The "just" is doing all the lifting. You can reductively describe any information processing system in a way that makes it sound like it couldn't possibly produce the outputs it demonstrably produces. "The sun is just hydrogen atoms bumping into each other" is technically accurate and completely useless as an explanation of solar physics.
Edit: Case in point, a mere 10 minutes later we got someone making that exact argument in a sibling comment to yours! Nature is beautiful.
This is a thought-terminating cliche employed to avoid grappling with the overwhelming differences between a human brain and a language model.
Or maybe there's even a medium term scratchpad that is managed automatically, just fed all context as it occurs, and then a parallel process mulls over that content in the background, periodically presenting chunks of it to the foreground thought process when it seems like it could be relevant.
All I'm saying is there are good reasons not to consider current LLMs to be AGI, but "doesn't have long term memory" is not a significant barrier.
Its even more ridiculous than me pretending I understand how a rocket ship works because I know there is fuel in a tank and it gets lit on fire somehow and aimed with some fins on the rocket...
> I have worked in a startup wherein we heavily finetuned Deepseek, among other smaller models, running on our own hardware.
Are you serious with this? I could go make a lora in a few hours with a gui if I wanted to. That doesn't make me qualified to talk about top secret frontier ai model architecture.
Now you have moved on to the guy who painted his honda, swapped out some new rims, and put some lights under it. That person is not an automotive engineer.
Intelligence is the ability to reason about logic. If 1 + 1 is 2, and 1 + 2 is 3, then 1 + 3 must be 4. This is deterministic, and it is why LLMs are not intelligent and can never be intelligent no matter how much better they get at superficially copying the form of output of intelligence. Probabilistic prediction is inherently incompatible with deterministic deduction. We're years into being told AGI is here (for whatever squirmy value of AGI the hype huckster wants to shill), and yet LLMs, as expected, still cannot do basic arithmetic that a child could do without being special-cased to invoke a tool call.
Our computer programs execute logic, but cannot reason about it. Reasoning is the ability to dynamically consider constraints we've never seen before and then determine how those constraints would lead to a final conclusion. The rules of mathematics we follow are not programmed into our DNA; we learn them and follow them while our human-programming is actively running. But we can just as easily, at any point, make up new constraints and follow them to new conclusions. What if 1 + 2 is 2 and 1 + 3 is 3? Then we can reason that under these constraints we just made up, 1 + 4 is 4, without ever having been programmed to consider these rules.This is not even wrong.
>Probabilistic prediction is inherently incompatible with deterministic deduction.
And his is just begging the question again.
Probabilistic prediction could very well be how we do deterministic deduction - e.g. about how strong the weights and how hot the probability path for those deduction steps are, so that it's followed every time, even if the overall process is probabilistic.
Probabilistic doesn't mean completely random.
https://en.wikipedia.org/wiki/Not_even_wrong
Personally I think not even wrong is the perfect description of this argumentation. Intelligence is extremely scientifically fraught. We have been doing intelligence research for over a century and to date we have very little to show for it (and a lot of it ended up being garbage race science anyway). Most attempts to provide a simple (and often any) definition or description of intelligence end up being “not even wrong”.
Human Intelligence is clearly not logic based so I'm not sure why you have such a definition.
>and yet LLMs, as expected, still cannot do basic arithmetic that a child could do without being special-cased to invoke a tool call.
One of the most irritating things about these discussions is proclamations that make it pretty clear you've not used these tools in a while or ever. Really, when was the last time you had LLMs try long multi-digit arithmetic on random numbers ? Because your comment is just wrong.
>What if 1 + 2 is 2 and 1 + 3 is 3? Then we can reason that under these constraints we just made up, 1 + 4 is 4, without ever having been programmed to consider these rules.
Good thing LLMs can handle this just fine I guess.
Your entire comment perfectly encapsulates why symbolic AI failed to go anywhere past the initial years. You have a class of people that really think they know how intelligence works, but build it that way and it fails completely.
They still make these errors on anything that is out of distribution. There is literally a post in this thread linking to a chat where Sonnet failed a basic arithmetic puzzle: https://news.ycombinator.com/item?id=47051286
> Good thing LLMs can handle this just fine I guess.
LLMs can match an example at exactly that trivial level because it can be predicted from context. However, if you construct a more complex example with several rules, especially with rules that have contradictions and have specified logic to resolve conflicts, they fail badly. They can't even play Chess or Poker without breaking the rules despite those being extremely well-represented in the dataset already, nevermind a made-up set of logical rules.
I thought we were talking about actual arithmetic not silly puzzles, and there are many human adults that would fail this, nevermind children.
>LLMs can match an example at exactly that trivial level because it can be predicted from context. However, if you construct a more complex example with several rules, especially with rules that have contradictions and have specified logic to resolve conflicts, they fail badly.
Even if that were true (Have you actually tried?), You do realize many humans would also fail once you did all that right ?
>They can't even reliably play Chess or Poker without breaking the rules despite those extremely well-represented in the dataset already, nevermind a made-up set of logical rules.
LLMs can play chess just fine (99.8 % legal move rate, ~1800 Elo)
https://arxiv.org/abs/2403.15498
I don‘t like to throw the word intelligence around, but when we talk about intelligence we are usually talking about human behavior. And there is nothing human about being extremely good at curve fitting in multi parametric space.
