llm

Published Soursen pieces tagged llm, page 2, with the standfirst and the beat. Every fact is checked against the source.

AI & GPU compute

I Tried Making AI Image Generation 90x Cheaper. Here Is What I Learned About GPU Quantization.

The bill had a comma in it.

13 July 20269 min read

AI & GPU compute

Why You Should Stop Using Ollama in 2026 and What to Use Instead

Let me start with a confession. I used Ollama for a long time, and I recommended it to almost everyone who asked me about running AI models locally. It deserved that recommendation when it launched.

9 July 202611 min read

AI & GPU compute

Japan Gave the World Robots, Bullet Trains, and PlayStation. So Why Is It Losing the AI Race?

This is the question that does not quite make sense on the surface.

30 June 20269 min read

AI & GPU compute

The AI Industry Built a $5 Trillion Debt Pile on One Bet. That Bet Is Not Paying Off.

Something shifted in the AI narrative over the past few months, and it is worth paying attention to. The CEOs who spent years telling anyone who would listen that AI would eliminate entire categories of human work have gone quiet on that point.

23 June 202612 min read