this post was submitted on 26 Sep 2024
277 points (98.3% liked)
Asklemmy
43781 readers
950 users here now
A loosely moderated place to ask open-ended questions
Search asklemmy ๐
If your post meets the following criteria, it's welcome here!
- Open-ended question
- Not offensive: at this point, we do not have the bandwidth to moderate overtly political discussions. Assume best intent and be excellent to each other.
- Not regarding using or support for Lemmy: context, see the list of support communities and tools for finding communities below
- Not ad nauseam inducing: please make sure it is a question that would be new to most members
- An actual topic of discussion
Looking for support?
Looking for a community?
- Lemmyverse: community search
- sub.rehab: maps old subreddits to fediverse options, marks official as such
- [email protected]: a community for finding communities
~Icon~ ~by~ ~@Double_[email protected]~
founded 5 years ago
MODERATORS
you are viewing a single comment's thread
view the rest of the comments
view the rest of the comments
When AI gets applied to robot bodies, real world results will be able to trim out bad knowledge. Currently because AI only feeds on internet content, all the AI has to eat is human content and AI content.
AI will drift away from accuracy until it gets embodied at which point it will start to get more accurate.
Just have a nuclear reactor to run the LLM on the robot and you're all set
Real world experience can help, but what we have now is also too stupid to recognize when it's succeeding or failing. It just greedily gobbles up inputs and feedback indiscriminately.
There's currently no way to know if the necessary advancement, to advance independently of humans, is 2 years or 2000 years away.
Even so, nature tells us that advancement probably isn't coming at all. It's not needed, so long as there are billions of humans available to partner with.