I’ve been re reading my response and my bad, I meant “artificial neurons were inspired from neurons”, not to behave like, they have little in common.

If you were to train an LLM on nothing but math and texts about math, then asked it an art question, it would respond somewhat nonsensically with math.

If you asked an human that speak german and nothing else, a question in english, it would also respond in german (that they cant understand you).
LLMs sometimes (not often enough) do respond they don’t know.

that suggests to me either you know how it was trained, or are making assumptions about how it was trained.

The general way it is trained is known, specifics and technics are not known, but the public do know how one of the flagship model was trained, the training process of deepseek r1 was documented in their research paper: https://arxiv.org/pdf/2501.12948

Which I did read a chunk when it was released.

And if the idea that there would be nobody training an LLM to do addition in this manner, I invite you to glance the Wikipedia article on addition.

The LLMs have multiple way to do additions, i’ll showcase two as an example, I asked ChatGPT 4.1 to solve a big addition. Here it’s output:

 I need to add these two large numbers carefully.

First, I'll write them one under the other, aligned at the right:

                  21,500,996,531,314,339
+ 96,357,764,557,683,165,397,721
----------------------------------


Working from right to left:

Units place: 9 + 1 = 10. Write 0, carry 1.
Tens place: 3 + 2 + 1(carry) = 6. Write 6, no carry.
Hundreds place: 3 + 7 = 10. Write 0, carry 1.
Thousands place: 4 + 7 + 1(carry) = 12. Write 2, carry 1.
Ten thousands place: 1 + 9 + 1(carry) = 11. Write 1, carry 1.
Hundred thousands place: 3 + 3 + 1(carry) = 7. Write 7, no carry.
Millions place: 1 + 5 = 6. Write 6, no carry.
Ten millions place: 5 + 6 = 11. Write 1, carry 1.
Hundred millions place: 3 + 1 + 1(carry) = 5. Write 5, no carry.
Billions place: 5 + 3 = 8. Write 8, no carry.
Ten billions place: 9 + 8 = 17. Write 7, carry 1.
Hundred billions place: 9 + 6 + 1(carry) = 16. Write 6, carry 1.
Trillions place: 0 + 7 + 1(carry) = 8. Write 8, no carry.
Ten trillions place: 0 + 5 = 5. Write 5, no carry.
Hundred trillions place: 5 + 5 = 10. Write 0, carry 1.
Quadrillions place: 1 + 4 + 1(carry) = 6. Write 6, no carry.
Ten quadrillions place: 2 + 6 = 8. Write 8, no carry.

For the remaining places, I just copy the digits from the second number:
7, 7, 5, 3, 6, 9

Therefore, the answer is:
96,357,786,056,867,851,761,260

You can notice, the whole reasoning is correct, but it wrote the wrong response, I can expand more on this if you want (I do some research on it on my free time)
This reasoning of decomposing the addition was of course learned from training data.
Now, the trigonometry used to calculate additions that i talked earlier, is not for writing a “reasoning” but when it try to write the correct response. It was created by the backpropagation trying to find a local minimum that can solve additions in order to more accuratly predict the next token.

so I would point out that technically LLMs have “tensors” not “neurons”.
I get that tensors are designed to behave like neurons, and this is just me being pedantic. I know what you mean when you say neurons, just wanted to clarify and be consistent. No shade intended.

Artificial neurons were made to behave like neurons: https://en.wikipedia.org/wiki/Artificial_neuron
And the terminology used, is neurons, cf the paper i sent earlier about how they do additions: https://arxiv.org/pdf/2502.00873

I don’t think you can disconnect how an LLM was trained from how it operates

You can, heck the example I gave show exaclty this:

If you train an LLM to use trigonometry to solve addition problems, I think you will find the LLM will do trigonometry to solve addition problems.

It was not trained to do trigonometry to solve addition problem, it was trained to respond to additions, trigonometry is how the statiscal part, the backpropagation, found a way to make the neurons solve additions.

In general if that is how the LLM is coming to its next token, then the training data must be really heavily weighted in that manner.

You are mixing up stuff, the way LLM are trained does not impose anything about how the neurons gets organised to get better score at inferrence.

You are mistaking how LLMs are trained to how they work.
It’s not because it’s been trained with statistics, that they compute, or think using statistics.
For example, to do additions, internally LLMs do trignonometry: https://arxiv.org/abs/2502.00873
They do probably use statistics for tons of stuff internally, but humans do too: guessing, bias, tendency, preferences.
Anthropics researcher found that their LLMs have “features” for concepts.

So you think you need words to be able to think ? Monkeys, birds, human babies are unable to think then ?

Is the argument that LLMs are thinking because they make guesses

No, it’s that you can’t root the argument that they don’t think over the fact they make stuff up, because humans too. You could root it in the amount of things it guess wrong, but it’s extremely hard to measure.
Again, I’m not claiming that they think, but that we don’t know until one or the other is proven.
Right now, thinking one, or the other is true, is belief.

How did you concluded that from theses 2 messages.

