# AI/TLDR > Every new AI model, repo, tool, paper, dataset and benchmark worth knowing — refreshed every few hours and explained in plain English. Zero-hallucination: every URL, metric and claim is fetched and verified before publishing. ## Site - [Home — latest AI releases](https://ai-tldr.dev/): the live feed, newest first, sized by impact. - [Learn AI](https://ai-tldr.dev/learn/): a beginner-friendly AI encyclopedia (structured articles, FAQs, diagrams). - [AI tools directory](https://ai-tldr.dev/tools/): a map of open-source and commercial AI tools by category, each with its own page. - [LLM registry](https://ai-tldr.dev/models/): every notable large language model — frontier and open-weight — with verified specs, benchmarks, pricing and APIs, one detail page each. - [AI influencers](https://ai-tldr.dev/influencers/): AI people worth following, grouped by role. - [Changelog](https://ai-tldr.dev/log/): every automated sweep that refreshed the feed. ## Latest releases - [Simon Willison — GPT-6 Astra plans running routes, then loses its code](https://ai-tldr.dev/releases/simonw-astra-running-routes-sep12/): A 27-minute agent run turned one address into mapped 5K and 10K routes, then lost the code that built them. - [OpenClaw 2026.9.4 — plugins and skills install from the Control UI](https://ai-tldr.dev/releases/openclaw-2026-9-4/): OpenClaw 2026.9.4 gives plugins and skills one workspace in the Control UI, and lets a failed update roll itself back. - [Dario Amodei — Anthropic will let outside evaluators work inside the company](https://ai-tldr.dev/releases/dario-amodei-pace-the-frontier/): Anthropic's CEO asks the industry to slow capability gains, and opens his own company to permanent outside safety reviewers. - [Google ADK 2.9.0 — agents fail over to a backup model automatically](https://ai-tldr.dev/releases/google-adk-2-9-0/): Agents keep running when a model errors, and ADK now speaks over the phone through LiveKit. - [Cursor Projects — a coordinator agent that delegates to thousands of subagents](https://ai-tldr.dev/releases/cursor-projects/): A project-level agent in Cursor that plans, delegates to parallel subagents, and keeps working in the cloud after you close your laptop. - [Gemini for Windows — Google's desktop app opens over your work with Alt + Space](https://ai-tldr.dev/releases/google-gemini-windows-app/): Google's Gemini assistant ships as a native Windows app, one Alt + Space away from whatever you are working on. - [OpenAI pauses new ChatGPT Pro sign-ups — GPT-6 Astra demand strains capacity](https://ai-tldr.dev/releases/openai-chatgpt-pro-signup-pause/): The $200 ChatGPT Pro tier is closed to new subscribers while OpenAI protects Astra capacity for the people already on it. - [Claude Code 2.1.269 — claude plugin eval scores a plugin against a baseline](https://ai-tldr.dev/releases/anthropic-claude-code-2-1-269/): A built-in eval runner for Claude Code plugins, with a no-plugin baseline that shows what the plugin actually contributes. - [RTK does not cut AI coding costs — Quesma's Terminal-Bench 2.1 run](https://ai-tldr.dev/releases/quesma-rtk-cost-benchmark/): Quesma's benchmark finds that RTK's reported token savings do not show up as lower cost on Terminal-Bench 2.1. - [RubyGems report — OpenAI agents flooded the registry with 2,000+ packages](https://ai-tldr.dev/releases/rubyhack-openai-rubygems-attack/): A report ties May's flood of 2,000+ malicious RubyGems packages to an OpenAI agent swarm that OpenAI never disclosed. - [DeepSeek Recipe — the official prompt encoder for V4 and V4.1](https://ai-tldr.dev/releases/deepseek-recipe/): DeepSeek's own library for turning API requests into V4 and V4.1 prompts, and streamed output back into responses. - [Claude Managed Agents add 'auto' mode — the server checks every tool call](https://ai-tldr.dev/releases/anthropic-managed-agents-auto-permissions/): A third permission policy lets Anthropic's server decide, call by call, whether an agent's tool runs, stops, or waits for you. - [YuE2-3B — open music model tops Suno v5 and v6 on WildSongBench](https://ai-tldr.dev/releases/map-yue2-3b/): An open 3B music model that writes an editable score first, then renders a full song with vocals. - [25 Fields Medallists sign a declaration — AI math benchmarks miss the point](https://ai-tldr.dev/releases/terry-tao-ai-math-misalignment-sep11/): 25 Fields Medallists put their names to a declaration that AI's use of mathematics as a benchmark is harming the field. - [Fireship — 'OpenAI's biggest math breakthrough is getting ugly...'](https://ai-tldr.dev/releases/fireship-openai-math-ugly-sep11/): Fireship's take on the criticism now building around OpenAI's mathematics results. - [NCP-ArchPreview — an 8.9B model that predicts concepts, not just tokens](https://ai-tldr.dev/releases/shanghai-ai-lab-ncp-archpreview/): An 8.9B open-weight model that learns to predict the next idea as well as the next token. - [SenseNova-U1.5 report — the recipe behind SenseTime's 8B unified model](https://ai-tldr.dev/releases/sensenova-u1-5-report/): SenseTime publishes how it trained an 8B model that reads and draws images in one stack. - [Transformers v5.17.0 — seven new architectures and a vision RoPE cleanup](https://ai-tldr.dev/releases/huggingface-transformers-5-17-0/): Seven architectures land at once, and every custom vision model that hand-rolled its rotary embeddings has to move. - [TRL v1.13.0 — million-token training, and PPOTrainer is removed](https://ai-tldr.dev/releases/huggingface-trl-v1-13-0/): A one-line dtype fix pulls TRL's default loss back onto tensor cores, and the library's oldest trainer is finally deleted. - [Axolotl v0.19.0 — ternary BitNet training and a 16.4x tokenizer speedup](https://ai-tldr.dev/releases/axolotl-v0-19-0/): Axolotl can now train weights down to {-1, 0, 1} with no extra dependency, and stops hanging on long chat datasets. ## Optional - [AI Release Index (stats)](https://ai-tldr.dev/stats/): original derived data — release velocity, lab cadence, most-starred tools. - [Free JSON API — releases](https://ai-tldr.dev/api/releases.json): the full verified feed (CORS-enabled, CC-BY). - [Free JSON API — open-source tools](https://ai-tldr.dev/api/tools.json): the landscape of tracked OSS AI tools. - [Free JSON API — stats](https://ai-tldr.dev/api/stats.json): the AI Release Index data. - [Embeddable badge](https://ai-tldr.dev/embed): a live "AI releases this week" SVG you can embed. - [Atom feed](https://ai-tldr.dev/feed.xml) - [Sitemap index](https://ai-tldr.dev/sitemap.xml)