#fp16

The latest AI and developer news about #fp16, with the original source and a short summary.

Notes on Pretraining Parallelization and Failed Training Runs

It identified broken causality and numerical bias as the core causes of pretraining failures.

·1·2026.05.17 04:00

What We Learned This Week (20 min read)

Frontier models are harder to prevent **distillation** of than expected, and large-scale pretraining can easily collapse from small **bias** and **causality** violations.

·1·2026.04.17 09:00