emax: teaching "better prompting" has diminishing returns, cost/carbon is a systems problem

emax · x · 2026-09-26

In an X thread, emax argues that managing AI cost and carbon footprint is fundamentally a UX and infrastructure problem, not a user-education problem. Teaching non-experts to "prompt better" for specific models has diminishing returns; it's more valuable to learn to use AI systems generally. He calls expecting non-expert users to estimate or minimize token carbon footprint "bonkers," and adds there's greenfield work for systems designers optimizing cost and carbon, with existing token routers likely good enough at first.

Original post →

More from AGI Musings

AGI Musings channel →