We all know intelligence is spikey/jagged - but do we know _why_ it is? TL;DR is everyone is cheating in AI (and because we're releasing 🤫 soon, we're making precommitments to do better 💪)

Sutton's bitter lesson is roughly: scale > algorithms Sharing the wisdom that anyone who has made ML useful knows: neither scaling nor algorithms is the be-all and end-all

Wrote about the incredibly important KV cache: - Why are people spending so much on agents - How is inference so profitable all of a sudden (+ why it won't last) - Why routing doesn't work P.S. they also constrain the design space for coding agents A LOT