Consciousness may be an illusion born from the ability of self reflection.
Also, like i showed before, you may act before consciously taking the decision of it.
https://en.m.wikipedia.org/wiki/Neuroscience_of_free_will
Theses study with the one presented by cgpgray, indicate that maybe we do stuff then we come up with a reasonable explanation after.

And the brain is made out of neurons that sends electric signals between them and operate muscles.
That doesnt explain how the brain think.

I’m saying that’s fine, but it should be able to reason that it doesn’t know the answer, and say that.

That is of course a big problem. They try to guess too much stuff, but it’s also why it kinda works. Symbolics AI have the opposite problem, they are rarely useful, because they can’t guess stuff, they are rooted in hard logic, and cannot come up with a reasonable guess.
Now humans also try to guess stuff and sometimes get it wrong, it’s required in order to produce results from our thinking and not be stuck in a state where we don’t have enough data to do anything, like a symbolic AI.

Now, this is becoming a spectrum, humans are somewhere in the middle of LLMs and symbolics AI.
LLMs are not completely unable to say what they know and doesnt know, they are just extremely bad at it from our POV.

The probleme with “does it think” is that it doesn’t give any quantity or quality.

I think when you ask an LLM how many 'r’s there are in Strawberry and questions along this line you can see they can’t form judgments.

Like a LLMs you are making the wrong affirmation based lacking knowledge.
Current LLMs input, and output tokens, they dont ever see the individual letters, they see tokens, for straberry, they see 3 tokens:

They dont have any information on what characters are in this tokens. So they come up with something. If you learned a language only by speaking, you’ll be unable to write it down correctly (except purely phonetical systems), instead you’ll come up with what you think the word should be written.

I would also add that if you “form judgments” you probably don’t need to be reminded you formed a judgment immediately after forming one.

You come up with the judgment before you are aware of it: https://www.unsw.edu.au/newsroom/news/2019/03/our-brains-reveal-our-choices-before-were-even-aware-of-them--st

can tell it it made a mistake and it will blindly change it’s answer whether it made a mistake or not. That also doesn’t feel like it’s able to reason or make judgments.

That’s also how the brain can works, it come up with a plausible explanation after having the result.
See the experience which are spoken about here: https://www.youtube.com/watch?v=wfYbgdo8e-8

I showed the same behavior in humans of some behavior you observed in LLMs, does this means that by your definition, humans doesnt think ?

Claiming LLMs can’t think with the current informations available, and calling that not a belief, is claiming to have a response to this philosophy question.
The only sensible answer is saying you don’t know, or being aware and communicating that your statement is a belief.

No it is here because that’s what they claim.
Nobody yet know how it work, we don’t know how LLMs process information.
Anyone who claim it really think, or it isn’t thinking, is believing, this is not something the current ML field know.

You think computation is thinking ?
I asked for your definition of thinking.
The OP talked about belief, then made a statement using a word that is not precisely defined.
If you think computation is thinking then by your definition the LLM is thinking.
But that’s your definition of thinking.

I asked for your definition, I cannot prove something if we do not agree on a definition first.
You also missread what I said, I did not said AI were thinking.
The burden of proof is on the one who made an affirmation.
I’m not the one who made an affirmation which field experts doesn’t know the answer.
But depending of your definition of thinking, some can be answered.

What called my attention is that assessments of AI are becoming polarized and somewhat a matter of belief.

Proceed to write a belief as a statement in the following paragraph

If you think LLMs doesnt think (I won’t argue that they arent extremely dumb), please define what is thinking, before continuing, and if your definition of thinking doesn’t apply to humans, we won’t be able to agree.

The video very clearly answers this. Like, multiple times.

No, they made affirmation, that’s not a proof.
For the first location, they say the loss of water pression AND the sediments are due to the datacenter.
They are getting their water from a well, if a well runs out, you get more sediments.
Is this your “clear answers” ?

We know that it is a tremendous amount of water because we can estimate and we can see the data of towns literally going into extreme droughts right next to data centers.

If this come from your video again, i again doubt your statements.

Datacenters dont make water magically disapear, it have to go somewhere.
You would see a release pipe, so the water is restituted, or vapor cloud, which should be very visible.
But we dont see any vapor cloud.

AI is not the whole cloud, it’s a fraction of the cloud.
The MIT Press article is from 2022, citing 2019 data. Datacenter tech and heat reuse extremely intensified the last years, so this data is clearly out of date.

Go explain to these people why “bigger DCs are actually better”:

Tell me where there is any proof this is meta fault ? Because they are near the datacenter ? Do you have any idea of the amount of water a datacenter consume ?

Not at all. Bigger datacenter usually use lower carbon energy source, try to lower energy spent for cooling and try to recycle the heat.
New datacenter reuse more and more the heat they generate:
https://blog.equinix.com/blog/2024/06/05/what-is-data-center-heat-export-and-how-does-it-work/

Real water polluter are PFAS producer, not datacenters.

Not every datacenter is like elon datacenters.