Geoffrey Hinton on current and future AI risks and concerns

3 Brysonbw 1 6/18/2025, 10:53:16 PM youtube.com ↗

Comments (1)

moritzwarhier · 9h ago
I believe that Geoffrey Hinton is much more intelligent than I am, by a large margin.

But it stands out to me that he appears as if he'd voluntarily taken the perspective of an AI-fear evangelist (because it bolsters his reputation).

I cannot judge if it is legitimate to choose this category ("fear-evangelist"). I did it anyway, deliberately.

Generative AI, LLMs, Machine Learning: I'm not an expert in any of these fields. I heard a lecture on Neural Networks, Genetic algorithms, AI history at university, but that was in 2013.

It's interesting how some approaches that felt like moon-shots at the time turned out to be massively capable with the emergence of large transformer models. At least that's my understanding (regarding NNs).

If someone says to me: "you are replaceable by AI", I don't fret much. It might be correct.

But to be honest, I am also frustrated about the fact that AI gloom and doom (be it white-collar job elimination or paperclip singularity) hasn't changed much since 2022 or so.

Agree that it feels we are at the beginning of a disruptive period, where work that was previously considered inseparable from human domain experts and specialized programmers, is being challenged by AI.