Do shocking performance gains of smaller open models in the last few months upset the Big Tech AI oligopoly?
No offense, obviously, because I appreciate that this post is a thoughtful accounting of the current situation with AI/LLM but man the mumbo jumbo levels here are at near crypto currency levels. This stuff is so far afield from anything that could be useful in the real world.
"7. What undesirable societal effects can we predict with confidence in the near term and are there realistic mechanisms to mitigate any of these effects?"
This one certainly suffers from a deficit of insights with actionable value. I took a crack at the first half of the point in a post on the economics of efficiency gains. Most of the article is about framing the problem, then describing the context, and then delivering specific predictions about the job market. Didn't bother going into solutions, but perhaps someone working towards some will find the analysis useful.