2026-09-02
The AI decision most founders think they are making is which model, which vendor, rent or own. The decision they have not made yet is the one that makes all the others possible.
2026-08-22
The trade-off is not speed against depth. It is speed against your ability to tell when the answer is wrong. One is a preference. The other is a failure mode.
2026-04-23
LLMs are fast. But speed has a trade off. When you don't have the depth to know when the LLM is wrong, you're fragile.
2026-02-08
A day of talks from OpenAI, Hugging Face, Mistral, GitHub and others on the Model Context Protocol: what's working, what's still broken, and where this is all going.