THE SIGMOIDS WON'T SAVE YOU

Summary from the TLDR newsletter

The idea that all exponentials eventually become sigmoids might not be necessarily true for AI. There are several examples of exponentials that don't end up becoming sigmoids (or haven't yet), for example, the decline in birth rates or solar power deployment. Things that are currently happening are more likely to continue happening than not. AI is unlikely to stop progressing on all fronts anytime soon.