the "infinite iterations" pitch assumes infinite compute.


every call costs money, runs through someone else's infrastructure, and burns tokens whether the output is brilliant or garbage.
iteration is only cheap if you already know what you're looking for.
> someone with 10 years of taste built up runs 3 iterations and ships.
> someone without it runs 300 and still can't tell which one's good.
and no one priced this in yet:
the more AI outputs, the more human verification you need, so 10x the drafts means 10x the people checking them.
... except for the person who already knows what good looks like.
AI didn't remove the bottleneck. it moved it from production to judgment, which doesn't scale with headcount. it scales with taste.
taste isn't a soft skill anymore. it's a cost function.
This page may contain third-party content, which is provided for information purposes only (not representations/warranties) and should not be considered as an endorsement of its views by Gate, nor as financial or professional advice. See Disclaimer for details.
  • Reward
  • Comment
  • Repost
  • Share
Comment
Add a comment
Add a comment
No comments
  • Pin