5 ms·
On the Imminence and Danger of AI
- api 12y agoI'm the author. IMHO that post is not fully baked, but I figured I'd share it anyway. I think the best part is actually the argument against "exploding" AI and runaway self-improvement. If I had the time I'd give that a more thorough treatment with better references to relevant work and theorems.