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

Google Developers BlogLabs

Reproducing OLMo 3 7B Pre-training in MaxText: case study of large scale training on TPUs

September 25

The MaxText team successfully reproduced AI2’s OLMo 3 7B language model from scratch on Google Cloud TPUs using JAX/XLA, precisely matching the original PyTorch-on-GPU reference across pre-training…

Models & releases Chips & compute

Read at Google Developers Blog ↗

One letter a week, in your inbox.

Latest

Reproducing OLMo 3 7B Pre-training in MaxText: case study of large scale training on TPUs

Related

LabsEnterprise-Grade Precision for Long-Context Multimodal Embedding Inference on Cloud TPU Google Developers Blog

AsiaThailand approves first chip plan, targets $80b Tech in Asia

AsiaDeepSeek Details DSec Elastic Compute: Agentic-Training Sandboxes at ~3M/Day, 380K+ Concurrent Pandaily

AsiaAlibaba reportedly testing QwenBook as a tablet-like agent computer TechNode

SiliconAccelerating Dropless MoE Training in JAX with NVIDIA Transformer Engine NVIDIA Technical Blog

SiliconMeta Superintelligence – Leadership Compute, Talent, and Data SemiAnalysis