A beautiful explanation of what LLMs cannot do. Choice sneer:

If you covered a backhoe with skin, made its bucket look like a hand, painted eyes on its chassis, and made it play a sound like “hnngghhh!” whenever it lifted something heavy, then we’d start wondering whether there’s a ghost inside the machine. That wouldn’t tell us anything about backhoes, but it would tell us a lot about our own psychology.

Don't have time to read? The main point:

Trying to understand LLMs by using the rules of human psychology is like trying to understand a game of Scrabble by using the rules of Pictionary. These things don’t act like people because they aren’t people. I don’t mean that in the deflationary way that the AI naysayers mean it. They think denying humanity to the machines is a well-deserved insult; I think it’s just an accurate description.

I have more thoughts; see comments.

no comments (yet)

sorted by: hot top controversial new old
there doesn't seem to be anything here
3 comments in c/technology@lemmy.world view that post
[–] 12 points 11 months ago (2 children)

(That’s also why I see no point in using AI to, say, write an essay, just like I see no point in bringing a forklift to the gym. Sure, it can lift the weights, but I’m not trying to suspend a barbell above the floor for the hell of it. I lift it because I want to become the kind of person who can lift it. Similarly, I write because I want to become the kind of person who can think.)

This is one of the best quotes that I can grab from the page.

The second being:

Markus Strasser, an entrepreneur who tried to start one of those companies that’s like “we’ll put every scientific paper in the bag and then ??? and then profit”, eventually abandoned the effort, saying that “close to nothing of what makes science actually work is published as text on the web.”

and last one:

For instance, if you show humans computer code that has security vulnerabilities, they do not suddenly start praising Hitler. But LLMs do. So yes, I would worry about putting the nuclear codes in the bag.

The paper cited in it is amazing, in the sense of confusingly surprising. Gonna give it a read.

  • source
  • hideshow 2 child comments
  • [–] 4 points 11 months ago (1 child)

    The thing I see LLMs more useful for is searching.

    Imagine asking a robot where a book with certain contents is. I wish for that instead of a robot that does my homework, because I suck a little with searching.

    I also saw them useful for science: putting all selected, finetuned scientific articles and papers into an LLM to see what it can do on a specific scientific field. So imagine my surprise seeing this is only doable for private associations, since there is nothing "good" on the public web.

  • source
  • parent
  • hideshow 1 child comment
  • [–] 2 points 11 months ago

    that would be great, but they'd have to redo it a bit, so you can peruse the sources to the text it produces from your prompt. I don't think they store that info, but that would be super awesome.

  • source
  • parent