John R Fraser

Missionary, teacher, theologian. Author of The Growing Disciple: Seven Attributes of a Transformed Life.

[John R Fraser @John_R_Fraser]

@otoge_3107 @ripgarythegoat @gmkurtzer I see you've had more experience with him than I have. First time for me, haha.

[John R Fraser @John_R_Fraser]

You put some references but not one quote from Scripture where it says we are not under the law except for the Ten Commandments. You claim you posted it, but that's not true. So I want the reference AND THE WORDS where it says that we are not under the Law except for the decalogue.

[John R Fraser @John_R_Fraser]

Try reading back where I explained how consciousness can't be proven or disproven, but is only inferred. Your whole problem is thinking that consciousness is the kind of thing that can be proven, like a physical property or something. That's just not true. Consciousness is not an object that can be studied in a lab using experiments.

But hey, while we're at it, I would like you to disprove the proposition that I am the only conscious being in the universe. If you can't disprove that, then I win. Deal?

[John R Fraser @John_R_Fraser]

I'm not saying we have WEAKER evidence for LLMs. I'm saying that we have no evidence whatsoever. They are not the same kind of thing as we are so the grounds for the reasonable inference that we make for other people is entirely absent. It's not just a little bit weaker, it's just non-existent. Period. So your analogy fails on the launch pad.

[John R Fraser @John_R_Fraser]

Me: It's illegitimate to demand proof that LLMs are not or can't be conscious because we can't prove consciousness in other beings one way or the other, we just infer it because they're the same kind of thing that we are. LLMs are not the same kind of thing that we are, so there's no reason to infer it to begin with.

You: Yeah, but you can't prove that LLMs can't be conscious, so I win.

[John R Fraser @John_R_Fraser]

Me: It's illegitimate to demand proof that LLMs are not or can't be conscious because we can't prove consciousness in other beings one way or the other, we just infer it because they're the same kind of thing that we are. LLMs are not the same kind of thing that we are, so there's no reason to infer it to begin with.

You: Yeah, but you can't prove that LLMs can't be conscious, so I win.

[John R Fraser @John_R_Fraser]

@ripgarythegoat @gmkurtzer Made a couple of typos in there, getting ready to go pick up supper for me and the wife 🙂.

[John R Fraser @John_R_Fraser]

You just keep saying the same thing over and over in different words, basically ignoring my arguments. I'll leave this as my last summary because I don't think you're going to suddenly change your tact.

We infer consciousness in other people by analogy, not by "proof" that other people are conscious. By all appearances they are the same kind of thing that we are, so it's reasonable to think they experience things the way we do.

LLMs are NOT the same kind of thing that we are, but a lot of people don't really know how they work and don't know that and may make an inference to consciousness ("understanding") based on outputs alone, but that inference is not warranted the way it is for other people.

Then along comes someone like you, and you say, "well, you can't prove they aren't conscious, can you?" No, but that's because consciousness can't be proven either way even in other people. So it's not actually for anything unless you want to argue that we should take LLMs as conscious unless proven otherwise. But the relevant inference is not available for LLMs because they aren't the same kind of thing as people and don't produce outputs the same way. So there is actually no reason to infer consciousness to begin with.

And yes, you are saying you win if proven wrong because you insist that someone prove that consciousness in LLMs is impossible. But we have no more reason to infer consciousness in an LLM than we do in a calculator. The output is much more complex and better at mimicking human outputs, but the relevant reasons for inferring consciousness to being with are absent.

So the analogy fails. And it will keep failing even after this guy comes back on and says the same stuff he's been saying over and over and over again.

[John R Fraser @John_R_Fraser]

My definition of question begging was not "any argumentative move my opponent won't accept," but has to do with an argument among philosophers over when begging the question occurs (philosophers do not make the blunder of directly assuming the conclusion in a premise but do frequently accuse each other of begging the question in other ways). Anyways, I thought maybe you were philosophically savvy enough to be aware of that. I guess not.

Going back to the o.p., I think it is aimed at people who simply don't know how AI works and believe that machines do, in fact, have understanding because they think they work the same way human brains do. That's how we infer consciousness in other people, so if AI worked the same way as human brains then that would be a rational inference. His point is that producing outputs in the way that AI does is NOT analogous to human thought. And he's right.

