357
repeat after me: LLMs are not AI.
this post was submitted on 23 Nov 2024
357 points (89.2% liked)
Technology
59708 readers
4228 users here now
This is a most excellent place for technology news and articles.
Our Rules
- Follow the lemmy.world rules.
- Only tech related content.
- Be excellent to each another!
- 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, to ask if your bot can be added please contact us.
- Check for duplicates before posting, duplicates may be removed
Approved Bots
founded 1 year ago
MODERATORS
LLMs are one version of AI. It's just one tiny part of AIs that are used every day, from chess bots to voice transcription, but they also are AI.
I would replace the word version with aspect. LLMs are merely one part of the puzzle that would be AI. Essentially what's been constructed is the mouth and the part of the brain that can form words but without any of the reasoning or intelligence behind what the mouth says.
The same goes for the art AIs. They can paint pictures based on input but they can't reason how those pictures should look. Which is why it requires so much tweaking to get them to output something that doesn't look like it came out of a Lovecraft novel.
load more comments
(3 replies)
I think you are confusing AI with AGI.
https://en.m.wikipedia.org/wiki/Artificial_general_intelligence
Not at all. AI is something that uses rules, not statistical guesswork. A simple control loop is alreadu basic AI, but the core mechanism of LLMs is not (the parts before and after token association/prediction are). Don't fall for marketing bullshit of some dumbass silicon valley snake oil vendors.
It's pretty obvious that they will hit a ceiling.
Quick buck is over. And now it's time again for base research to create better approach.
I really wish we had a really advanced AI with reasonable resource consumption within my lifetime. I don't think it's unreasonable as we have got really far in the last 30 years of computational technology.
I really wish we had a really advanced AI with reasonable resource consumption within my lifetime.
You only wish that for as long as it doesn't happen. Have you looked at the world we live in? Such tools would be controlled by the same billionaire dipshits for their personal gain as all social media is being used already.
load more comments
(8 replies)
We've come a long way in computing, but the computational power difference between a human brain and a computer is significant. LLMs were just a smart way to have computers learn pattern recognition. While important, it isn't anything close to artificial general intelligence (AGI), which is what the term AI usually means.
Yeah. AI may grind for a while but hardly anyone has put the current stuff to work, yet. We will be feeling the benefits of what is released right now for a decade to come. I am working on a very rudimentary application that will use ML at work and it won't come out for 12 more months, and it hardly does anything but make the most obvious decisions 10m times faster than I can. But it's going to fundamentally change our labor model.
There are regular folks applying amazing technologies that go way beyond content generation.
The tech may grind but the application of that tech is barely getting its feet and should run hard for a decade.
You’re not using a typewriter.
How to tell someone uses Hacker News without asking:
The problem isn't with the AI. It's with how it's being treated. It's currently being sold as if it were general intelligence. Which it's not. It should instead be treated like it's a mindless tool. Something that is inert on its own. Useful for some things but only in a limited sense. Unfortunately the companies, who have spent millions of dollars developing these things, are trying to sell it as the "do-all" artificial intelligence that people have grown up seeing in sci-fi media. Which it 100% is not.
Every company have always oversell their own products. This is not new.
Coca Cola is also just a carbonated sweet drink and it's being sold as happiness, socialization and the meaning of Christmas in a bottle.
Companies oversell, it's called marketing. It's shit practice but it's not nothing new.
That does not make the technology worse (or better). Current AI technology has its uses. With a big problem in how resource hungry it is. But it's fairly useful.
Your point is valid. Companies do use marketing to sell their products by using lots of outrageous claims. And my problem isn't really what the companies it's mostly with the people who are buying that bullshit.
P.S. your Coca-Cola example would have been better if you had reached back to their origins when they were sold as a tonic that cures just about everything. What they're being sold as right now is just a soda and all of the current marketing around it is just nostalgia bait which everybody uses for everything especially around Christmas time.
So long and thanks for all the fish habitat?
view more: next ›