Post #2117568
2026-04-24 22:42 UTC
Sure... for static photos of inanimate objects image recognition its ok. It is easily fooled by false perspectives, weird lighting and odd angles but whatever.
The problem is that AI is a shit show as soon as you try to adapt it for real world use. The capability of these things are beyond exaggerated because tech bros lie and bias test results (because it makes them very rich).
Such speculation of using AI for this or that is part of the scam. Best not to do it.
Replies (1)
-
@leoj@piefed.social 2026-04-27 12:57
I mean this discussion is centered around static location and a target that can be color primed for recognition, so I think most of your points fall flat. Sticking your fingers in your ears and pretending it is fully worthless is not a productive take, but I understand it is the cultural zeitgeist around here - I'm willing to go against a wave when I think its partially wrong.