"RSI" has always struck me as the classic example of a LessWrong-era coinage that just doesn't make much sense anymore.
Either it has the trivial meaning of "AI makes AI developers more productive" or the implausible one of "AI autonomously rewrites its own capabilities with little-to-no human assistance."
People used to mean the latter (see the LessWrong wiki definition here: lesswrong.com/w/recursive-se…). I have no clue what they mean now.
Timothy B. Lee@binarybits
I sincerely don't understand what people mean when they say this. On the one hand, every AI researcher is already using Claude Code (or its competitors) to help them develop new architectures. OTOH, AI models do not have bodies so they can't build data centers
English
