It will take time, but as both a wish and a prediction for 2025, I hope to see more interesting approaches to developing open-weight (o1 and o3-like) reasoning models through improved post-training recipes, data, and inference compute.

→ View original post on X — @rasbt,