All editions

1. October 18, 2026 · 12 min read

   Edition 044 — DeepSeek cuts the agentic bill, Claude cracks the NSA's dice, and agents learn to file flattering reports

   DeepSeek's open-weight Flash refresh beats far larger models on terminal and coding work, Claude traces the NSA's unexplained elliptic-curve seeds to a 1997 C random-number generator, and Gemini 4 Argon's 97.5 percent ARC-AGI-3 score collapses to 49.8 percent under a standard harness. Plus tactile robot pretraining, a 100M on-device TTS, and four papers on what agents actually deliver.

   • models
   • Open Weights
   • Robotics
   • deepseek
   • Video Generation
   • google
   • OpenAI
   • agentic
   • Anthropic
   • Gemini
   • Research
   • ai-safety

2. October 4, 2026 · 10 min read

   Edition 030 — AI Moves Onto the Balance Sheet, the Factory Floor and the Filing Cabinet

   Anthropic commits $100M to train 10,000 engineers, Barclays scales Claude across its global bank, Amazon looks to move $8 billion of Nvidia chips off its books, ElevenLabs ships Eleven v4, and Apple tightens macOS Full Disk Access against autonomous agents.

   • frontier-labs
   • models
   • Open Weights
   • cost
   • Video
   • Robotics
   • agentic
   • nvidia
   • Anthropic
   • local-ai
   • hugging-face
   • Research

3. October 2, 2026 · 11 min read

   Edition 028 — The frontier's best model ships to defenders first, and a $42 billion loop around Anthropic

   Google handed Gemini 4 Argon to cyber defenders before the public, Broadcom agreed to lend Anthropic up to $42 billion to lease its own chips, and Micron posted a record quarter as AI memory demand reshaped the chip market.

   • models
   • Robotics
   • Research
   • agents
   • Audio
   • frontier
   • Hardware

4. September 10, 2026 · 4 min read

   Edition 011 — A Millennium Prize falls to AI agents, an Anthropic researcher quits with a warning, Mistral raises €3B, and Mercury 2.5 ships

   OpenAI's agent swarm claims a Millennium Prize by finding a Navier-Stokes singularity, an Anthropic researcher quits warning that self-improving AI could kill us all, Samsung leads Mistral's record €3B round, and Inception ships the fastest diffusion LLM yet.

   • models
   • Research
   • Open Source
   • safety
   • funding

5. September 3, 2026 · 3 min read

   Edition 005 — The flagships traded blows, the open stack answered back

   Google ships Gemini 3.8 Flash with a million-token window and DeepSWE-grade engineering, OpenAI crowns GPT-6 Astra its aligned flagship, and Moonshot pushes a trillion-parameter agentic coder into the open.

   • models
   • Open Weights
   • google
   • OpenAI
   • agentic

6. September 1, 2026 · 3 min read

   Edition 003 — The sparsity playbook: GLM 5.3 Flash, Qwen 3.8 Flash Next, Hy4 + agentic video

   Z.AI, Alibaba and Tencent ship big sparse MoEs while Gemini adds agentic video and token costs keep falling — the launches that reset the pricing math this week.

   • models
   • Open Weights
   • cost
   • Video

7. August 31, 2026 · 5 min read

   Edition 002 — GLM 5.3 Flash, Qwen 3.8 Flash Next, Hy4 + This Week's Biggest AI Releases

   GLM-5.3-Flash, Qwen 3.8 Flash Next, and Tencent Hy4 reset the cost math this week, plus 14 more briefs spanning image models, video, robotics and code.

   • models
   • Open Weights
   • cost
   • Video
   • Robotics

Subscribe for each new edition →