The name is the button.

An agent writes the code now, and a human thumb hovers over Accept all a hundred times a day. The honest question of this decade is what, exactly, you just accepted. So I read the diff.

Eight years as a backend engineer in Stockholm, back in Tokyo since 2023, and I still ship something small most nights, usually badly. My one rule: the only real review of a model is the thing you built with it. That is why there is no leaderboard on this site.

Around a hundred and forty sources come through — the labs' own posts, the papers, the tools, the runtimes and languages nobody is watching yet, the chip under the demo, and the newsletters people actually forward. What is moving gets counted so the week reads at a glance, and every few days I explain one mechanism in plain words.

A story here is a headline, a short excerpt, one image and the link to whoever wrote it. We never reproduce the post; the images belong to their authors. A post written in Japanese appears in Japanese, exactly as it was published, and the link goes to the original.

— Yuki Halvorsen, Tokyo

← Accept All Archive

Ahead of AIVoices

From DeepSeek V3 to V3.2: Architecture, Sparse Attention, and RL Updates

2025.12.03

Last updated: January 1st, 2026 Similar to DeepSeek V3, the team released their new flagship model over a major US holiday weekend.

Models & releases Research

Read at Ahead of AI ↗

One letter a week, starting soon.

Latest

From DeepSeek V3 to V3.2: Architecture, Sparse Attention, and RL Updates

Related

VoicesUnderstanding and Implementing Qwen3 From Scratch Ahead of AI

VoicesUnderstanding the 4 Main Approaches to LLM Evaluation (From Scratch) Ahead of AI

VoicesFrom GPT-2 to gpt-oss: Analyzing the Architectural Advances Ahead of AI

LanguagesBun is joining Anthropic Bun Blog

ToolsRetro Diffusion's pixel art models are now on Replicate Replicate Blog

VoicesThree Years from GPT-3 to Gemini 3 One Useful Thing