What you probably mean is that it is not a mind in the sense that it is not conscious. It won't cringe or be embarrassed like you do, it costs nothing for an LLM to be awkward, it doesn't feel weird, or get bored of you. Its curiosity is a mere autocomplete. But a child will feel all that, and learn all that and be a social animal.
Whereas the child does what exactly, in your opinion?
You know the child can just as well to be said to "just do chemical and electrical exchanges" right?
The comparison is therefore annoying
I see your "flat plane of silicon" and raise you "a mush of tissue, water, fat, and blood". The substrate being a "mere" dumb soul-less material doesn't say much.
And the idea is that what matters is the processing - not the material it happens on, or the particular way it is.
Air molecules hitting a wall and coming back to us at various intervals are also "vastly different" to a " matrix multiplication routine on a flat plane of silicon".
But a matrix multiplication can nonetheless replicate the air-molecules-hitting-wall audio effect of reverbation on 0s and 1s representing the audio. We can even hook the result to a movable membrane controlled by electricity (what pros call "a speaker") to hear it.
The inability to see that the point of the comparison is that an algorithmic modelling of a physical (or biological, same thing) process can still replicate, even if much simpler, some of its qualities in a different domain (0s and 1s in silicon and electric signals vs some material molecules interacting) is therefore annoying.
"Annoying" does not mean "false".
Aside from a priori bias, this assumption of absurdity is based on what else exactly?
Biological systems can't be modelled (even if in a simplified way or slightly different architecture) "with silicon arrangements", because?
If your answer is "scale", that's fine, but you already conceded to no absurdity at all, just a degree of current scale/capacity.
If your answer is something else, pray tell, what would that be?
Any definition of intelligence that does not axiomatically say "is human" or "is biological" or similar is something a machine can meet, insofar as we're also just machines made out of biology. For any given X, "AI can't do X yet" is a statement with an expiration date on it, and I wouldn't bet on that expiration date being too far in the future. This is a problem.
It is, in particular, difficult at this point to construct a meaningful definition of intelligence that simultaneously includes all humans and excludes all AIs. Many motivated-reasoning / rationalization attempts to construct a definition that excludes the highest-end AIs often exclude some humans. (By "motivated-reasoning / rationalization", I mean that such attempts start by writing "and therefore AIs can't possibly be intelligent" at the bottom, and work backwards from there to faux-rationalize what they've already decided must be true.)
Good thing I didn't make that claim!
> Ignoring refutations you don't like doesn't make them wrong.
They didn't make a refutation of my points. They asserted a basic principle that I agreed with, but assume acceptance of that principle leads to their preferred conclusion. They make this assumption without providing any reasoning whatsoever for why that principle would lead to that conclusion, whereas I already provided an entire paragraph of reasoning for why I believe the principle leads to a different conclusion. A refutation would have to start from there, refuting the points I actually made. Without that you cannot call it a refutation. It is just gainsaying.
> Any definition of intelligence that does not axiomatically say "is human" or "is biological" or similar is something a machine can meet, insofar as we're also just machines made out of biology.
And here we go AGAIN! I already agree with this point!!!!!!!!!!!!!!! Please, for the love of god, read the words I have written. I think machine intelligence is possible. We are in agreement. Being in agreement that machine intelligence is possible does not automatically lead to the conclusion that the programs that make up LLMs are machine intelligence, any more than a "Hello World" program is intelligence. This is indeed, very repetitive.
If you are prepared to accept that intelligence doesn't require biology, then what definition do you want to use that simultaneously excludes all high-end AI and includes all humans?
By way of example, the game of life uses very simple rules, and is Turing-complete. Thus, the game of life could run a (very slow) complete simulation of a brain. Similarly, so could the architecture of an LLM. There is no fundamental limitation there.
I literally did provide a definition and my argument for it already: https://news.ycombinator.com/item?id=47051523
If you want to argue with that definition of intelligence, or argue that LLMs do meet that definition of intelligence, by all means, go ahead[1]! I would have been interested to discuss that. Instead I have to repeat myself over and over restating points I already made because people aren't even reading them.
> Not even that current models are not; you seem to be claiming that they cannot be.
As I have now stated something like three or four times in this thread, my position is that machine intelligence is possible but that LLMs are not an example of it. Perhaps you would know what position you were arguing against if you had fully read my arguments before responding.
[1] I won't be responding any further at this point, though, so you should probably not bother. My patience for people responding without reading has worn thin, and going so far as to assert I have not given an argument for the very first thing I made an argument for is quite enough for me to log off.
Human brains run on probabilistic processes. If you want to make a definition of intelligence that excludes humans, that's not going to be a very useful definition for the purposes of reasoning or discourse.
> What if 1 + 2 is 2 and 1 + 3 is 3? Then we can reason that under these constraints we just made up, 1 + 4 is 4, without ever having been programmed to consider these rules.
Have you tried this particular test, on any recent LLM? Because they have no problem handling that, and much more complex problems than that. You're going to need a more sophisticated test if you want to distinguish humans and current AI.
I'm not suggesting that we have "solved" intelligence; I am suggesting that there is no inherent property of an LLM that makes them incapable of intelligence.