
The article discusses the inevitability of an open model consortium, with companies like Nvidia potentially supporting open models but facing pressures to pull back due to competition and financial constraints. Future business models will involve releasing smaller, fine-tunable models rather than the absolute frontier models, as maintaining high-cost, near-frontier research becomes increasingly difficult. The shift toward smaller, custom models is expected to grow, while the number of companies willing to release fully open, near-frontier models will decline.
Reflections on AI's writing ability and how AI models get more capable.

After a few long years of finding time to document my lessons from training open models, my post-training book is done!

Musings on model alignment, what determines safety, and where we go from here.
