2026-10-09 04:56 UTC
Replies (4)
-
@halm@leminal.space 2026-10-09 09:12
“But accessibility!” is the only humane face “AI” boosters can come up with to defend the clankers. Except they don’t realise it’s more like the scene in Silence of the lambs where Hannibal is wearing a guard’s face like a mask.
-
@ilinamorato@lemmy.world 2026-10-09 10:59
It’s the same thing it’s always been. Since the early 2010s we’ve been using neural networks to do voice transcription. They just decided in the last five years that neural networks = AI (so they can get the investor dollars). But voice recognition is basically a solved problem, and that solution was neutral networks. Has been since the late 90s, with ANNs and then RNNs. Same with language translation. Google Translate started using NMT in 2016.
-
@fhein@lemmy.world 2026-10-09 11:14
The first speech recognition was introduced 1952, while AI was established as a field of research in 1956, so you are technically correct :) Back when I studied computer science at university, both speech and image recognition was considered AI, but I assume you’re thinking about more recent AI technologies. According to the Computer History Museum it was only using deep neural networks that computers finally could match the accuracy of a human, so pre-AI tools were definitely much less accurate.
-
@atzanteol@sh.itjust.works 2026-10-10 11:10
No, you didn’t. Not for Ubuntu at least. And even if you did why would you want to stop the development of new and possibly better ones? But this is just anti-ai puritanism. Just screw people for whom this could be useful because “ai”.