- Sep 28, 2026One Control PointAt my day job, I was asked to be more efficient with AI spend. The honest answer was a semantic router in front of every model, so I pointed my own coding agent at one. It broke. Five bugs later, the fixes are upstream, which means everybody gets them.
- Inference
- vLLM
- Open Source
- +2 more
- Sep 10, 2026The Same Brain, Another PassA normal model is twelve editors who each read your draft once. A looped model is one editor who reads it twelve times. I spent this week working out what that buys you, what it costs, and where it stops working.
- Transformers
- Inference
- AI Research
- +1 more