4 ms·

> And your brain does it with chemistry. Alledgedly. > Define your test. A pass is a model with no human training data input that is able to reason on human

> And your brain does it with chemistry. Alledgedly. > Define your test. A pass is a model with no human training data input that is able to reason on human content. > Every single test that existed when I was a kid has been bested. The mechanical turk fooled humans for 80 years and passed all the tests at its time. As it turns out, the tests were not good enough. > Coming up with a test is itself a challenge, these days. My test is simple to understand and conceive. We just don't have any valid candidate to put through it. My requirements disqualify LLMs.

ben_w 2y ago
> Alledgedly If you can't agree on our brains being chemical in nature, you're engaged in magical thinking and we have no more to discuss. > A pass is a model with no human training data input that is able to reason on human content. What does it mean to "reason on human content"? Both "reason" and separately "human content"? The former because some argue that "reason" is the mechanical application of logic, where a single transistor alone would be sufficient to pass; others require it to be a conscious application, at which point you've got to deal with there being about 40 different definitions of "consciousness". The latter, because a model with no human training data will not be able to read any human language for the same reason that no human alive can read Linear A… but also, our content includes the games we play, and AI trained on the rules of Go with zero examples of our games beat human Go players at the game of Go, which definitely seems to me to be doing things on "human content". > My test is simple to understand and conceive. We just don't have any valid candidate to put through it. My requirements disqualify LLMs. Where did you write this "simple to understand and conceive" test? Because what you've given here is not even a test. If you can't come up with a test, there's nothing to distinguish you personally from the AI that you're dismissing.
- alganet 2y ago
  > If you can't agree on our brains being chemical in nature, you're engaged in magical thinking and we have no more to discuss. I agree that our brains have a lot of chemistry. I can't deny that. Science shows we can mess up brains by messing up their chemistry. So far, though, we still don't understand how it works. We can't use chemistry to improve it, just to break it or fix it. That alone should be a huge red flag for the reductionist point of view. It's our best hypothesis, but is not good enough to travel disciplines to IT. We don't know enough to compare the two, hence the skepticism. > What does it mean to "reason on human content"? Pass the turing test should be enough. I'm just changing the requirements (no tiny humans inside, complete or otherwise). > no human alive can read Linear A If we had a living human fluent in Linear A, I believe we could decipher it. Machines have us, living humans speaking a living language, to learn. > If you can't come up with a test, there's nothing to distinguish you personally from the AI that you're dismissing. I can. Open the machine and prove me that there are no tiny humans or human pieces inside.
  - ben_w 2y ago
    > We can't use chemistry to improve it, just to break it or fix it. That alone should be a huge red flag for the reductionist point of view. Caffeine and theobromine are both commonplace cognitive enhancers, AKA nootropics. There's plenty that aren't available over the counter, too, like amphetamines. > Pass the turing test should be enough. LLMs already do that, and it's a problem because it's a fraud opportunity. > If we had a living human fluent in Linear A, I believe we could decipher it. Machines have us, living humans speaking a living language, to learn. How is this supposed to be different from what LLMs already do? > I can. Open the machine and prove me that there are no tiny humans or human pieces inside. That reads like an LLM — and not even a good one, I mean like the GPT-2 based AIDungeon — wrote it, given there's obviously no tiny human inside a computer. What did you actually mean? Be specific, be testable.
    - alganet 2y ago
      > Caffeine and theobromine We still don't fully understand how they work. Of course, we didn't make them, nature did. > amphetamines Those break more than enhance. > LLMs already do that As I said, disqualified for being a cauldron of souls. > How is this supposed to be different from what LLMs already do? My requirements enforce a no human content during training. It's exactly that requirement that makes the test easy. If you allow human content to be part of it, then tests become ellusive. > What did you actually mean? Be specific, be testable. I'm saying that LLMs are the spiritual successor to the mechanical turk. The entire concept falls apart if you take the human part of it out. There are living humans that had their works put inside and are publicly saying "I don't want to be inside this machine". What more do you want? Yes, they are impressive. I already said that. It's just not intelligence, it's an almost mythological trick we should have been more guarded against.
