this post was submitted on 15 Jul 2024
826 points (97.7% liked)

memes

9806 readers
6 users here now

Community rules

1. Be civilNo trolling, bigotry or other insulting / annoying behaviour

2. No politicsThis is non-politics community. For political memes please go to [email protected]

3. No recent repostsCheck for reposts when posting a meme, you can only repost after 1 month

4. No botsNo bots without the express approval of the mods or the admins

5. No Spam/AdsNo advertisements or spam. This is an instance rule and the only way to live.

Sister communities

founded 1 year ago
MODERATORS
[email protected]
[email protected]
[email protected]
826
AI taking over the world (lemmyf.uk)
submitted 3 months ago by [email protected] to c/[email protected]
40 comments fedilink hide all child comments
 
you are viewing a single comment's thread
view the rest of the comments
[–] [email protected] 11 points 3 months ago* (last edited 3 months ago) (1 children)

AI not having all that much to catch up to.

That's what you think. But we will sooner have fusion power than AGI. Comparing it with the developement of personal computers, we're currently at the stage of punchcard weaving machines.

[–] [email protected] -2 points 3 months ago (1 children)

And you base this on what? Both things are very hard to predict even for experts in the respective fields.

[–] [email protected] 6 points 3 months ago (1 children)

Experts in both fields have significant paychecks on the line for people believing them. Of the two, fusion is the only one showing (very minor) measurable progress.

We can't even define what success would look like for AGI.

[–] [email protected] -1 points 3 months ago

AGI is software with human like intelligence that can self-improve and the last two years is very measurable progress. We went from essentially nothing to multimodal models that can run on consumer hardware. OpenAIs new model, if they are to believed, is apparently much better at reasoning and can do long term research. I've also seen a few papers in the wild talking about self teaching methods and framework.

To be clear, I don't know which will come first. It's hard to know if the next leap is just a step or if there's a giant chasm laying infront. I do know that it's a lot easier to prototype with AI then fusion and there's a lot more people working on it, both behind closed doors and publically on the internet. Fusion doesn't have this advantage.

Your statement is basically a shot in the dark imo.