Comment by lukebuehler
4 days ago
If cannot the say they are "thinking", "intelligent" while we do not have a good definition--or, even more difficult, unanimous agreement on a definition--then the discussion just becomes about output.
They are doing useful stuff, saving time, etc, which can be measured. Thus also the defintion of AGI has largely become: "can produce or surpass the economic output of a human knowledge worker".
But I think this detracts from the more interesting discussion of what they are more essentially. So, while I agree that we should push on getting our terms defined, I think I'd rather work with a hazy definition, than derail so many AI discussion to mere economic output.
Heres a definition. How impressive is the output relative to the input. And by input, I don't just mean the prompt, but all the training data itself.
Do you think someone who has only ever studied pre-calc would be able to work through a calculus book if they had sufficient time? how about a multi-variable calc book? How about grad level mathematics?
IMO intelligence and thinking is strictly about this ratio; what can you extrapolate from the smallest amount of information possible, and why? From this perspective, I dont think any of our LLMs are remotely intelligent despite what our tech leaders say.
Hear, hear!
I have long thought this, but not had as good way to put it as you did.
If you think about geniuses like Einstein and ramanujen, they understood things before they had the mathematical language to express them. LLMs are the opposite; they fail to understand things after untold effort, training data, and training.
So the question is, how intelligent are LLMs when you reduce their training data and training? Since they rapidly devolve into nonsense, the answer must be that they have no internal intelligence
Ever had the experience of helping someone who's chronically doing the wrong thing, to eventually find they had an incorrect assumption, an incorrect reasoning generating deterministic wrong answers? LLMs dont do that; they just lack understanding. They'll hallucinate unrelated things because they dont know what they're talking about - you may have also had this experience with someone :)
> So the question is, how intelligent are LLMs when you reduce their training data and training? Since they rapidly devolve into nonsense, the answer must be that they have no internal intelligence
This would be the equivalent of removing all senses of a human from birth and expecting them to somehow learn things. They will not. Therefore humans are not intelligent?
> LLMs dont do that; they just lack understanding.
You have no idea what they are doing. Since they are smaller than the dataset, they must have learned an internal algorithm. This algorithm is drawing patterns from somewhere - those are its internal, incorrect assumptions. It does not operate in the same way that a human does, but it seems ridiculous to say that it lacks intelligence because of that.
It sounds like you've reached a conclusion, that LLMs cannot be intelligent because they have said really weird things before, and are trying to justify it in reverse. Sure, it may not have grasped that particular thing. But are you suggesting that you've never met a human that is feigning understanding in a particular topic say some really weird things akin to an LLM? I'm an educator, and I have heard the strangest things that I just cannot comprehend no matter how much I dig. It really feels like shifting goalposts. We need to do better than that.
1 reply →
You might be interested in reading about the minimum description length (MDL) principle [1]. Despite all the dissenters to your argument, what your positing is quite similar to MDL. It's how you can fairly compare models (I did some research in this area for LLMs during my PhD).
Simply put, to compare models, you describe both the model and training data using a code (usual reported as number of bits). The trained model that represents the data within the fewest number of bits is the more powerful model.
This paper [2] from ICML 2021 shows a practical approach for attempting to estimate MDL for NLP models applied to text datasets.
[1]: http://www.modelselection.org/mdl/
[2]: https://proceedings.mlr.press/v139/perez21a.html
Animals think but come with instincts which breaks the output relative to the input test you propose. Behaviors are essentially pre-programmed input from millions of years of evolution, stored in the DNA/neurology. The learning thus typically associative and domain-specific, not abstract extrapolation.
A crow bending a piece of wire into a hook to retrieve food demonstrates a novel solution extrapolated from minimal, non-instinctive, environmental input. This kind of zero-shot problem-solving aligns better with your definition of intelligence.
I'm not sure I understand what you're getting at. You seem to be on purpose comparing apples and oranges here: for an AI, we're supposed to include the entire training set in the definition of its input, but for a human we don't include the entirety of that human's experience and only look at the prompt?