      - ben_w 2y ago
        > As I said, disqualified for being a cauldron of souls. Would you like help moving those goal-posts for your supposedly easy to understand test? This claim: > My requirements enforce a no human content during training. It's exactly that requirement that makes the test easy. If you allow human content to be part of it, then tests become ellusive. Is directly incompatible with this one: > If we had a living human fluent in Linear A, I believe we could decipher it. Machines have us, living humans speaking a living language, to learn If there's no human content allowed, then we here today can't use a living human fluent in Linear A to decipher it. Nor, for that matter, even so much as the mere historical record of Linear A without any human. You are demanding AI pass a standard that literally no human would ever be capable of, because if I simply switch the nouns you get: "A pass is a human with no Linear A training data input that is able to reason on Linear A content". > There are living humans that had their works put inside and are publicly saying "I don't want to be inside this machine". What more do you want? I want you to be specific and to be testable. You still have a test which is… have you ever had a client who wanted you to build "Uber for aircraft" or something else equally vague? They didn't even realise they didn't have a plan. That's what I see from you. Your position is pre-paradigmatic. Consider for example: a human who grows up with no human works put inside their head is what is called a "feral child" — they do not magically understand anything of the rest of civilisation, this is because that understanding has to be taught by… observing humans as they talk to each other, mimicking them, trying to talk with them to get feedback, etc., which is what the LLMs are all doing. It looks like a double standard to me, that you demanding that an AI, who is raised by actual wolves does better than a human who is raised by actual wolves. What you have actually said so far, would refuse to recognise as intelligent an AI that is given control of an android body and had to read books by physically holding them and turning the pages with robot hands, which it operates in human timescales, and which gets there in the same timeframe as an actual human. It would also rule out an actual human, because we learn from the works of other humans.
Hizonner 2y ago
> A pass is a model with no human training data input that is able to reason on human content. OK, then humans fail hard. A baby raised with "no human training data" (no conversation, no active teaching, no social interaction, obviously no books or videos) would be a total imbecile... or dead.
- alganet 2y ago
  You're using "training" in a looser way than me. LLMs have training data and context. I'm talking about training data. Conversation, interaction... those happen within an LLM context window. LLMs learn nothing from those experiences. What they can learn within a context window is super limited and disposable. You can't really compare the two. Humans are much, much more.
  - ben_w 2y ago
    > Conversation, interaction... those happen within an LLM context window. LLMs learn nothing from those experiences. False. Why do you think e.g. ChatGPT has buttons for indicating if an answer is good or bad? Even when a specific version is frozen, this is a choice rather than a fundamental requirement, and the data from human feedback is used to continue training the system as a whole.
    - alganet 2y ago
      Ah, yes, the typical learning experience. You are born and fed a firehose of all the content available in the world. Then, when you need to learn something, you just firehose it all again from scratch with some added notes you made on a napkin. Can we at least agree that this cannot be compared with a human baby?
      - ben_w 2y ago
        > Ah, yes, the typical learning experience. You are born and fed a firehose of all the content available in the world. You are indeed born being fed a firehose of experience. Is it all the content available in the world? No, it's much less ordered than that. Also much less content. If your point were to have been "current AI are thick because they need a lot of examples to learn anything", you'd have a much easier time with your claims. (Some people still argue that point because vision is high bandwidth, but nevertheless it would be easier to argue than setting the bar so high that no human could possibly get there). > Then, when you need to learn something, you just firehose it all again from scratch with some added notes you made on a napkin. I have no idea what this is supposed to mean in this analogy. > Can we at least agree that this cannot be compared with a human baby? https://pmc.ncbi.nlm.nih.gov/articles/PMC27565/ https://pmc.ncbi.nlm.nih.gov/articles/PMC27565/ You can compare almost anything, if you want to. But why would it matter, that it's not like a baby? A baby treated the way you propose treating an AI, would still be somewhere between "feral" and "dead". That, even in isolation, is enough to show that your test is not well-conceived.

