AI timelines can grow old. Reading through the last highly publicized exercise, it's as if we were stuck in the early 2020s: ever larger models, unlocking ever more impressive emerging capacities. GPT-3 steampunk. We are already in a different cycle: pretraining as we know it is ending. Reasoning, reinforcement learning, mid and post-training are drawing most of the focus. The first productivized agents will nearly certainly be relatively small models. In short, it's time to reopen the timelines...