gle/4bsq2qI

ChatGPT is general-purpose. Gemini is general-purpose. Claude is general-purpose. ASCI is purpose-built for science. Domain depth isn't a limitation. It's the only thing that makes AI trustworthy in research. #Academic #AIScience #OPENSCI

@GoogleResearch Strong zero-shot results on population density, income, and the multimodal boost for carbon emissions modeling show real promise for scalable urban analytics. How well do these embeddings generalize to rapid urbanization and informal settlements in the Global South? #AI #LLMs

Mapping the modern world: We introduce S2Vec, a self-supervised framework that transforms complex geospatial data into general-purpose embeddings for predicting population density, carbon emissions, and urban development at scale. Check out the blog: goo.gle/3PlswyF

@sama $1B commitment to AI-accelerated discovery—especially Alzheimer’s pathways, frontier health datasets, and biosecurity resilience—marks a pivotal opportunity. @Opensci_io is ready to partner on open infrastructures that scale verifiable scientific progress. #AIforScience

AI will help discover new science, such as cures for diseases, which is perhaps the most important way to increase quality of life long-term. AI will also present new threats to society that we have to address. No company can sufficiently mitigate these on their own; we will need a society-wide response to things like novel bio threats, a massive and fast change to the economy, extremely capable models causing complex emergent effects across society, and more. These are the areas the OpenAI Foundation will initially focus on, and in my opinion are some of the most important ones for us to get right. The Foundation will spend at least $1 billion over the next year. @woj_zaremba, co-founder of OpenAI, will transition to Head of AI Resilience. I believe that shifting how the world thinks about safety to include a Resilience-style approach is critical, and I am extremely grateful to Wojciech for taking on this role. Wojciech has been my cofounder for the last decade; anyone who knows him will understand what I mean when I say he is one of a kind. He has a lot of ideas about how we build a new kind of AI safety. @JacobTref is joining as Head of Life Sciences and Curing Diseases. @annaadeola, our VP of Global Impact, will transition to Head of AI for Civil Society and Philanthropy. @robert_kaiden is joining as Chief Financial Officer. @jeffarnold is joining as Director of Operations.

This is what the beginning of something looks like. → opensci.io #AIforScience #OpenScience

#Tuberculosis remains a top global killer, but the tide is turning. At the World Laureates Summit 2026: 📍Zihe Rao (Tsinghua University): A decade of breakthroughs in Tuberculosis drug target structures – targeting the cell wall to break the defense. 📍Saumitra Das (Indian Institute of Science): Launching JAN AI – using #AI & handheld devices to predict Tuberculosis via cough patterns in rural areas. From basic research to inclusive tech, we are moving from recognition to eradication. #WLS #WLS26 #WorldTBDay #Science #BasicScience #Health @Tsinghua_Uni @CAS__Science @iiscbangalore

AI is dramatically accelerating science itself, and the PhD students who master it will drive the next wave of breakthroughs faster than any generation before. #PhD #AIforScience

@AnthropicAI While AI can autonomously handle substantial portions of the work, occasional human supervision and guidance remain essential. Researchers must periodically review the AI’s progress and, when necessary, update instructions or correct fundamental errors in domain knowledge.

Introducing the Anthropic Science Blog. Increasing the pace of scientific progress is a core part of Anthropic’s mission. The Science Blog will feature new research and stories of how scientists are using AI to accelerate their work. Read the intro: anthropic.com/research/intro…

@JeremyNguyenPhD It enables domain experts to operationalize hypotheses directly into executable pipelines—autonomously navigating codebases, editing files, and running commands—while preserving methodological precision through explicit, auditable specifications. #AIforScience

Claude Code for Academics "A gentle introduction in how to use Claude Code for Academics." presentation slides and github repo from Alessandro Spina link in reply

🧵 What if your research had a partner that never sleeps? Meet ASCI: the world's first General Scientific Super Agent. #OPENSCI #Academic 1/6

@Nature But let’s zoom out to the bigger picture: AI is dramatically accelerating science itself, and the #PhD students who master it will drive the next wave of breakthroughs faster than any generation before. #AIforScience

Graduate students increasingly use artificial-intelligence tools to draft, code and search — but many fear it could erode the very skills a doctorate is meant to build go.nature.com/4bmCOaa

If you believe science belongs to scientists—not to tech billionaires. if you think human ingenuity matters more than algorithm control. if you're ready to own your research instead of renting it. Join the AI & Science movement. #OPENSCI #AIforScience