You are simply claiming that the AI method might still produce conscious thought ("understanding") even though it operates completely differently. And you think everybody else has the burden of proof to prove you wrong, and has to prove what it is in human beings that produces consciousness, and if they don't then you win. So yeah, that's your problem. Your analogy still fails.

[John R Fraser @John_R_Fraser]

First you say the reductio does not depend on "LLMs might be conscious" being true, then you say that I need to prove that consciousness is absent in LLMs! In other words, you just admitted that in fact the reductio DOES depend on LLMs possibly being conscious, because you insist that I prove they aren't!

I don't know if you realize this or have ever thought about it, but it's impossible to "prove" consciousness in another entity one way or the other. You can't prove other people are conscious, you can only know your own consciousness. We infer consciousness in other people by analogy. But I don't experience your consciousness or anyone else's, just my. So your demand that I "prove" that LLMs aren't or can't be conscious or you win is an utterly ridiculous requirement. Fun debating tactic, maybe, but not a legitimate one.

The thing is, your supposed great gotcha analogy is literally question-begging. It is only going to be accepted by people who accept your underlying premise that LLMs are or might be conscious. I'm using the accepted definition of question-begging as making any move in an argument which will not be accepted by someone who does not already agree with your conclusion. And I do not.

You can say that the o.p. is only going to be accepted by people who already accept his conclusion, and that would be a fair point. But that also means that if the o.p. fails then so does your rebuttal.

[John R Fraser @John_R_Fraser]

@ripgarythegoat @gmkurtzer If you think "my reductio might be successful because LLMs might be conscious," is a real gotcha, I just don't think it is. And yes, I think that's your issue. That's not a reduction of my position, that's always been my position with respect to people like you.

[John R Fraser @John_R_Fraser]

Well, if you think the description of LLMs leaves open the possibility that they are conscious, that's your issue. You reductio is a nothingburger but I think you've maybe realized that by now. You thought you had a gotcha and you didn't. But congrats on the apparently successful engagement farming!

[John R Fraser @John_R_Fraser]

You really don't think his point had anything to do with the actual details of the mechanism in question? Or that the mechanism behind LLMs if fully articulable and UNDERSTOOD (I use that word deliberately) by people? Or that it excludes understanding on the part of the LLM? You really think you have given a full and complete description of human thought? Don't think so.

I'll just say for my part that understanding requires consciousness. And if you think you're explained consciousness in your reductio somewhere, I don't even know what to tell you except maybe take some classes in philosophy, preferably philosophy of mind.

[John R Fraser @John_R_Fraser]

Yes, great verse. You are teaching the man-made doctrine that the Torah is divided into different parts (which I have already pointed out several times). That is not taught in Scripture, anywhere. Paul does not say we are not under the Torah except for the 10 commandments. That was taught by John Calvin and others, it is not in the Bible. So you should really take Mat. 15:9 to heart, because you are breaking it with what you are saying about the Torah.

[John R Fraser @John_R_Fraser]

@Gaba70707 @tereoentole @royallebach I'll take Scripture. What I won't take is false, man-made doctrines being read into Scripture. That's called eisegesis, and that's what you're doing.

[John R Fraser @John_R_Fraser]

The inference you claimed to be testing was, “Here is the mechanism, therefore understanding is absent.” Those are YOUR words, not mine.

So yes, your reductio does, in fact, require that both use the same mechanism. If the mechanism is different and now you are claiming that of course they are, then your reductio fails.

[John R Fraser @John_R_Fraser]

No, the analogous move against humans would not work unless you can demonstrate that in fact the mechanism is the same for both. Which is why I said that you assume that humans and LLMs are the same kind of thing (ie. that they work the same way). Thanks for proving me right, btw.

[John R Fraser @John_R_Fraser]

@ripgarythegoat @gmkurtzer The problem with your attempted reductio is that it assumes at the outset that humans and LLMs are the same kind of thing, and that if one has understanding then so does the other. So I don't know how you think you've proven anything.

[John R Fraser @John_R_Fraser]

@Gaba70707 @tereoentole @royallebach None of the verses that you gave say that, friend. Not a single one. You are reading your man-made false doctrine into them.

[John R Fraser @John_R_Fraser]

@ripgarythegoat @gmkurtzer Funny that you say a human doesn't understand anything and then proceed to explain a bunch of stuff as if you did understand it - and as if your readers should be able to understand it. Weird.