6 ms·

Then again, I think the opposite is (at this moment) equally unfalsifiable. Humans have on average 86 billion neurons, sperm whales have bigger brains and their

by lewhoo 3y ago

Then again, I think the opposite is (at this moment) equally unfalsifiable. Humans have on average 86 billion neurons, sperm whales have bigger brains and their average neuron count is 200 billion. African elephants have 257 billion neurons. There are some reasons to believe we are kind of special.

gumballindie 3y ago
To me it sounds as if ai workers are trying to oversimply how humans and brains generally speaking work to fit their narrative. A bit among the lines of "if my grandmother had wheels, she would be a bicycle". Emulating neurons doesn't make intelligence.
- logicprog 3y ago
  The "LLMs will attain intelligence/consciousness if we just throw enough neurons at the problem" crowd really seems to be engaging in cargo cult thinking to me. Not that I think it would be impossible to make a conscious or intelligent mind using simulated neural networks — I don't think that there's some kind of "special mystical property" about humans like a soul or something — it's just that the structure and training of LLMs is fundamentally wrong for producing intelligence, that's all. It's an easy, simple route that gives you startling results for the first 30% but it's a dead end in the long run for achieving what they want, yet they're insisting that if they just keep pantomiming hard enough real intelligence will come.
- danielbln 3y ago
  While I also don't think autoregressive LLMs will ever gain sentience, I also don't subscribe to the other side of the argument that it;s all a dead end and true(tm) intelligence can never be achieved. Fact of the matter is, we don't know. We don't know why we are sentient, so how can we argue what may or may not lead to sentience when we have a lacking understanding of the inner working of current gen LLMs? A little more humility on what we don't know is in order, from all sides.
  - cameronh90 3y ago
    My guess is we won't know when we've developed a sentient AI until we're well past the point that it's undeniable, and then in hindsight we might be able to identify where it began to transition from non-sentient to sentient. Sentience isn't even well-defined and I'm not sure we can even point to any unique quality of any individual human as indicative of sentience. We certainly can't agree on what point an individual human becomes sentient, as evidenced by the abortion debate. At best we seem to have some statistical evidence based on what we've achieved as a species being superior to what oak trees have achieved, but at an individual level, I don't know how to prove to anyone that I'm not just a very advanced LLM. There is another sense where sentience is just being used to mean "that which is human", and by definition, nothing aside from humans will ever qualify as that.
    - gumballindie 3y ago
      Sentience is a property of animal life. Electrons, software, rocks, etc don't posses sentience, so the notion that ai which is software may be sentient is nonsense.
      - thek3nger 3y ago
        Animal life is just a different permutation of electrons, carbon and “software”.
      - cameronh90 3y ago
        This is just extending the "sentience is that which is human" definition to animals. What about animals makes them sentient? Until we can answer that, people are just going to be talking past each other. Even if you forget about AI, whether animals are sentient, and, if so, which animals are sentient, is a big argument in biology/ethics/law that's been going on for at least half a century. The UK recently passed a law that declares animals to be sentient, but not all invertebrates are considered sentient in that law. Their justification is that they don't contain a central nervous system, but what special property about a CNS confers sentience?
- grantcas 3y ago
  [dead]

by lewhoo 3y ago

gumballindie 3y ago

To me it sounds as if ai workers are trying to oversimply how humans and brains generally speaking work to fit their narrative. A bit among the lines of "if my grandmother had wheels, she would be a bicycle". Emulating neurons doesn't make intelligence.

logicprog 3y ago
The "LLMs will attain intelligence/consciousness if we just throw enough neurons at the problem" crowd really seems to be engaging in cargo cult thinking to me. Not that I think it would be impossible to make a conscious or intelligent mind using simulated neural networks — I don't think that there's some kind of "special mystical property" about humans like a soul or something — it's just that the structure and training of LLMs is fundamentally wrong for producing intelligence, that's all. It's an easy, simple route that gives you startling results for the first 30% but it's a dead end in the long run for achieving what they want, yet they're insisting that if they just keep pantomiming hard enough real intelligence will come.
danielbln 3y ago
While I also don't think autoregressive LLMs will ever gain sentience, I also don't subscribe to the other side of the argument that it;s all a dead end and true(tm) intelligence can never be achieved. Fact of the matter is, we don't know. We don't know why we are sentient, so how can we argue what may or may not lead to sentience when we have a lacking understanding of the inner working of current gen LLMs? A little more humility on what we don't know is in order, from all sides.
- cameronh90 3y ago
  My guess is we won't know when we've developed a sentient AI until we're well past the point that it's undeniable, and then in hindsight we might be able to identify where it began to transition from non-sentient to sentient. Sentience isn't even well-defined and I'm not sure we can even point to any unique quality of any individual human as indicative of sentience. We certainly can't agree on what point an individual human becomes sentient, as evidenced by the abortion debate. At best we seem to have some statistical evidence based on what we've achieved as a species being superior to what oak trees have achieved, but at an individual level, I don't know how to prove to anyone that I'm not just a very advanced LLM. There is another sense where sentience is just being used to mean "that which is human", and by definition, nothing aside from humans will ever qualify as that.
  - gumballindie 3y ago
    Sentience is a property of animal life. Electrons, software, rocks, etc don't posses sentience, so the notion that ai which is software may be sentient is nonsense.
    - thek3nger 3y ago
      Animal life is just a different permutation of electrons, carbon and “software”.
    - cameronh90 3y ago
      This is just extending the "sentience is that which is human" definition to animals. What about animals makes them sentient? Until we can answer that, people are just going to be talking past each other. Even if you forget about AI, whether animals are sentient, and, if so, which animals are sentient, is a big argument in biology/ethics/law that's been going on for at least half a century. The UK recently passed a law that declares animals to be sentient, but not all invertebrates are considered sentient in that law. Their justification is that they don't contain a central nervous system, but what special property about a CNS confers sentience?
grantcas 3y ago
[dead]