Well I thought it was gonna be about the hostile deep-fakes but it's about career threat for creatives.
I do wish the robots were doing more mundane or labour intensive tasks and leave the art to humans.
I do wish the robots were doing more mundane or labour intensive tasks and leave the art to humans.
We happen to be in a weird little valley right now where some tasks are still impossible for artificial beings to do better than us. This is a very short term aberration.
If we are still here in a couple generations, when artificial beings have surpassed our output in every way that matters, then we’ll have a choice to make about what we produce for fun. Not because we’re better at it or because it is profitable, just because we want to.
Writing stories, making music, making movies, plumbing, carpentry… the only humans who do these things will be the humans who feel like it.
I'd say all tasks when done for money and under pressure (deadlines, people depending on you, etc) become less enjoyable.
I enjoy coding in my free time. Doing it for work - not so much.
Last time I checked it was an algorithm that needs to be run and fed with input to perform calculations. Not a being?
The only thing that is a "being" about it currently is the nature of the beings that "donated" the training data (without being, hem, asked).
There comes a point at which abstraction becomes too general to be useful, and most attempts to equate LLMs with human beings fall into that category for me. Yes, one could describe the intricacies and complexities of human thought and reasoning as equivalent to prompting a stateless stochastic token matching machine but all that proves is how much of a reach the premise is.
If you ask me, no you cannot do that even if you know you're stretching the limits of that "abstraction". It just doesn't fly even in the wildest theory. Even if an LLM has a world model that could be compared to a human one, the whole underlying nature of experience is completely missing, hence rendering even this narrow comparison extremely wonky.
Otherwise, I like the way you've put it into words though!
Doubt. Explain, for instance, pain. How do you create that algorithmically? Or if you don't like to explain pain, how about love?
You see, that's the problem with being caught up in the "technical person" identity completely. Reducing yourself down to an algorithm is not exactly healthy if you ask me. But it sure explains a lot about where the world is headed and why.
I'm not going to ask you to lift things heavier than you can, or to do math or programming that's beyond your known capabilities.
My goal is going to be to find out what you're best at, then optimize my tasks to be aligned with your training and objectives.
This is pretty much what I do with AI also.
That doesn't explain what pain is. My question was how you would algorithmically create that? If I am as much a being as an algorithm, as you insulted, then sure thing there must be an algorithmic nature underlying my (or anyone's) pain?
I'm not sure what answer would communicate the idea clearest, but:
* If you want the AI version of pain, there are examples of attacks that cause it to bug out and malfunction. It's clearly not operating correctly, and I think someone famously got Bing to say it wants to go on a rampage? Here's blogspam from last year form a quick search: https://nypost.com/2023/02/16/bing-ai-chatbots-destructive-r...
* If you want to know why people or animals experience pain, it's because it's an important sensor to exist in nature. AI has been trained to avoid controversial topics and will push back if you try to get it to talk about something it knows but doesn't want to engage with. That's sort of pain?
* If you want AI with literal neurons that can feel "pain" in a biological sense, there are also organoids, but I don't know much about them.
The world only cares about inputs and outputs.
I meant what I typed. Claiming that a human being "is an algorithm if you abstract enough" is an insult. I hope you can see that?
> If you want the AI version of pain
> If you want to know why people or animals experience pain
> If you want AI with literal neurons that can feel "pain"
Look. I find it fascinating how a simple question can be evaded with so much insistence and gaslighting.
I don't want any of that. I didn't ask for that.
I said an LLM is not a being. It's an algorithm. You claimed that "So are you, if you abstract enough."
If I am to be reduced to an algorithm, then all I want is an explanation of how I am supposed to be that if you can't explain the algorithmic nature of the most fundamental aspects of a (human) being. If you can't, maybe you should consider that your claim is not only wrong, but potentially a sophisticated intellectual form of pseudo-legitimized dehuminization? (And we shouldn't be calling LLMs "artificial beings".)
> The world only cares about inputs and outputs.
This way of thinking is the source of all the issues we are facing right now. It screams "I need more suffering". You have been conditioned into believing that. But it doesn't stand up to the simplest test!
For a system to experience what we describe as pain, it needs to receive sensory input which it predicts signals that something is damaging its internal state, and then model/describe/represent itself as something that has this experience. Such as system would, like you do, believe it is experiencing pain.
Do you have a source to underpin that claim? (Honest question.)
> For a system to experience what we describe as pain, it needs to receive sensory input which it predicts signals that something is damaging its internal state, and then model/describe/represent itself as something that has this experience.
And then the system would feel pain? It doesn't sound convincing. I can model everything I want in my brain, but that certainly doesn't equate experiencing the feeling of pain. What about a headache? Do I describe myself without a headache and the difference to the actual "model" is my pain experience? It sounds a bit too left-brained to me, sorry.
> Such as system would, like you do, believe it is experiencing pain.
How do you know? I think that's a pretty bold claim. Also, pain is not a believe, it's an experience of a feeling?
The system would believe it is experiencing pain. Its model of itself would use the pain concept to describe itself, and it would believe it is experiencing it. And this description would be accurate since it is getting damaged.
I would argue that pain, like any interpretation of sensory input, are believes. "experience of a feeling" is equal to interpreting sensory input in a certain way.
But due to the way we are constructed, we cannot unbelieve certain things. It is like when you see a tree, you cannot decide to not see a tree after you have seen it. Or when you see a 2D projection of a 3D cube. You have to see the cube, even if you know it is not there since it is actually just a bunch of lines on a paper. This makes this type of belief different from other types of beliefs we hold which we easily can change our mind about, like what the weather will be like tomorrow.
It's the same thing with pain. You cannot stop believing in the interpretation of the sensory data that we describe as pain. And this "pain belief" is tied to certain behaviors, like trying to avoid it, since it is built into our reward system that it is negative. The reason we cannot easily "unbelieve" pain is that it would be very dangerous if we could simply ignore it. This is also why we are wired so that love is so hard to unbelieve, reproduction would fail in a species that could easily change their mind on the "love belief" about their offspring or partners.
If you think about phantom limb pain, it is quite clear that it is just a belief/model, and in that case an inaccurate one.
In other words: "Pain isn't real"? Are you serious?
Again, is there any reliable source for that other than a whole book that probably doesn't even talk about what you claim? If it is in the book, can you give a detailed reference where to find it? And if so, does the book provide references to research underpinning those claims?
Or did you just come up with that on your own? It sounds absolutely ridiculous to me. Any attempt to "model" the experience is prone to fail by lack of evidence because pain is, by definition, subjective.
If you just search for the definitions of belief, emotion and pain, you will quickly find that their meanings are distinct. Claiming that pain is a belief is absurd.
Of course it is real. When I feel pain from an injury, the injury is very real, my pain sensor's signals are real and my brain's interpretation of those signals is real, and my behavior of trying to avoid further injury is real.
> Again, is there any reliable source for that other than a whole book that probably doesn't even talk about what you claim? If it is in the book, can you give a detailed reference where to find it? And if so, does the book provide references to research underpinning those claims?
Lisa Barrett is a very famous professor of neuroscience, so sure there is a lot of research in that book :) Here is a short video where she explains it https://www.youtube.com/watch?v=0QfCvIJRtE0 I also recommend her interviews with Lex Fridman.
She talks more about emotions as predictions, but we mean the same thing.
Here is a highly cited paper where she explains the theory: https://pubmed.ncbi.nlm.nih.gov/27798257/
> Or did you just come up with that on your own? It sounds absolutely ridiculous to me. Any attempt to "model" the experience is prone to fail by lack of evidence because pain is, by definition, subjective.
Your pain is very subjective yes, but that just means that it is created by your internal model of the world and yourself, which has a unique subjective perspective of the world.
> If you just search for the definitions of belief, emotion and pain, you will quickly find that their meanings are distinct. Claiming that pain is a belief is absurd.
These words are normally not used in the way I use them, but the point is that emotions are outputs of our model of ourselves. These outputs you can call beliefs, predictions or interpretations if you like.
Unless you really claim that modeling "pain" as some kind of variable in an algorithm can be equivalent to a biological being feeling pain?
I will certainly have a look at the research, but I still believe it's not even talking about that point?
Edit: Where exactly in the cited paper is the claim that emotions are a belief? Can't find it.
Yes, I do. I guess it all boils down to whether or not you think the hard problem of consciousness is actually a problem or not. I doubt it is, but it is a totally respectable position if you do :)
> Where exactly in the cited paper is the claim that emotions are a belief? Can't find it.
