4 ms·
I think an AI can learn everything it needs from books and other text. But others claim that it can never understand qualia like the feel of grass between your
by tlb 4y ago
I think an AI can learn everything it needs from books and other text. But others claim that it can never understand qualia like the feel of grass between your toes or the sensation of the color red without some direct experience.
Does an AI really need direct experience of such things to be useful in the world? Or can it just take our (written) word that grass feels good between your toes and make any grass-toe related decisions on that evidence? Assuming we're just interested in having AI solve useful problems for us like curing cancer, I don't see the need for it.
It could still be that some important facts about people are missing from the literature, because we consider them too obvious to write down. That seems unlikely to me, given the amount of psychological research on trivial factors. But if we find a system making mistakes, we can probably add some written explanations to the training set it'll get fixed.
- akasakahakada 4y agoAs far as I concern, "feelings" or "sensations" is just a vector in finite dimensional space. No magic behind. Taste can already be replicated. https://www.meiji.ac.jp/cip/english/news/2021/enjsp3000000jrgy.html https://www.meiji.ac.jp/cip/english/news/2021/enjsp3000000jr...
- Jensson 4y agoYou are wrong, you need the entire computational network that processes that data and creates signals for the brain depending on various circumstances such as nutrient levels and so on. Low on protein? You will find meat tastier etc, that is a part of the feelings and taste, so it can't be replicated with a simple vector. I don't really see how we could copy all of that logic to a computer, we would need some super brain scan or something, and without it the computer wont be able to understand the feeling. ChatGPT works on text since we produced so much text that for it to process and associate. We can't do the same thing for taste or feelings. Only viable alternative I can see is that we try to reverse engineer it, or we put permanent brain scans on people and see what they taste at what circumstances and how it affects their feelings over time, and then train a model on that data, that could make it possible.
- akasakahakada 4y ago> You will find meat tastier etc, that is a part of the feelings and taste, so it can't be replicated with a simple vector. What is the logic? Didn't you just literally if seeing(meat): bias += 1 ???
- Jensson 4y agoYeah, now write down all such biases for taste and feelings. You will find it quite hard to get something that replicates the taste and feelings of a human, that is why we create models to generate that code based on tons of data. The problem is that we don't have tons of data for these things so we wont be able to create such a model until we do.
- akasakahakada 4y agoThis is just based on unknown factors, and the difficulty of getting the dataset. But this do not stop the thing that taste of feeling can be expressed as a vector and that is its representation inside the neural network. But then I think I get what you mean. Yeah the thing behind taste is a matrix, you multiply the current state of the body+food and it returns what you feel by your tongue. Since matrix modelize all kinds of interaction between all other factors. This should clears the problem. But still the output is a vector. Taste is a vector. But different people can have different matrix to be multiplied with. In physics they call this kind of matrix "observable". If you know physics you get the idea.
- Jensson 4y agoYeah, that makes more sense. I just wanted to point out that taste isn't just the input but also the experience of that input. Then you could also argue that replicating human feelings or taste isn't very important, and it isn't really except if you want it to understand humans. I think it will be hard to understand humans well without also being able to predict their feelings or taste, we humans have that built in but a computer would need data to predict it.
- jacknews 4y agoI think a real sense of 3d space and time would be very beneficial, but you are quite right, there's no need for AIs to actually experience lower level sensations in order to still be useful, and able to discuss them, etc. In the same way as with emotions. Some humans do not feel emotions, but can still 'model' them, and act to accommodate other people's emotions. In fact I'd argue that we want to make our AIs psychopathic, in a good way. We don't want sad, or indignant AI, or especially, angry AI.