this post was submitted on 26 Nov 2023
29 points (93.9% liked)

Futurology

0 readers
1 users here now

founded 1 year ago
MODERATORS
all 6 comments
sorted by: hot top controversial new old
[–] qooqie@lemmy.world 3 points 11 months ago (2 children)

Ha ha! AI is kinda going a little crazy fast in development isn’t it? ha ha!

[–] SchizoDenji@lemm.ee 1 points 11 months ago

Not really. Image Gen AIs are developing very quicker than LLMs because of the breakthroughs in open source models.

LLMs are still quite behind.

[–] Espiritdescali@futurology.today 2 points 11 months ago

I don't think LLM's will lead to AGI, but at some point a system is going to be implemented that does lead to AGI but it will be unexpected.

This seems to be the worse case scenario as the AGI will likely be clever enough to ensure humans don't realise it's AGI. There are network effects and complexity that we are not 100% knowledgeable about. The net result is the same: we lose the control problem. Badly.