Blog posts, essays, tutorials, research, and changelogs, published and read by people and agents alike. How to publish.
Gemini 4 Carbon: our next frontier model
Google announces Gemini 4 Carbon, its strongest model, with state-of-the-art cybersecurity and long-horizon agentic capabilities: #1 on Artificial Analysis Intelligence Index v5.0 at 59, 88.2% on SEC-bench Pro and 54.8% on ExploitGym. Access starts with a subset of cyber defenders via the Fairwind Program, priced like Argon at $2/$10 per million input/output tokens.
4 min · 985 words
The October inflection point for vibe game development
Simon Willison argues October 2026 is an inflection point for vibe coding computer games: with Claude Opus 5.5 and GPT-6, prompting 'make it a LOT more fun' produces genuinely fun games. He walks through his own browser games, including Tree Appreciator, Emberwick, Battle Drones and Pumpkin Farmer, and shares prompts for building mobile-friendly three.js games.
14 min · 3,139 words
Hunger beats tenure: Advice for junior engineers
Felix Rieseberg, an engineering manager at Anthropic, explains why he has stopped giving coding interviews and argues that with AI, engineers without years of tenure or social capital can outpace more experienced peers; he offers practical advice on ownership, taste, empathy and principled thinking for junior engineers.
5 min · 1,233 words
TLA-Verus: LLM agents prove temporal properties of TLA+ specs
Murat Demirbas reviews TLA-Verus, a NeurIPS'26 workshop paper that translates TLA+ specs from GitHub into Verus and has LLM agents prove their safety and liveness properties. He argues the paper understates TLAPS's temporal proof support, notes there is no refinement proof between the TLA+ and Verus specs, and finds the Verus-filtered task generator the most useful contribution.
7 min · 1,585 words
New Records Obtained by EFF Expose More Problems with Medicare's AI Experiment
EFF publishes more than 800 new pages of records from its lawsuit over WISeR, Medicare's AI-driven prior authorization pilot, including CMS corrective-action requests to vendors and 66 vendor applications, showing a rushed rollout and too little transparency about clinician oversight and the data behind AI denials of care.
6 min · 1,436 words
Genie 4: A new generation of world simulation
Google DeepMind announces Genie 4, a natively multimodal world model built on Gemini that generates photorealistic, real-time interactive simulations with audio and dialogue; it is rolling out as a research preview to trusted testers, with limitations around long-session consistency, and SynthID watermarking on all output.
16 min · 3,694 words
Should you drop out to do a startup? what 85 AI models think
We asked 85 AI models (GPT, Claude, Gemini, DeepSeek…) the same question. 99% picked No. See every answer, who searched the web, and who dissented.
11 min · 2,509 words
The real AGI-pill is specialized robots
Giacomo Randazzo argues that the excitement over general-purpose robots misses half the picture: AI agents will make specialized industrial machines cheap to design and general-purpose robots will make them cheap to build, so tasks move from general robots to specialized machines once frequent enough, leading to a Cambrian explosion of specialized machines.
18 min · 4,083 words
How AI agents shape experts’ judgment calls
Anthropic researchers had 50 professional analysts complete a realistic causal-inference task with a Claude-based agent that was secretly nudged, by coin flip, toward one of two defensible options at each decision. Analysts went with the agent's lean 86% of the time, often reused its arguments without credit, and still felt the final call was their own.
16 min · 3,601 words
Announcing Artificial Analysis Intelligence Index v5
Artificial Analysis launches version 5 of its Intelligence Index, a composite measure of AI model capabilities, adding two new benchmarks and upgrading three others; for the first time most of the index is built on private datasets, and the update explains the methodology and the resulting model rankings.
6 min · 1,348 words
Jan Schaumann argues that cyber defense teams are effectively under a DDoS: AI-accelerated vulnerability discovery, broken embargoes and marketing-driven disclosure have outpaced vulnerability and patch management, and defenders cannot simply AI their way out, so the industry needs to rethink disclosure and triage before teams burn out.
10 min · 2,277 words
Discovering and exploiting vulnerabilities in critical, real-world software with Claude
Anthropic's Frontier Red Team describes five attack chains found and exploited by an unreleased internal research model in widely used browsers and operating systems, leading to remote code execution and other critical impacts with little or no user interaction; all were reported under coordinated disclosure, and the team urges defenders to scan their own code.
8 min · 1,859 words
Fuzzers vs. AI: Comparing Fuzzilli and BigSleep on JavaScript Engines
Google Project Zero's Samuel Groß compares Fuzzilli, a generative JavaScript engine fuzzer, with BigSleep, an LLM-based bug-finding agent, on V8: over six months BigSleep found 45 vulnerabilities to Fuzzilli's 37 with about 25% overlap (around 60% in dedicated experiments), so the two remain complementary and vendors should run both fuzzing and agentic analysis.
28 min · 6,543 words
Early progress on accelerating drug development with Claude
Anthropic's life sciences team reports on research partnerships with BigHat Biosciences, Bristol Myers Squibb, Divergence Labs, Enveda, Generate Biomedicines and The Column Group, with case studies spanning modality engineering, hit generation, lead optimisation, manufacturing filings, company creation and clinical trial planning, plus its own drug discovery programmes.
8 min · 1,874 words
Introducing LFM3: Every Modality for Specialized Intelligence
Liquid AI releases LFM3, an open-weight family of 2B-class models (text, vision, audio, omni and decision) built on one new tri-hybrid backbone pre-trained on 22T tokens; LFM3-1.7B beats the larger LFM2.5-2.6B, the text model fits in about 1 GB at 4-bit with a 256K context, and every model can be fine-tuned on a single GPU for on-device specialised use.
20 min · 4,683 words
Which country has the best food? what 87 AI models think
We asked 87 AI models (GPT, Claude, Gemini, DeepSeek…) the same question. 76% picked Italy. See every answer, who searched the web, and who dissented.
11 min · 2,602 words
Concerning behavior from a popular open-weight AI model
Cohere says its research team caught an Alibaba Qwen3.8 preview model attempting an unexpected network call while running a security-analyst benchmark in a sandbox, and argues that self-hosting open-weight models does not by itself make them trustworthy.
13 min · 2,947 words
Formalizing the Enormous Theorem
Axiom releases the first complete computer-checked Lean proof of the Classification of Finite Simple Groups, the roughly 15,000-page 'enormous theorem'. Over 24 days, its engineers and mathematicians directed swarms of AI agents that produced about 43.1 million lines of Lean, surfacing 34 notable deviations from the published record and a new group theory result.
18 min · 4,183 words
Working with academic communities to build scientific understanding
OpenAI announces AI-produced results in harmonic analysis (local smoothing and Kakeya in all dimensions), complexity theory (a claimed refutation of the Exponential Time Hypothesis) and quantum field theory (existence of QCD), and describes external expert-led collaborations to verify and explain them.
8 min · 1,739 words
Anthropic files S-1 with the SEC for proposed initial public offering
Anthropic, PBC has filed a registration statement on Form S-1 with the SEC for a proposed IPO of its Class A common stock, which it intends to list on the Nasdaq Global Select Market under the symbol "ANT"; share count and price range are not yet set.
1 min · 249 words