> but for a human we don't include the entirety of that human's experience and only look at the prompt?
When did I say that? Of course you look at a human's experience when you judge the quality of their output. And you also judge their output based on the context they did their work in. Newton wouldn't be Newton if he was the 14th guy to claim that the universe is governed by three laws of motion. Extending the example I used above, I would be more impressed if an art student aced a tough calc test than a math student, given that a math student probably has spent much more time with the material.
"Intelligence and "thinking" are abstract concepts, and I'm simply putting forward a way that I think about them. It works very much outside the context of AI too. The "smartest" colleagues I've worked with are somehow able to solve a problem with less information or time than I need. Its usually not because they have more "training data" than me.
That an okay-ish definition, but to me this is more about whether this kind of "intelligence" is worth it, not whether it is intelligence itself. The current AI boom clearly thinks it is worth to put that much input to get the current frontier-model-level of output. Also, don't forget the input scales across roughly 1B weekly users at inference time.
I would say a good definition has to, minimally, take on the Turing test (even if you disagree, you should say why). Or in current vibe parlance, it does "feel" intelligent to many people--they see intelligence in it. In my book this allows us to call it intelligent, at least loosely.
There are plenty of humans that will never "get" calculus, despite numerous attempts at the class and countless hours of 1:1 tutoring. Are those people not intelligent? Do they not think? We could say yes they aren't, but by the metric of making money, plenty of people are smart enough to be rich, while college math professors aren't. And while that's a facile way of measuring someone's worth or their contribution to society (some might even say "bad"), it remains that even if someone cant understand calculus, some of them are intelligent enough to understand humans enough to be rich through some fashion that wasn't simply handed to them.
I don't think it's actually true that someone with:
1. A desire to learn calculus 2. A good teacher 3. No mental impairments such as dementia or other major brain drainers
could not learn calculus. Most people don't really care to try or don't get good resources. What you see as an intelligent mathematician is almost always someone born with better resources that was also encouraged to pursue math.
1 reply →
Yeah, that's compression. Although your later comments neglect the many years of physical experience that humans have as well as the billions of years of evolution.
And yes, by this definition, LLMs pass with flying colours.
I hate when people bring up this “billions of years of evolution” idea. It’s completely wrong and deluded in my opinion.
Firstly humans have not been evolving for “billions” of years.
Homo sapiens have been around for maybe 300’000 years, and the “homo” genus has been 2/3 million years. Before that we were chimps etc and that’s 6/7 million years ago.
If you want to look at the entire brain development, ie from mouse like creatures through to apes and then humans that’s 200M years.
If you want to think about generations it’s only 50/75M generations, ie “training loops”.
That’s really not very many.
Also the bigger point is this, for 99.9999% of that time we had no writing, or any kind of complex thinking required.
So our ability to reason about maths, writing, science etc is only in the last 2000-2500 years! Ie only roughly 200 or so generations.
Our brain was not “evolved” to do science, maths etc.
Most of evolution was us running around just killing stuff and eating and having sex. It’s only a tiny tiny amount of time that we’ve been working on maths, science, literature, philosophy.
So actually, these models have a massive, massive amount of training more than humans had to do roughly the same thing but using insane amounts of computing power and energy.
Our brains were evolved for a completely different world and environment and daily life that the life we lead now.
So yes, LLMs are good, but they have been exposed to more data and training time than any human could have unless we lived for 100000 years and still perform worse than we do in most problems!
12 replies →
This feels too linear. Machines are great at ingesting huge volumes of data, following relatively simple rules and producing optimized output, but are LLMs sufficiently better than humans at finding windy, multi-step connections across seemingly unrelated topics & fields? Have they shown any penchant for novel conclusions from observational science? What I think your ratio misses is the value in making the targeted extrapolation or hypothesis that holds up out of a giant body of knowledge.
