28 free resources on this topic. Everything here is free and hand-picked. You can also search within this topic.
METR reviews Anthropic’s assessment of automated research risks. It agrees the reviewed models posed low direct acute risk, but disputes how the public report described acceleration and early-warning thresholds. METR emphasizes uncertainty and contestable judgments.
METR assesses whether AI is accelerating research and development at Anthropic. It estimates faster capability progress on well-specified tasks and attributes much of the acceleration to AI use, while stressing uncertainty. Anthropic redacted evidence that would help outsiders independently assess the findings.
An updating snapshot of the math community's gut reactions to OpenAI's October 6, 2026 announcement, collected by the Proofs and Prompts blog.
Statement from the Advisory Group on Mathematics and Artificial Intelligence responding to OpenAI's release of manuscripts claiming solutions to high-profile math problems.
Intelligence science lab; its site describes it as working on intelligence science.
OpenAI's public GitHub repository, released October 6, 2026, of manuscripts the company says contain solutions to a number of high-profile mathematics problems. See the AHM statement for mathematicians' response.
Allen Institute for AI (Ai2) post from October 7, 2026 on its byte-level language model research appearing in Nature, with released checkpoints applying its "byteifying" method to Qwen 3 8B and Llama 3 8B.
Epoch AI's early results from InnovationEval, published October 7, 2026. Epoch reports that the two frontier models tested, Claude Fable 5 and GPT-5.6 Sol, fell far short of the original human-developed technique and that their write-ups overstated their results.
Free website for following new research papers, built around arXiv. It shows trending papers with summaries, lets readers comment on papers, and offers literature-review answers grounded in the underlying papers. Browsing is free; an account is needed to comment or save.
Rich Sutton's March 2019 essay arguing that the biggest lesson from 70 years of AI research is that general methods leveraging computation ultimately prove most effective, by a large margin. Sutton walks through prominent cases — computer chess, Go, speech recognition, and computer vision — where approaches built on human knowledge were outdone by search and learning methods that scaled with computing power. He attributes the pattern to the steadily falling cost of computation and warns that relying on human knowledge tends to complicate methods in ways that limit scaling.
A nonprofit research institute focused on basic AI research into how minds arise from computation. Its site frames minds as computational entities capable of goal-directed adaptive behavior in open, uncertain worlds, and lists open questions such as how minds learn in real time from sensorimotor experience and how synthetic and human minds can cooperate. The institute builds on The Alberta Plan for AI Research, says all results will be shared publicly with no intellectual property restrictions, and is recruiting research fellows.
The personal site of François Chollet, software engineer and AI researcher, co-founder of Ndea and the ARC Prize. It collects his books (including Deep Learning with Python), software, papers, essays and talks, and describes his research interests: the nature of abstraction, autonomous abstraction algorithms, and democratizing AI development.
The home page of the Stanford Artificial Intelligence Laboratory, a center for AI research, teaching, theory and practice since its founding in 1963. The site lists faculty, research groups, affiliated centers, courses, events and lab news.
The personal site of Noah Shinn, a programmer and researcher who was one of the first employees and a researcher at Sierra. His papers include tau-bench, a benchmark for tool-agent-user interaction (ICLR 2025), work on code-editing instruction following, type prediction, and the Reflexion paper on language agents with verbal reinforcement learning.
A free, open platform where machine-learning researchers submit and peer-review papers in public. Much of the field's conference review (including major AI venues) runs on it, and anyone can read the papers, reviews and rebuttals.
Epoch AI report on a spatial reasoning benchmark built around finding mistakes in photos of partially assembled IKEA furniture — a proxy for economically important visual reasoning tasks like repair work. Top scores jumped from 28% to 80% in 10 months; Chinese open-weight models trail the frontier by about 7 months.
The Decoder report on comments from Boris Power, OpenAI's Head of Applied Research, that most of the company's research targets future model generations, and that user awareness of AI capabilities lags model performance.
An interactive case study walking through a two-day investigation that recovered the machine settings for MVUEH, an 82-letter German Army Enigma message from July 10, 1941, with source comparisons, experiments and checks readers can reproduce. Published by an individual on a ChatGPT-hosted site; its claims have not been independently reviewed.
The free research library and blog of Snorkel AI, the company spun out of Stanford's Snorkel project on programmatic labelling. The papers and posts explain how training data for AI models is actually built — labelling, evaluation sets, and the 'environments' used to train agents. Useful if you want to understand the unglamorous data work behind model quality, which is where a lot of the real jobs are.
Why I recommend it: Research and blog posts are free to read with no signup. Read it for the how, not the verdict: Snorkel sells data services to frontier AI labs, so posts arguing that better data beats bigger models are also a sales case. Everything else on the site is a paid enterprise product — 'request dataset samples' means a sales call.
An essay arguing for building a large model of the natural world — weather, oceans, ecosystems — as its own kind of general intelligence, separate from language models.
Why I recommend it: One argument, well written, worth reading as a counterweight to the LLM-only view of where AI is heading. It is a manifesto, not peer-reviewed research.
Stanford professor who built ImageNet, co-directs Stanford's Human-Centered AI institute and co-founded the AI4ALL diversity pipeline program. Her faculty page collects the work; her Google Scholar list has the papers themselves.
From the site: Fei-Fei Li is part of Stanford Profiles, official site for faculty, postdocs, students and staff information (Expertise, Bio, Research, Publications, and more). The site facilitates research and collaboration in academic endeavors.
Why I recommend it: Start with ImageNet if you want to understand why the last decade of AI happened when it did. Google Scholar refuses automated visits, so that link may show no picture here.
Microsoft's research division: published papers, open datasets and tools, plus its internship and residency programs.
From the site: Explore research at Microsoft, a site featuring the impact of research along with publications, products, downloads, and research careers.
Why I recommend it: The publications are free to read and the programs page lists real entry routes into research work. Both are more useful than the marketing pages.
Montreal research institute founded by Yoshua Bengio: publications, research teams, and programs for students and visiting researchers.
From the site: Mila is a Montreal-based artificial intelligence research institute that brings together researchers from Université de Montréal, McGill University, Polytechnique Montréal and HEC Montréal.
Why I recommend it: Check the students and programs pages if you want to move toward research work. Academic institutes publish their entry routes more openly than companies do.
Yann LeCun's position paper arguing that today's language models are the wrong architecture, and sketching what he thinks should replace them. Free to read.
Why I recommend it: The serious technical case against scaling language models further. Dense, but it is the argument itself rather than a summary of it.
A free explorer for new arXiv research with plain-language paper summaries, topic pages and video overviews, so you can follow AI research without reading raw papers.
From the site: Your first stop to discover and learn about new arXiv research. Detailed paper summaries, video overviews, and more — no prompting required.
Why I recommend it: The fastest way I know to keep up with AI research when you are not a researcher. Free to browse.
TypeSafe AI announcement from founder Diogo Almeida (formerly OpenAI) introducing System One models and Jev, aimed at cheaper automation rather than better chat.
Why I recommend it: Worth skimming to track where new AI labs are placing bets — useful context for interviews at AI companies.
Research-driven lab studying memory and judgment in AI agents, publishing work on agent memory systems.
Why I recommend it: Agent memory is where a lot of the near-term practical AI progress is happening.
Accessible breakdowns of AI and machine learning research papers for a general audience. Hosted by Károly Zsolnai-Fehér.
In plain terms: This YouTube channel offers simple video breakdowns of artificial intelligence and machine learning research papers. You can watch these quick guides to stay informed about new AI developments and understand how the technology is evolving.
Why I recommend it: Short, visual, and it makes research feel approachable.
