so grok4 takes us like 2x on humanity’s last exam the problem is almost nobody knows what that means and it’s probably optimized to the problem set

i dabble with ai using open identity data from crypto https://alexpaden.tech

prev startups & racecars

Grok 4 is a product of genius not benchmark gaming

you probably don’t think about liquidity enough // it is what it is  // build value, not valuation

every new model is a product of genius i'll believe that as true when the other models are no longer needed (which isn't currently true)