I wrote up some (late) predictions for 2026, mostly as a way to guide my own thinking and priorities. But would love feedback and where I may be miscalibrated. Overall: 2026 is the year we will finally see the contours of society begin to visibly change in response to AI. I discuss three reasons: 1. The proliferation of advanced and autonomous AI agents in economic and social spaces. 🦞is a small glimpse, a trial run, in a controlled sandbox. Things are going to get a lot wilder, buckle up. 2. Recursive improvement of science and research. 3. Solving continual learning (this I'm most uncertain about). For each case, I stress the urgency of developing rules, governance, and institutions that facilitate the benefits of these changes, and mitigate the risks (which are...hopefully obvious). That's what I'll be working on this year. Here is the post: