Comment by sfn42
15 hours ago
> we teach machines to learn
I feel like this statement is too anthropomorphizing. We don't teach a computer anything. We write normal computer programs that can process data(training), turning it into what is essentially a configuration file(weights/parameters) for another normal computer program(inference engine). Upon receiving a request, the inference engine is invoked and passed the input, then it processes that input and its configuration(weights), producing an output. "Thinking" is just feeding the output back in as an input, allowing the model to revise it and refine it before it's returned to the user.
There is no teaching, nor learning. There is just data processing. There is nothing happening between requests to the inference engine. It's a relatively simple system altogether, but when the training data set is large enough it produces unbelievably good outputs.
With this in mind I really don't understand how people think these programs can be conscious. They're not doing anything fundamentally different than any other computer program. It's just crunching numbers. You might as well argue that a computer game or 7zip or a chess bot or whatever is conscious.
I know you specifically didn't claim they're conscious, I just thought it was worth mentioning because I think it's ridiculous how many people will seriously argue this as if there is any difference between "ai" and other programs. There isn't. It's just a set of instructions for doing computation, as computers have always done.
3D rendering is actually quite similar. There is a bunch of files (input/configuration) describing 3D meshes, textures, lighting etc which are fed into a rendering engine which processes all that data and turns it into a 2D image. Similarly to AI, once the input files are sophisticated enough, this can produce mindblowingly realistic and beautiful images, and you can run it in a loop to make video or make it interactive, moving the camera around in real time etc. But at the core of it all is just a bunch of files containing a bunch of numbers, and a program that reads those files and does computation with them.
The current leading materialistic model of consciousness is essentially that human consciousness is a computation in itself, and that the brain is basically a very very complex computer. So it's not at all illogical for someone who holds this view and who fully understands that AI is just a regular computer program to also believe that it could be conscious.
Of course, I'm not saying that you have to believe any of this - even defining what consciousness is is famously elusive. And there are a lot of people who think consciousness is not materialistic to begin with, or that we have not yet discovered the scientific nature of it.
The computational theory of mind is the leading theory for pure cognition. There is no leading theory of consciousness, although some people also favor the computational theory of mind for that as well but this is far less popular.
I hope we never do prove that, because as soon as that happens it will be used for eugenics.
I definitely agree that consciousness is mysterious. But I think the assumption that AI is conscious begs the question "then why isn't 7zip conscious?".
Is a simple calculator conscious? How about a mechanical calculator? Is an abacus conscious? Are they only conscious when executing a sufficiently sophisticated program? What if I hypothetically had infinite time and patience, and performed the inference on paper with a pencil. Would it then be conscious?
Well, the exact same question applies to biology. If humans are conscious, is all life conscious? If it's just the brain that's conscious, is every part of the brain conscious? Are individual neurons conscious? How about other cells in your body, or tissues or organs? Our stomach has more cells and more neurons than many full animals - is our stomach conscious?
The answer in the "consciousness is a computation" model is that consciousness can arise in any computational medium. But the model doesn't say that any computation is conscious - it says that consciousness is any of a set of specific computations, any computation with properties X, Y, Z whatever. So it's perfectly possible (according to this model) that a GPU running game graphics is not conscious, while a GPU running some future even more advamced LLM is conscious. Or a network of GPUs + storage + whatever.
Good examples. We don't know what human intelligence, understanding or consciousness is but we do know what all of those examples are. And we do know what an LLM is. Like you said, it's a data processing machine and we know what algorithms it uses. It's more productive to discuss it in those engineering terms because we've done plenty of data processing, statistics, control systems, curve-fitting etc. before. There's no need to make this jump to consciousness/magic except for marketing or fantasy purposes.
We humans have the habit of falling into the trap of anthropocentric views and explanations of the universe and everything within it. The history of AI is the history of moving goalposts. I don't particularly care if they are conscious or not. Humans have a history of treating even actual, conscious, intelligent beings as chattel or food. So I don't really see why our attitudes would change if we somehow realized that LLMs are conscious.
I have been thinking about this for a while in the context of the quest for AGI. If all we want is something to take over some parts of our intellectual labor, then LLMs are already there. They can communicate with you in a language you understand and do stuff that you command them to. Often, on their own.
AGI will have its own telos. An LLM does not (at least I refuse to believe that it does).
So, yeah, you could think of LLMs as programs processing data. But there is definitely some emergent behavior you notice that emerges when the LLM runs. How you interpret all this depends on your view on consciousness and intelligence.
I also don't really care whether they're conscious, but I do think it's worth challenging the claim that they are.
LLMs are programs processing data. Lots of programs exhibit emergent behavior. A simple example is Conway's game of life. It can be programmed in less than an hour, exhibits lots of interesting behavior. In fact you could run LLMs inside of it if you wanted to.
I'm not denying that LLMs are unbelievably cool and useful. It still blows my mind when I send a half-assed complex screenshot and a vague description and it just gets it and fixes the issue in seconds. It's like magic. But it isn't magic, it's just some relatively simple algorithms processing a crapload of data, and I think it's useful to understand and remember that.
> "There is nothing happening between requests to the inference engine." This is true for vLLM/SGLang/etc but LayerScale is making use of idle GPUs. A new novel inference engine for agentic workloads. Our advanced scheduler enables double the workload on the same number of GPUs (e.g. prioritization of on-demand queries and background processing of batch workloads simultaneously). Let me know if you want to try it out!
> I feel like this statement is too anthropomorphizing.
That's number one mistake preventing people from getting a good grasp of how LLMs work and how to use them effectively.
That is, not anthropomorphizing is the mistake. It gives people bad mental models.
LLMs are not people in general, but behaviorally and in terms of system design, they may as well be. They have same failure modes and cognitive characteristics humans have, and are very much unlike regular, non-ML software tools.
Not really sure what you mean, I have the view I presented and I'm able to use LLMs effectively. The understanding that they're not human allows me to treat them as such - I don't say please or sorry or thanks or other pleasantries, I just instruct them in the simplest most efficient way I can. Humans would find me rude, be annoyed if I give them lots of work for little benefit, and say the same things over and over etc.