Founders

Some advice I keep hearing is to build things now that can only be done poorly with ai, but that could be made immediately 100x better by simply swapping in a future/more capable LLM underneath that meets the threshold your product needs to feel magical. Makes a ton of sense when you compare things like v0 today vs v0 on release.

chess, markets, internet money, past: cofounder @evaluatexyz, @doodles current: @grandmaisters

Need to make sure the 100x improvement doesn’t imply openAi automatically swallows your product, though. 
I guess that’s implied by your condition that it works poorly with todays AI

Interested in: startups + Piano + Fitness (running + barbell movements)

Great point. So many AI products become obsolete so fast.