ben_w 2y ago

> Alledgedly If you can't agree on our brains being chemical in nature, you're engaged in magical thinking and we have no more to discuss. > A pass is a model with no human training data input that is able to reason on human content. What does it mean to "reason on human content"? Both "reason" and separately "human content"? The former because some argue that "reason" is the mechanical application of logic, where a single transistor alone would be sufficient to pass; others require it to be a conscious application, at which point you've got to deal with there being about 40 different definitions of "consciousness". The latter, because a model with no human training data will not be able to read any human language for the same reason that no human alive can read Linear A… but also, our content includes the games we play, and AI trained on the rules of Go with zero examples of our games beat human Go players at the game of Go, which definitely seems to me to be doing things on "human content". > My test is simple to understand and conceive. We just don't have any valid candidate to put through it. My requirements disqualify LLMs. Where did you write this "simple to understand and conceive" test? Because what you've given here is not even a test. If you can't come up with a test, there's nothing to distinguish you personally from the AI that you're dismissing.

alganet 2y ago
> If you can't agree on our brains being chemical in nature, you're engaged in magical thinking and we have no more to discuss. I agree that our brains have a lot of chemistry. I can't deny that. Science shows we can mess up brains by messing up their chemistry. So far, though, we still don't understand how it works. We can't use chemistry to improve it, just to break it or fix it. That alone should be a huge red flag for the reductionist point of view. It's our best hypothesis, but is not good enough to travel disciplines to IT. We don't know enough to compare the two, hence the skepticism. > What does it mean to "reason on human content"? Pass the turing test should be enough. I'm just changing the requirements (no tiny humans inside, complete or otherwise). > no human alive can read Linear A If we had a living human fluent in Linear A, I believe we could decipher it. Machines have us, living humans speaking a living language, to learn. > If you can't come up with a test, there's nothing to distinguish you personally from the AI that you're dismissing. I can. Open the machine and prove me that there are no tiny humans or human pieces inside.
- ben_w 2y ago
  > We can't use chemistry to improve it, just to break it or fix it. That alone should be a huge red flag for the reductionist point of view. Caffeine and theobromine are both commonplace cognitive enhancers, AKA nootropics. There's plenty that aren't available over the counter, too, like amphetamines. > Pass the turing test should be enough. LLMs already do that, and it's a problem because it's a fraud opportunity. > If we had a living human fluent in Linear A, I believe we could decipher it. Machines have us, living humans speaking a living language, to learn. How is this supposed to be different from what LLMs already do? > I can. Open the machine and prove me that there are no tiny humans or human pieces inside. That reads like an LLM — and not even a good one, I mean like the GPT-2 based AIDungeon — wrote it, given there's obviously no tiny human inside a computer. What did you actually mean? Be specific, be testable.
  - alganet 2y ago
    > Caffeine and theobromine We still don't fully understand how they work. Of course, we didn't make them, nature did. > amphetamines Those break more than enhance. > LLMs already do that As I said, disqualified for being a cauldron of souls. > How is this supposed to be different from what LLMs already do? My requirements enforce a no human content during training. It's exactly that requirement that makes the test easy. If you allow human content to be part of it, then tests become ellusive. > What did you actually mean? Be specific, be testable. I'm saying that LLMs are the spiritual successor to the mechanical turk. The entire concept falls apart if you take the human part of it out. There are living humans that had their works put inside and are publicly saying "I don't want to be inside this machine". What more do you want? Yes, they are impressive. I already said that. It's just not intelligence, it's an almost mythological trick we should have been more guarded against.
    - ben_w 2y ago
      > As I said, disqualified for being a cauldron of souls. Would you like help moving those goal-posts for your supposedly easy to understand test? This claim: > My requirements enforce a no human content during training. It's exactly that requirement that makes the test easy. If you allow human content to be part of it, then tests become ellusive. Is directly incompatible with this one: > If we had a living human fluent in Linear A, I believe we could decipher it. Machines have us, living humans speaking a living language, to learn If there's no human content allowed, then we here today can't use a living human fluent in Linear A to decipher it. Nor, for that matter, even so much as the mere historical record of Linear A without any human. You are demanding AI pass a standard that literally no human would ever be capable of, because if I simply switch the nouns you get: "A pass is a human with no Linear A training data input that is able to reason on Linear A content". > There are living humans that had their works put inside and are publicly saying "I don't want to be inside this machine". What more do you want? I want you to be specific and to be testable. You still have a test which is… have you ever had a client who wanted you to build "Uber for aircraft" or something else equally vague? They didn't even realise they didn't have a plan. That's what I see from you. Your position is pre-paradigmatic. Consider for example: a human who grows up with no human works put inside their head is what is called a "feral child" — they do not magically understand anything of the rest of civilisation, this is because that understanding has to be taught by… observing humans as they talk to each other, mimicking them, trying to talk with them to get feedback, etc., which is what the LLMs are all doing. It looks like a double standard to me, that you demanding that an AI, who is raised by actual wolves does better than a human who is raised by actual wolves. What you have actually said so far, would refuse to recognise as intelligent an AI that is given control of an android body and had to read books by physically holding them and turning the pages with robot hands, which it operates in human timescales, and which gets there in the same timeframe as an actual human. It would also rule out an actual human, because we learn from the works of other humans.

