AI compute

AI chips, inference, post-training, and the stack underneath.

A topic hub for the physical and software layers behind AI: power, silicon, HBM, GPUs, frameworks, model weights, inference serving, and agents.

Interactive map

The AI stack

From electrons to tokens, layer by layer.

Decision guide

AI chip buyer's guide

H100, H200, B200, GB300, TPU, Trainium, MI300X, Huawei 910C, and Groq by workload.

Training pipeline

The RL post-training map

SFT, RLHF, DPO, PPO, GRPO, and RLVR in one pipeline.

Reference

AI, term by term

CUDA, CoWoS, HBM, vLLM, SGLang, MoE, and the vocabulary around AI infrastructure.

Essay

Nine layers between electrons and tokens

The compute stack as a dependency chain, with failure modes exposed.