this post was submitted on 19 Nov 2023
496 points (86.9% liked)
Technology
58011 readers
3069 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
you are viewing a single comment's thread
view the rest of the comments
view the rest of the comments
At this point I want a calendar of at what date people say "AI could never" - like "AI could never explain why a joke it's never seen before is funny" (such as March 2019) - and at what date it happens (in that case April 2022).
(That "explaining the joke" bit is actually what prompted Hinton to quit and switch to worrying about AGI sooner than expected.)
I'd be wary of betting against neural networks, especially if you only have a casual understanding of them.
I mean the limitations of LLMs are very well documented, they aren't going to advance a whole lot more without huge leaps in computing technology. There are limits on how much context they can store for example, so you aren't going to have AIs writing long epic stories without human intervention. And they're fundamentally incapable of originality.
General AI is another thing altogether that we're still very far away from.
Nearly everything you wrote is incorrect.
As an example, rolling context windows paired with RAG would easily allow for building an implementation of LLMs capable of writing long stories.
And I'm not sure where you got the idea that they were fundamentally incapable of originality. This part in particular tells me you really don't know how the tech is working.
A rolling context window isn't a real solution and will not produce works that even come close to matching the quality of human writers. That's like having a writer who can only remember the last 100 pages they wrote.
The tech is trained on human created data. Are you suggesting LLMs are capable of creativity and imagination? Lmao - and you try to act like I'm the one who's full of shit.
That's why you pair it with RAG.
They are trained by iterating through network configurations until there's diminishing returns on how accurately they can complete that human created data.
But they don't just memorize the data. They develop the capabilities to extend it.
So yes, they absolutely are capable of generating original content that's not in the training set. As has been demonstrated over and over. From explaining jokes not found in the training data, solving riddles not found in it, or combining different concepts to result in a new synthesis not found in the original data.
What do you think it's doing? Copy/pasting or something?
Now.
A year ago even the boring stuff was impossible. Six months ago it'd do everything okay but fumble the details. Today? Sometimes the only reason AI art stands out is that models like central framing and eye contact.
Six months from now, I don't know and you don't either. You can posture about the indomitable human et cetera and ignore how it mirrors past declarations about what's possible right now. The joke goes, "AI is whatever hasn't been done yet." And buddy, that category never gets bigger.
If there are rules, a deep enough network can discern them. And we are a lot less complex than we'd like to think.
I'm only describing the present. You're the one saying you'll always always always be able to tell. Feel free to test that bravado, internet stranger.
The simple fact is, we don't understand intelligence, and we keep being wrong about how much can be faked. Sometimes by underestimating ourselves - sometimes by overestimating ourselves. We do all this civilization nonsense with three pounds of electrified meat. Some parts, even important or popular parts, can be very formulaic. People have been generating so-so sheet music since an electric calculator was an investment. Now we're talking about direct audio output of dead singers with new lyrics from dead writers. No kidding it's flawed. But it's already jawdropping when it works properly, and there is no reason to imagine this rush of progress will suddenly stop.
Really, say that out loud: "this is the best technology will ever be."
I'm not convinced you'd recognize yourself in a mirror.
You're aggressively dismissive of cocksure predictions... which you're doing, and I'm not.
You're incensed by language making disagreement personal... which you used, and I only turned around.
You think "I don't know" is damning failure, but "you don't either" is somehow contradictory.
You are the one certain of what's going to happen.
That is somehow not the most absurd part of this conversation.
Naked projection and complete nonsense.
Success is impossible. You're not listening to yourself and you don't care what words mean.
Meanwhile, keep an eye on this rapidly-changing technology, and try not to act surprised if your absolute certainty turns out to be misplaced.
"For the new and special, humans will always be required. End of line."
But I'm the one making predictions. By saying "I don't know and you don't either."
"All of the bravado of all of the Internet strangers isn’t going to suddenly make me 'believe.'"
When I never asked you to believe anything, besides human fallibility.
"Tough guy talk from internet strangers only impresses children on Xbox live."
In response to an invitation to check your assumptions with a survey of examples... using the 'bravado of internet strangers' phrase you had just used.
"Do you often find this approach successful when the bloviating and juvenile attempts at intimidation don’t work?"
Which is outright fantasy, considering I cannot even guess what part of this was supposed to be "intimidation."
And now you want to pretend I'm the one who threw a tantrum and started with insults.
What a sincere and worthwhile conversation this has been.
Meanwhile:
Yeah, you will believe it, when you see it. All I ask is that you remember you said it was literally impossible. But that's a tall order when you can't remember what you said two comments prior.
It's your comments, troll.
It's a detailed effort to do the opposite of what you keep lying about.
But you don't care what's true. You just want to feel innately superior.
You are the one offended - by repetition of your own words.
You don't even know my opinion. You refuse to read anything. You just make shit up.
One of us has been trying to bring this conversation to a point, and it was never you.
I wrote more. You didn't read it. You were in fact a childish troll about not reading it.
You keep projecting and it's getting dull.
I keep hoping you'll care what words mean, guy who also keeps replying to blunt criticism. Unknowing hypocrite.
How this went is, you keep calling me names, and I keep pointing to the obvious hypocrisy in that.
But you don't care what words mean. You're stuck in a loop. You're never going to care about the point, no matter what anyone says, because you've got your conclusion and any challenge to it must be a personal insult.
You are a broken individual and a waste of everyone's time. Good luck.