2026-09-23 22:30 UTC
RE: https://social.coop/@cwebber/117321242779306632
One of the excuses that comes up again and again about GenAI/LLMs is "It's okay to use as long as a human is checking the output."
In the real world, as the thread attests, humans *aren't* checking outputs, because their money-grabbing bosses are telling them to hurry up, time is money, why so slow etc.
If GenAI/LLMs are used to make something, the most realistic assumption has to be that it will be used as sloppily as the company can get away with.
(And even if this was solved, it would still leave environmental problems, labour theft, energy theft, water theft, deskilling, dependence on fascist CEOs etc.)
#AI #GenAI #LLM
Replies (2)
-
@alpacamale@social.cologne 2026-09-24 01:23
@FediThing@social.chinwag.org There's also review fatigue. Even a diligent software engineer with all the time in the world and the most understanding boss will get tired or sloppy after reading a thousand lines of code.
-
@slotos@toot.community 2026-09-24 02:57
@FediThing@social.chinwag.org Checking output has never worked. But it’s no surprise this bullshit comes from the software industry, where QA Engineer is a title of someone forced to do QC work.