Post #3797926
2026-07-14 00:27 UTC
"much token usage, or ‘chewing,’ is not from supercharged engineers creating lots of code, but people converting PDFs into presentation slides."
In other words, AI doing "grunt work" is too expensive. You cannot make this up.
https://www.404media.co/companies-are-throttling-employees-ai-use-because-its-too-expensive/
Replies (1)
-
@shac@ioc.exchange 2026-07-14 02:20
@jilleduffy@mastodon.social I see many instances of running the same prompt to do a task repeatedly that would be much better served by having the AI produce a script once that would do the same task repeatedly. People use LLMs as the only tool when it is expensive and unpredictable, instead of using it to build cheap and reliable tools and processes.