Are you aware of anything novel, produced by an LLM?
For more on this perspective, see the paper On the measure of intelligence (F. Chollet, 2019). And more recently, the ARC challenge/benchmarks, which are early attempts at using this kind of definition in practice to improve current systems.
Is the millions of years of evolution part of the training data for humans?
Millions of years of evolution have clearly equipped our brain with some kind of structure (or "inductive bias") that makes it possible for us to actively build a deep understanding for our world... In the context of AI I think this translates more to representations and architecture than it does with training data.
1 reply →
The discussion about “AGI” is somewhat pointless, because the term is nebulous enough that it will probably end up being defined as whatever comes out of the ongoing huge investment in AI.
Nevertheless, we don’t have a good conceptual framework for thinking about these things, perhaps because we keep trying to apply human concepts to them.
The way I see it, a LLM crystallises a large (but incomplete and disembodied) slice of human culture, as represented by its training set. The fact that a LLM is able to generate human-sounding language
Not quite pointless - something we have established with the advent of LLMs is that many humans have not attained general intelligence. So we've clarified something that a few people must have been getting wrong, I used to think that the bar was set so that almost all humans met it.
Almost all humans do things daily that LLMs don't. It's only if you define general intelligence to be proficiency at generating text instead of successfully navigating the world while pursuing goals such as friendships, careers, families, politics, managing health.
LLMs aren't Data (Star Trek) or Replicants (Blade Runner). They're not even David or the androids from the movie A.I.
What do you mean? Almost every human can go to school and become a stable professional at some job, that is the bar to me, todays LLM cannot do that.
1 reply →
I think it has a practical, easy definition. Can you drop an AI into a terminal, give it the same resources as a human, and reliably get independent work product greater than that human would produce across a wide domain? If so, it's an AGI.
Doesn't sound like AGI without physical capabilities. It's not general if it's bound to digital work.
2 replies →
I agree that the term can muddy the waters, but as a shorthand for roughly "an agent calling an LLM (or several LLMs) in a loop producing similar economic output as a human knowledge-worker", then it is useful. And if you pay attention to the AI leaders, then that's what the defintion has become.
Personally I think that kind of discussion is fruitless, not much more than entertainment.
If you’re asking big questions like “can a machine think?” Or “is an AI conscious?” without doing the work of clarifying your concepts, then you’re only going to get vague ideas, sci-fi cultural tropes, and a host of other things.
I think the output question is also interesting enough on its own, because we can talk about the pragmatic effects of ChatGPT on writing without falling into this woo trap of thinking ChatGPT is making the human capacity for expression somehow extinct. But this requires one to cut through the hype and reactionary anti-hype, which is not an easy thing to do.
That is how I myself see AI: immensely useful new tools, but in no way some kind of new entity or consciousness, at least without doing the real philosophical work to figure out what that actually means.
I agree with almost all of this.
IMO the issue is we won't be able to adequately answer this question before we first clearly describe what we mean of conscious thinking applied to ourselves. First we'd need to clearly define our own consciousness and what we mean by our own "conscious thinking" in a much, much clearer way than we currently do.
If we ever reach that point, I think we'd be able to fruitfully apply it to AI, etc., to assess.
Unfortunately we haven't been obstructed from answering this question about ourselves for centuries or millennia, but have failed to do so, so it's unlikely to happen suddenly now. Unless we use AIs to first solve that problem of defining our own consciousness, before applying it back on them. Which would be a deeply problematic order, since nobody would trust a breakthrough in the understanding of consciousness that came from AI, that is then potentially used to put them in the same class and define them as either thinking things or conscious things.
Kind of a shame we didn't get our own consciousness worked out before AI came along. Then again, wasn't for the lack of trying… Philosophy commanded the attention of great thinkers for a long time.
I do think it raises interesting and important philosophical questions. Just look at all the literature around the Turing test--both supporters and detractors. This has been a fruitful avenue to talk about intelligence even before the advent of gpt.