https://ai-tldr.dev/releases/google-gemini-3-8-flash/ AI/TLDR en 2026-09-02T12:00:00+00:00 Gemini 3.8 Flash — Google's best coding model yet, plus a cyber twin https://ai-tldr.dev/releases/qwen-3-8-max-0902/ AI/TLDR en 2026-09-02T12:00:00+00:00 Qwen3.8-Max-0902 — Alibaba's 2.4T flagship gets a coding refresh https://ai-tldr.dev/releases/carloslfu-slotstream/ AI/TLDR en 2026-09-01T12:00:00+00:00 Slotstream v0.2.0 — a 104GB model on a 48GB Mac, plus speculative decoding https://ai-tldr.dev/releases/google-pics/ AI/TLDR en 2026-09-01T12:00:00+00:00 Google Pics — an AI image tool in Workspace where you start from a prompt https://ai-tldr.dev/releases/manifold-gitspawn/ AI/TLDR en 2026-09-01T12:00:00+00:00 GitSpawn — a repo's git config can run code in Claude Code, Codex and Cursor https://ai-tldr.dev/releases/metr-security-update-2026/ AI/TLDR en 2026-08-31T12:00:00+00:00 METR discloses two breaches — $600K of model credits burned unnoticed https://ai-tldr.dev/releases/wes-roth-astra-critical-sep2/ AI/TLDR en 2026-09-02T12:00:00+00:00 Wes Roth — 'GPT-6 Astra Just Went CRITICAL' on OpenAI's frontier safeguards https://ai-tldr.dev/releases/world-labs-atlas/ AI/TLDR en 2026-09-01T12:00:00+00:00 Atlas — World Labs' omni model for text, image, video and 3D https://ai-tldr.dev/releases/wes-roth-fable-51-astra-sep1/ AI/TLDR en 2026-09-01T12:00:00+00:00 Wes Roth — 'Fable 5.1 just smoked ASTRA' on the new frontier models https://ai-tldr.dev/releases/openai-astra-critical-cyber-safeguards/ AI/TLDR en 2026-09-01T12:00:00+00:00 Astra hits Critical cyber capability — OpenAI locks it down before release https://ai-tldr.dev/releases/openai-chatgpt-healthcare-epic/ AI/TLDR en 2026-09-01T12:00:00+00:00 ChatGPT connects to Epic — clinicians can pull chart context into the chat https://ai-tldr.dev/releases/simonw-codex-libreoffice-sep1/ AI/TLDR en 2026-09-01T12:00:00+00:00 Simon Willison — the ChatGPT desktop app ships a full copy of LibreOffice https://ai-tldr.dev/releases/danluu-zitron-predictions/ AI/TLDR en 2026-09-01T12:00:00+00:00 Dan Luu — checking Ed Zitron's AI predictions against the numbers https://ai-tldr.dev/releases/anthropic-enterprise-frontier-safeguards/ AI/TLDR en 2026-09-01T12:00:00+00:00 Enterprise Frontier Safeguards — misuse checks run in your own cloud https://ai-tldr.dev/releases/anthropic-claude-fable-5-1/ AI/TLDR en 2026-09-01T12:00:00+00:00 Claude Fable 5.1 — Anthropic's new top model, with cache reads 75% cheaper https://ai-tldr.dev/releases/google-gemini-agentic-video/ AI/TLDR en 2026-09-01T12:00:00+00:00 Agentic video in Gemini — the model loads only the clips it needs https://ai-tldr.dev/releases/anthropic-claude-code-2-1-257/ AI/TLDR en 2026-09-01T12:00:00+00:00 Claude Code 2.1.257 — Claude Fable 5.1 becomes the default Fable model https://ai-tldr.dev/releases/fireship-ox-alpha-unmasked-sep1/ AI/TLDR en 2026-09-01T12:00:00+00:00 Fireship — 'A mysterious new model just took over the internet' https://ai-tldr.dev/releases/google-timesfm-3/ AI/TLDR en 2026-08-31T12:00:00+00:00 TimesFM-3 — Google's forecasting model handles many series at once https://ai-tldr.dev/releases/two-minute-papers-glm-5-3-sep1/ AI/TLDR en 2026-09-01T12:00:00+00:00 Two Minute Papers — 'GLM 5.3: Powerful AI Is Becoming Almost Free' https://ai-tldr.dev/releases/dod-genai-mil-chatgpt-grok/ AI/TLDR en 2026-08-31T12:00:00+00:00 ChatGPT Mil and Grok for Government — the Pentagon's AI portal adds two models https://ai-tldr.dev/releases/openai-codex-cli-0-152/ AI/TLDR en 2026-09-01T12:00:00+00:00 Codex CLI 0.152.0 — the planning tool is now off by default https://ai-tldr.dev/releases/wes-roth-apple-ai-hardware-sep1/ AI/TLDR en 2026-09-01T12:00:00+00:00 Wes Roth — 'Apple became an AI company OVERNIGHT' on the new Macs https://ai-tldr.dev/releases/deepseek-v4-flash-vision-exp-open-weights/ AI/TLDR en 2026-08-31T12:00:00+00:00 DeepSeek-V4-Flash-Vision-Exp weights go public — 305B multimodal MoE under MIT https://ai-tldr.dev/releases/anthropic-alignment-security-hardening/ AI/TLDR en 2026-08-31T12:00:00+00:00 Anthropic locks down its test sandboxes — a classifier now blocks escape attempts