Hizonner 2y ago

> A pass is a model with no human training data input that is able to reason on human content. OK, then humans fail hard. A baby raised with "no human training data" (no conversation, no active teaching, no social interaction, obviously no books or videos) would be a total imbecile... or dead.

alganet 2y ago
You're using "training" in a looser way than me. LLMs have training data and context. I'm talking about training data. Conversation, interaction... those happen within an LLM context window. LLMs learn nothing from those experiences. What they can learn within a context window is super limited and disposable. You can't really compare the two. Humans are much, much more.
- ben_w 2y ago
  > Conversation, interaction... those happen within an LLM context window. LLMs learn nothing from those experiences. False. Why do you think e.g. ChatGPT has buttons for indicating if an answer is good or bad? Even when a specific version is frozen, this is a choice rather than a fundamental requirement, and the data from human feedback is used to continue training the system as a whole.
  - alganet 2y ago
    Ah, yes, the typical learning experience. You are born and fed a firehose of all the content available in the world. Then, when you need to learn something, you just firehose it all again from scratch with some added notes you made on a napkin. Can we at least agree that this cannot be compared with a human baby?
    - ben_w 2y ago
      > Ah, yes, the typical learning experience. You are born and fed a firehose of all the content available in the world. You are indeed born being fed a firehose of experience. Is it all the content available in the world? No, it's much less ordered than that. Also much less content. If your point were to have been "current AI are thick because they need a lot of examples to learn anything", you'd have a much easier time with your claims. (Some people still argue that point because vision is high bandwidth, but nevertheless it would be easier to argue than setting the bar so high that no human could possibly get there). > Then, when you need to learn something, you just firehose it all again from scratch with some added notes you made on a napkin. I have no idea what this is supposed to mean in this analogy. > Can we at least agree that this cannot be compared with a human baby? https://pmc.ncbi.nlm.nih.gov/articles/PMC27565/ https://pmc.ncbi.nlm.nih.gov/articles/PMC27565/ You can compare almost anything, if you want to. But why would it matter, that it's not like a baby? A baby treated the way you propose treating an AI, would still be somewhere between "feral" and "dead". That, even in isolation, is enough to show that your test is not well-conceived.