Machines don't understand anytime.
this post was submitted on 10 Mar 2026
2 points (53.6% liked)
Technology
82488 readers
3968 users here now
This is a most excellent place for technology news and articles.
Our Rules
- Follow the lemmy.world rules.
- Only tech related news or articles.
- Be excellent to each other!
- Mod approved content bots can post up to 10 articles per day.
- Threads asking for personal tech support may be deleted.
- Politics threads may be removed.
- No memes allowed as posts, OK to post as comments.
- Only approved bots from the list below, this includes using AI responses and summaries. To ask if your bot can be added please contact a mod.
- Check for duplicates before posting, duplicates may be removed
- Accounts 7 days and younger will have their posts automatically removed.
Approved Bots
founded 2 years ago
MODERATORS
These assholes really want Terminator to happen.
once they figured out that we weren't willing to murder each other on their behalf, their next moves became pretty obvious
Can I have 2 billion dollars to build UNIVERSE models?
You're selling yourself short. $1B for the world, $2B for the solar system, $4B for the milky way, $8B for the local group, $16B for the universe.
The language learning model has no awareness and will never understand the the real world. You're being sold propaganda from techno oligarchs who want to sell you their garbage data.
Not that I have any skin of the game or what you’re saying, but his entire pitch is to build something that’s not an LLM
LeCun argues that most human reasoning is grounded in the physical world, not language, and that AI world models are necessary to develop true human-level intelligence. “The idea that you’re going to extend the capabilities of LLMs [large language models] to the point that they’re going to have human-level intelligence is complete nonsense,” he said in an interview with WIRED.
I forget the technique he’s going after but it’s supposed to be different from the regular LLM.
EDIT: if I paid more attention I’d see it’s in the article. World models.
I don't see anything specific that validated his claims either.
Almost like we don't know what the fuck human intelligence actually is.
It’s a cool idea, but I don’t know if it’s $1B cool
six months later Come on guys, just another billion, we're almost there I swear
I mean, if you could actually solve that problem, it would be worth a LOT more than 1 billion. I just don't remotely believe that you can solve it for 1 billion. Training costs alone would eat that and more.
Best I can do is $87!