Post #1127887

2026-04-03 11:09 UTC

If you can look past the clickbaity title (pretty sure Connor never actually says &quot;mutate&quot;), I think this is another v good discussion with Connor Leahy. He explains... - How LLMs work and how they apparently achieve general learning from patterns which was a &quot;holy grail&quot; of AI research - The risk of losing control; why we need to slow the heck down and regulate its development (I know it&#39;s not popular message, but think it&#39;s too important to easily dismiss) https://youtu.be/rf2KFVcKQdQ

Replies (2)

@peter@toot.cafe 2026-04-03 13:04

Some points lots of people neglect: - More important than how LLMs predict the next word is how the attention mechanism builds "understanding" of patterns. - It's not just human language, they work on tokens of different kinds & achieve multi-modal "understanding". - Losing control of AI doesn't require it to have emotions or sentience, just goals. - The AI companies are trying to achieve AGI/ASI through "recursive self-improvement", and AI is already being used to help develop AI.

Open ##1986153
@markhughes@mastodon.social 2026-04-03 11:33

@peter my take is that the second will be futile if the first is true. Therefore the first doesn't matter. I don't believe it is likely either but am happy to wait and see. There are though other reasons to regulate the shit out of LLMs, and they're all things that would be achievable and worth doing, if only the people hell bent on this weren't the same people who have captured our political systems. So IMO it won't happen and we'll just have to ride this out and hope.

Open ##1986155