Things With AI — Notes on AI

AI Engineering Series10 mins

How to Train a 1.6T Parameter MoE on a Budget: Inside DeepSeek-V4's Pre-Training Stack

The Math That Beat the Export Controls: DeepSeek-V4's Radical Training Efficiency

June 7, 2026

DeepSeek Efficiency

Part 1

How to Train a 1.6T Parameter MoE on a Budget: Inside DeepSeek-V4's Pre-Training Stack

The Math That Beat the Export Controls: DeepSeek-V4's Radical Training Efficiency

10 mins

The AI Knowledge Series

Part 1

RAG vs. Fine-Tuning — The Question Every AI Builder Gets Wrong

AI models don't know your private data. Two approaches have been the standard answer. In 2026, a third matters just as much.

5 min read Part 2

Inside RAG — How It Really Works (And Why Most Projects Stall Before They Ship)

The gap between the RAG concept and production reality is where most projects quietly fail. Here's what the explanations usually skip.

6 min read Part 3

Fine-Tuning in 2026 — Cheaper Than You Think, Harder Than You Think

Bloomberg spent roughly $10 million on it. You can now get started for under $100. What changed, and when does fine-tuning actually make sense?

6 min read Part 4

Agentic RAG — The Architecture That Made the Debate Beside the Point

Once you see RAG and fine-tuning operating inside a reasoning loop, the question of which one to choose starts to feel like asking whether a workshop needs better hammers or better saws.

6 min read