Matt Clifford

This is very true. I’ve probably shared “The Scaling Hypothesis” more than any other post. From 2020, but still essential reading:

Great post by @simonw that points to a lot of near-term AI safety / security challenges

Ian is one of the smartest and wisest people I know; this is worth a read👇

Fascinating work: “These models hold the promise to have context lengths of millions… or maybe even a billion!” Also shows huge power of new primitives in ML: even if there were *no* more research, remixing of existing ideas would take us a long way…

Fascinating and important post from Jack

Yes, lots of quirks between 3 and 3.5 because of training and safety methods (long and technical but interesting post on this: but would be very surprised if 4 didn't blast 3 out of the water on this (caveat: am often surprised!)

Very smart piece, worth a read

“Magic has not been a part of everyday life since the seventeenth century, so it’s no wonder we are a little out of practice” Excellent essay on how we process AI advances.

This is really excellent; thanks for writing it!

Important essay, both depressing and stimulating, by @adam_tooze on the UK economy in historical context

Very good piece by @patio11 nominally on SWIFT but much broader: great on “the financial system [as] a very blunt instrument for expressing society’s preferences”

Superb piece on Ukraine and geopolitical predictions by @slatestarcodex: