A buddy of mine is way into AI.
So much so he refers to it/them as LLMs.
He runs into errors so the time too.
When he points it out the robot overlord thanks him and says something like “your were right to bring that to my attention”.
Yeah, it’ll all work out fine.
A few weeks ago some AI proved a math conjecture (the cycle double cover conjecture) that had gone unproven for decades. That’s interesting, and the proof was apparently clever, but to me the most interesting aspect of the exercise was that the AI tried to give up a few times, and its minder essentially kept telling it to try harder. So we’re apparently at the “congratulations, you have a really smart but inconsistently motivated teenager” stage of development.
Oh yeah, things are fine
I saw a video where a guy told an AI to count from x to y going by 2s.
It kept giving up like you described.
I’m not a naysayer at all. Just find theses things interesting.
This is mostly on OpenAI’s shoulders for not securing their test environment sufficiently, but it proves that really protecting against these things is devilishly difficult. There still seems be a lot of what I’ll call carelessness in this industry. If biology labs operated this cavalierly, we’d probably all be dead by now.
That’s what I’ve read/ heard: the interfaces run out to the public have not been fully vetted for hallucinations etc.
Other products with more important tasks than generating goofy pictures of your friends have been more carefully scrutinized.
And the more specific a task you give it, the more likely it is to give a worthwhile response like DAS demonstrated above.
I am for sure saying nay.
They make their users dumber, they can’t but make already obscenely wealthy people wealthier while destroying the jobs of working people, they’re terrible for the environment, and—since they learn by consuming the internet, and since the internet is increasingly a product of their own design, they can’t but get worse. And here you have municipalities and states falling over themselves to erect data centers. I don’t get it at all. It looks to me like biggest collective rakestepping in American history.
They are fantastic for programming and helping put reports together (of material I’ve created so I know what its supposed to be).
I get that they’re convenient. It just seems like a no-brainer to me that the convenience is coming at entirely too steep a cost. Entire generations of young people aren’t going to have a chance to live regular human lives.
It’s fun to listen to the Dodger broadcasters express surprise that Tucker could fall into a deep slump.
If any other pitcher had this balk called on him I’d have some sympathy. Since it happened to Gerrit Cole I find it extremely amusing.
I was at the gym this morning and ESPN was on the tube, and they had two Banana Ballers on from the Indianapolis (?) Clowns. Imagine my surprise at learning that one of them was Jackie Bradley, Jr. I still think that his beyond-belief performance was some sort of cosmic joke.
Was Steve Pearce one of his teammates?
He might be (what was up with that team?), but he wasn’t on the show. Something tells me Pearce lacks the showmanship gene necessary for that league.
Did Kalas just say “unchartered territory”?
Oh fuck me, Abreu warming in the pen.
Edited: Wrong Thread.
I wonder if, at some point, Mark will realize…
Ah, fuck me.
Mike Daddino’s take on this was pretty funny. For anyone unfamiliar with his work, he is a Mets fan with a Youtube channel and from what I can make of it, he just sits and yells at his computer screen all day.
If I was smarter I would be able to insert the vid… but there you go.