61–70 of 279 posts
> If you want to be really good, you have to understand what’s being said. That’s the only way. This is simply not true. Predicting the next letter or word, or id you abstract it away from things that mean something to you, like the next color of a block in a long chain of colored blocks. You would realize that all we are doing is using statistics to predict what the next item might be. There simply is no need or req…
the human brain is just a statistical machine.
Earlier quoted context omitted.
Isn't that exactly how humans learn to respond to stimuli? Don't we just try to predict the best next response to everything? Yes, It's statistics but the fun part is nobody is writing this statistical function by hand.
LLMs don't have a concept of "best". Only most likely in what they've been trained on. I think LLMs ultimately just take imitation to a creative and sophisticated extreme. And imitation simply doesn't comprise the whole of human intelligence at all, no matter how much it is scaled up. The sophistication of the imitation has some people confused and questioning whether everything can be reduced to imitation. It can't.…
There are levels of sophistication in "imitation". It follows a gradient. At the low end of this gradient is a bad imitation.
At the high end of this gradient is a perfect imitation. Completely indistinguishable from what it's imitating.
If an imitation is perfect than is it really an imitation?
If I progressively make my imitation more and more accurate am I progressively building an imitation or am I progressively building the real thing?
See what's going on here? You fell for a play on words. It's a common trope. Sometimes language and vocabulary actually tricks the brain into thinking in a certain direction. This word "imitation" is clouding your thoughts.
Think about it. A half built house can easily be called an imitation of a real house.
This kind of was on my mind recently, I was wondering, when I speak am I just spitting out the next word that makes sense or is there more to it. I think there is for people, I can think before I speak, I can plan out my thought entirely before turning it into words. Each invocation of the chat bot/llm is a new set of probabilities. I can plan what my 2nd token output will be and stick to it. Llm models dont have the…
> when I speak am I just spitting out the next word that makes sense or is there more to it. There is more to it. Specifically you are doing so to advance towards a specific goal. LLMs don't have goals. They just pick from a list of likely tokens - based on their training data - at random to generate the next token.
> If you want to be really good, you have to understand what’s being said. That’s the only way. This is simply not true. Predicting the next letter or word, or id you abstract it away from things that mean something to you, like the next color of a block in a long chain of colored blocks. You would realize that all we are doing is using statistics to predict what the next item might be. There simply is no need or req…
Doing moderately technical translations, state of the art LLMs may get it about 99% right but you find about 1 word per page where they would have needed to “zoom out” to a wider context to choose the word correctly. Often its an important word that is missed.
Earlier quoted context omitted.
No human can solve the Riemann Hypothesis. Why do you expect an AI to do it before you consider it able to understand things?
OTOH, doing something that only humans have done thus far would be a huge step in demonstrating understanding. Does that mean when a computer outputs a new proof it understands?
Train the LLM on a bunch of things but avoid certain things...Things that humans already know about.
The you query the model about that thing. See if the model can come to the same conclusions humans do. You can actually do this right now with chatGPT.
The argument “a sufficiently capable autocomplete must contain a level of general intelligence” is correct but also not very useful. It is a lot like saying “a sufficiently fast horse can fly”. It is technically correct that when you take things to the extreme you can accomplish great things, but we may not reach those levels. We may require completely different technology to reach those levels of autocomplete, and w…
The argument is simpler than that. Prediction requires a model, completely accurate or not. There's a projection of the world in text. A model of the text data we feed it is a model of the world as humans see it. The trend of loss is more and more accurate models of the dataset. So it won't stop at any arbitrary competency level. Indeed, there are already a few abilities GPT possess that are deemed Super Human. It's…
ChatGPT has a model of all the text information on the internet, but it remains to be seen what the hard limits of this model are. Does this model let it do logic or predict the future well, or will no amount of training give it those abilities? Simply being good in one task doesn't imply a general ability to do everything, or even most of everything. LLM's would simply be the last advancement in a field with a lot of similar advancements.
I think this misses the point made by Hinton. The point is that, if a model is to reach a certain level of autocomplete, it needs to understand what it is talking about. Imagine the result of what required deep thought from you is included in the training data, along with billions of other such exchanges. At some point the only way to improve training score is to develop the ability to "understand" things. (That is,…
> it needs to understand what it is talking about "to understand" is a concept invented by the human brain to help it operate. It doesn't actually mean anything concrete in the physical world. See: the Chinese room argument (paradoxically it proves the converse of what Searle was arguing -- that there is no such thing as understanding, because you can emulate a system with identical outputs from the same set of input…
I think this misses the point made by Hinton. The point is that, if a model is to reach a certain level of autocomplete, it needs to understand what it is talking about. Imagine the result of what required deep thought from you is included in the training data, along with billions of other such exchanges. At some point the only way to improve training score is to develop the ability to "understand" things. (That is,…
> it needs to understand what it is talking about "to understand" is a concept invented by the human brain to help it operate. It doesn't actually mean anything concrete in the physical world. See: the Chinese room argument (paradoxically it proves the converse of what Searle was arguing -- that there is no such thing as understanding, because you can emulate a system with identical outputs from the same set of input…
Is not black or white. Probably most of what we do is System 1, most of the time we are mostly meme machines, and so is a good part of the activity on some sectors. But are able to make steps forward, intuition, hard step by step reasoning, finding connections between dots, etc. GPT can do some of that, and in some point of the road someone must decide if we reached somewhere else. Even if making the full road may or…