4 ms·

I kept assembling points I wanted to make in the comments, and then the article would cover them itself, so I'll just quote it: > He imagined that this machine

I kept assembling points I wanted to make in the comments, and then the article would cover them itself, so I'll just quote it: > He imagined that this machine, which he called “the great instrument of reason,” would be able to answer all questions and resolve all intellectual debate. “When there are disputes among persons,” he wrote, “we can simply say, ‘Let us calculate,’ and without further ado, see who is right.” > But as today’s data scientists devise ever-better algorithms for natural language processing, they’re having debates that echo the ideas of Leibniz and Swift: Even if you can create a formal system to generate human-seeming language, can you give it the ability to understand what it’s saying? This is the important part. Having the naivete to assume that a flawed mechanical system represents the whole of Truth, is worse than having no system at all.

Banana699 5y ago
> flawed mechanical system I don't like what "mechanical" is implying here, it gives the impression that humans have weird magic in their head that makes them understand language or search for truth (whatever approximate version they like anyway). It's just a flawed system, period. Your brain circuits are a flawed system, and the artifacts produced by the data science craze are flawed. They are not flawed equally, but they are all physical infromation processing systems trying to make sense of the weird jumble of signals coming to it from Outside.
- yawboakye 5y ago
  > Your brain circuits are a flawed system What a roundabout way to admit limitation both in knowledge and tools to investigate them. Except this one is more sinister because it almost closes the curtain on approaching this natural order (of brain circuitry) with both humility and curiosity.
  - Banana699 5y ago
    >What a roundabout way to admit limitation both in knowledge and tools to investigate them. Why? I view humans (and all biological beings really) as machines, fantastic contraptions of unimaginable complexity that push the materials they are made of to the absoulute limits. Their brains are the pattern-matching subsystem responsible for Command, Control and Communcications. Maybe expressing limitations in terms of circuits is roundabout to you, but it's perfectly natural for me. >it almost closes the curtain on approaching this natural order (of brain circuitry) with both humility and curiosity Again, why? is it because "Circuits" somehow imply rigidity or fixity ? that's not true at all, plenty of circuits can change their structure while they're operating. And even fixed circuits can be so complex and of so many signal paths that they seem to have a mind of their own, especially in a dynamic changing enviroment. I don't know about humility*, but the two grandfathers of neural networks, McCulloch and Pitts, were both very curious about the brain, and they used ideas from logic and cybernetics to study it. The Nobel-prize-winning Hodgkin–Huxley model is another framework that looks at the activity inside a neuron in circuit-theoritic terms such as capacitors and current sources. * : That said, "All of your views and opinions are the product of a few billion capacitors wired together" seems to me like a pretty humble way to go through life.
- inafewwords 5y ago
  Humans have a bit of irrationality that is based on their different POVs and experience. So the magic system is just that we have dynamic value systems. Shared culture through interconnected internet is making this less of a problem as the dynamic range is getting smaller or narrow enough to categorize
- brundolf 5y ago
  The problem comes with rigidity and certainty. Using a flawed system with the acknowledgement that it's flawed is well and good. But "mechanical" components like hard data and logic often give the illusion of being flawless and true in some absolute way. That doesn't make them useless, but it means you have to work extra hard to keep in mind the fact that they are and always will be imperfect.
mdp2021 5y ago
> Having the naivete Regulative ideals do not constitute a claim of achieved solution. Approximation of an ideal result is already an achievement.

Banana699 5y ago

> flawed mechanical system I don't like what "mechanical" is implying here, it gives the impression that humans have weird magic in their head that makes them understand language or search for truth (whatever approximate version they like anyway). It's just a flawed system, period. Your brain circuits are a flawed system, and the artifacts produced by the data science craze are flawed. They are not flawed equally, but they are all physical infromation processing systems trying to make sense of the weird jumble of signals coming to it from Outside.

yawboakye 5y ago
> Your brain circuits are a flawed system What a roundabout way to admit limitation both in knowledge and tools to investigate them. Except this one is more sinister because it almost closes the curtain on approaching this natural order (of brain circuitry) with both humility and curiosity.
- Banana699 5y ago
  >What a roundabout way to admit limitation both in knowledge and tools to investigate them. Why? I view humans (and all biological beings really) as machines, fantastic contraptions of unimaginable complexity that push the materials they are made of to the absoulute limits. Their brains are the pattern-matching subsystem responsible for Command, Control and Communcications. Maybe expressing limitations in terms of circuits is roundabout to you, but it's perfectly natural for me. >it almost closes the curtain on approaching this natural order (of brain circuitry) with both humility and curiosity Again, why? is it because "Circuits" somehow imply rigidity or fixity ? that's not true at all, plenty of circuits can change their structure while they're operating. And even fixed circuits can be so complex and of so many signal paths that they seem to have a mind of their own, especially in a dynamic changing enviroment. I don't know about humility*, but the two grandfathers of neural networks, McCulloch and Pitts, were both very curious about the brain, and they used ideas from logic and cybernetics to study it. The Nobel-prize-winning Hodgkin–Huxley model is another framework that looks at the activity inside a neuron in circuit-theoritic terms such as capacitors and current sources. * : That said, "All of your views and opinions are the product of a few billion capacitors wired together" seems to me like a pretty humble way to go through life.
inafewwords 5y ago
Humans have a bit of irrationality that is based on their different POVs and experience. So the magic system is just that we have dynamic value systems. Shared culture through interconnected internet is making this less of a problem as the dynamic range is getting smaller or narrow enough to categorize
brundolf 5y ago
The problem comes with rigidity and certainty. Using a flawed system with the acknowledgement that it's flawed is well and good. But "mechanical" components like hard data and logic often give the illusion of being flawless and true in some absolute way. That doesn't make them useless, but it means you have to work extra hard to keep in mind the fact that they are and always will be imperfect.

mdp2021 5y ago

> Having the naivete Regulative ideals do not constitute a claim of achieved solution. Approximation of an ideal result is already an achievement.