Comment by 0x20cowboy
6 months ago
LLM are a compressed version of their training dataset with a text based interactive search function.
6 months ago
LLM are a compressed version of their training dataset with a text based interactive search function.
Yes, but you're missing their ability to interpolate across that dataset at retrieval time, which is what makes them extremely useful. Also, people are willing to invest a lot of money to keep building those datasets, until nearly everything of economic value is in there.
not everything of economic value is data retrieval
Most economic value is not data retrieval
25 replies →
Because hypetrain.
Exactly I am so tired to hear about AI… And they are not even AI! I am also losing faith in this field when I see how much they all push so much hype and lies like this instead of being transparent. They are not AGIs not even AIs… For now they are only models and your definition is a good one
> they are not even AI!
From the "Artificial Intelligence" Wikipedia page: old-school Google search is AI, chess bots are AI, Google Maps is AI, the YouTube recommendation algorithm is AI...
LLMs are definitely AI.
LLMs are useful in that respect. As are media diffusion models. They've compressed the physics of light, the rules of composition, the structure of prose, the knowledge of the internet, etc. and made it infinitely remixable and accessible to laypersons.
AGI, on the other hand, should really stand for Aspirationally Grifting Investors.
Superintelligence is not around the corner. OpenAI knows this and is trying to become a hyperscaler / Mag7 company with the foothold they've established and the capital that they've raised. Despite that, they need a tremendous amount of additional capital to will themselves into becoming the next new Google. The best way to do that is to sell the idea of superintelligence.
AGI is a grift. We don't even have a definition for it.
I hate the "accessible to the layperson" argument.
People who couldn't do art before, still can't do art. Asking someone, or something else, to make a picture for you does not mean you created it.
And art was already accessible to anyone. If you couldn't draw something (because you never invested the time to learn the skill), then you could still pay someone else to paint it for you. We didn't call "commissioning a painting" as "being an artist", so what's different about "commissioning a painting from a robot?"
> I hate the "accessible to the layperson" argument.
Accessible to a layperson also means lowering the gradient slope of learning.
Millions of people who would have never rented a camera from a rental house are now trying to work with these tools.
Those publishing "slop" on TikTok are learning the Hero's Journey and narrative structure. They're getting schooled on the 180-degree rule. They're figuring out how to tell stories.
> People who couldn't do art before, still can't do art. Asking someone, or something else, to make a picture for you does not mean you created it.
Speak for yourself.
I'm not an illustrator, but I'm a filmmaker in the photons-on-glass sense. Now I can use image and video models to make animation.
I agree that your average Joe isn't going to be able to make a Scorsese-inspired flick, but I know what I'm doing. And for me, these tools open an entire new universe.
Something like this still takes an entire week of work, even when using AI:
https://www.youtube.com/watch?v=tAAiiKteM-U
There's lots of editing, rotoscoping, compositing, grading, etc. and the AI models themselves are INSANELY finicky and take a lot of work to finesse.
But it would take months of work if you were posing the miniatures yourself.
With all the thought and intention and work that goes into something like this, would you still say it "does not mean you created it"? Do you still think this hasn't democratized access to a new form of expression for non-animators?
AI is a creative set of tools that make creation easier, faster, more approachable, and more affordable. They're accessible enough that every kid hustling on YouTube and TikTok can now supercharge their work. And they're going to have to use these tools to even stay treading water amongst their peers, because if they don't use them, their competition (for time and attention) will.
I an not an expert but I have a serious counterpoint.
While training LLMs to replicate the human output, the intelligence and understanding EMERGES in the internal layers.
It seems trivial to do unsupervised training on scientific data, for instance, such as star movements, and discover closed-form analytic models for their movements. Deriving Kepler’s laws and Newton’s equations should be fast and trivial, and by that afternoon you’d have much more profound models with 500+ variables which humans would struggle to understand but can explain the data.
AGI is what, Artificial General Intelligence? What exactly do we mean by general? Mark Twain said “we are all idiots, just on different subjects”. These LLMs are already better than 90% of humans at understanding any subject, in the sense of answering questions about that subject and carrying on meaningful and reasonable discussion. Yes occasionally they stumble or make a mistake, but overall it is very impressive.
And remember — if we care about practical outcomes - as soon as ONE model can do something, ALL COPIES OF IT CAN. So you can reliably get unlimited agents that are better than 90% of humans at understanding every subject. That is a very powerful baseline for replacing most jobs, isn’t it?
Anthropomorphization is doing a lot of heavy lifting in your comment.
> While training LLMs to replicate the human output, the intelligence and understanding EMERGES in the internal layers.
Is it intelligence and understanding that emerges, or is applying clever statistics on the sum of human knowledge capable of surfacing patterns in the data that humans have never considered?
If this were truly intelligence we would see groundbreaking advancements in all industries even at this early stage. We've seen a few, which is expected when the approach is to brute force these systems into finding actually valuable patterns in the data. The rest of the time they generate unusable garbage that passes for insightful because most humans are not domain experts, and verifying correctness is often labor intensive.
> These LLMs are already better than 90% of humans at understanding any subject, in the sense of answering questions about that subject and carrying on meaningful and reasonable discussion.
Again, exceptional pattern matching does not imply understanding. Just because these tools are able to generate patterns that mimic human-made patterns, doesn't mean they understand anything about what they're generating. In fact, they'll be able to tell you this if you ask them.
> Yes occasionally they stumble or make a mistake, but overall it is very impressive.
This can still be very impressive, no doubt, and can have profound impact on many industries and our society. But it's important to be realistic about what the technology is and does, and not repeat what some tech bros whose income depends on this narrative tell us it is and does.
2 replies →
Indeed 90% problems can be solved by googling and that's what LLMs do. AGI is expected to be something more than a talking encyclopedia.
> Superintelligence is not around the corner. OpenAI knows this and is trying to become a hyperscaler / Mag7 company with the foothold they've established and the capital that they've raised.
+1 to this. I've often wondered why OpenAI is exploring so many different product ideas if they think AGI/ASI is less than a handful of years away. If you truly believe that, you would put all your resources behind that to increase the probability / pull-in the timelines even more. However, if you internally realized that AGI/ASI is much farther away, but that there is a technology overhang with lots of products possible on existing LLM tech, then you would build up a large applications effort with ambitions to join the Mag7.
LLMs require the sum total of human knowledge to ape what you can find on google, meanwhile Ramanujan achieved brilliant discoveries in mathematics using nothing but a grade school education and a few math books.
You phrase it as a diss but "Yeah LLM suck, they aren't even as smart as Ramanujan" sounds like a high praise to me.
Unfortunately LLMs fail even basic logic tests given to children so definitely not high praise. I'm just highlighting the absurd amount of data they need versus humans to highlight that they're just spitting out regressions on the training data. We're talking data that would take a human countless thousands of lifetimes to ingest. Yet a human can accomplish infinitely more with a basic grade school education.
4 replies →
yeah I've been thinking about them as stochastic content addressable memory. You can put as many next = userInput; while(true's) { next = mem[next]; } around them as you need in different forms. Single shot. Agents. etc and get wildly cool results out, but it's gated by some of the limitations there.