As I wrote in my previous response, you can call it predictions too. Predictions are usually beliefs about the future. In the predictive brain literature, it is also used as predictions about the present. I use belief as the output of some inference algorithm that needs to deal with uncertainty.
Here is a quote from the paper: "The brain continually constructs concepts and creates categories to identify what the sensory inputs are, infers a causal explanation for what caused them, and drives action plans for what to do about them. When the internal model creates an emotion concept, the eventual categorization results in an instance of emotion."
But you are right, the paper talks more about how emotional categories are created, and dodges the question of how the "subjective experience of having an emotion" emerges. In my mind, the step is not far though, and boils down to, as said above, how you view the hard problem or consciousness. That we "feel" stuff is a result of an algorithm/model that describes ourselves as having experiences, which is a good model, since how else would we describe ourselves? My belief is that the progress in AI and neuroscience will prove me right :)
Depends on what constitutes a "problem" in this context. But based on my interpretation of what you probably want to express with this sentence, yes, I think consciousness is a problem (for your theory). Otherwise, all of my previous comments wouldn't have made that much sense.
> Predictions are usually beliefs about the future.
I am not sure if I would agree that predictions and beliefs can simply be declared identical so easily. A belief doesn't need to be based on actual data and it doesn't have to be about the future while a prediction, by definition, is. Certain beliefs also seem to be emergent in humans. Otherwise, it would be difficult to explain the independent emergence of religion and belief in god. Beliefs are not necessarily emotions and that's why I have a hard time with the conflation of the two terms. "Predictions about the present" sounds more like a technical term to me. It doesn't make that much sense in the original meaning of the term "prediction". (I understand that it's used that way in many scientific works. Still, in this context, I think it's important to distinguish otherwise we end up redefining the meaning of words.)
> dodges the question of how the "subjective experience of having an emotion" emerges
Yea, it's convenient, right? But I think these are the central points of my argument that you brush away so swiftly. You won't get around subjective experience and consciousness. And how would you go about proving that an algorithm can have subjective experience and "feel" pain or emotions? And if you really believe that this is possible, how do you make sure that in your research nothing extremely unethical happens? By your definition, the Sims may evolve to something that could be protected by certain rights soon? ;)
I could imagine that in reality all of that isn't needed. Whether the algorithm really feels something or not - it doesn't matter as long as the expression is realistic enough for humans to believe that it feels something. Then you get the consciousness "injected" into your algorithm from the outside. Can you convince someone with a realistic sophisticated simulation that something is conscious even though it isn't? Probably. Still, it won't get you anywhere. Nevertheless, I find that to be a much more likely avenue than that we will ever be able to prove that an algorithm can experience feelings in a way that is equivalent to biological beings. I also don't think this kind of research is all that beneficial to us as humans, especially when mixed with advances in AI. But good luck with all of that and thanks for the references!
The thing is that I can't prove that you or anyone else is conscious either, including myself. With conscious acting AIs it will be the same, and these AIs will believe they are conscious in the same way as we do. So yes, we will have to treat them as if they are conscious.
See, that's the thing. It sounds extremely entitled to say that "you believe you are conscious" (but in reality you aren't). You don't know that. You have no evidence for it. You believe that we believe we are conscious. Your opinion is acceptable to me as long as you keep it limited to your own being and discuss it as an explicit hypothesis. I find it overreaching to put this as a general claim and make it sound as if it was obvious, self evident, and applicable to anyone else. It's not.
There’s nothing magic about it, other than that we don’t currently understand much about how it works or how to modulate it.
It’s like saying, “if we’re reducible to machines then explain itchiness”.
I think that's a very reductionist view point that differs from the official definition of pain?
> There’s nothing magic about it, other than that we don’t currently understand much about how it works or how to modulate it.
I didn't claim that it's magical. Just that you wouldn't be able to produce the experience of pain algorithmically?
Still waiting for someone to explain how it could be produced algorithmically. I asked GPT. It said it would be immoral to even think about doing so if we could.
> It’s like saying, “if we’re reducible to machines then explain itchiness”.
If you can explain how itchiness is created algorithmically, I would also take that just out of curiosity! Though not the same as pain, it would be a good starting point maybe.
We have all benefited from mass replication of mediums that before required someone to work for each performance.
I don't see why generation of content should be treated any different. And humans might always keep an edge on this. Well crafted coherent product with decent direction often feels better than a bland one. I really wonder how soon AI can replicate the first one.