To your last section - Matthew Crawford in Why We Drive actually makes reference to the idea of "the world is it's own best model" and how this is influencing a new wave of AI research.
Is this in Why We Drive? Phrased like that, it is quite realist, which is good. But if data-driven AI is blind to the fact that data is not the world, then you get a very postmodern AI.
Gotcha. I get the massive practical difference between neural networks and logic, but not the philosophical difference. Numbers, the things that neural networks operate on, aren't out there in the world, so there's still some modeling.
But that's quite similar to how our own sense organs work, so there's that.
To your last section - Matthew Crawford in Why We Drive actually makes reference to the idea of "the world is it's own best model" and how this is influencing a new wave of AI research.
Is this in Why We Drive? Phrased like that, it is quite realist, which is good. But if data-driven AI is blind to the fact that data is not the world, then you get a very postmodern AI.
The idea is more along the lines of "don't presuppose a model". Which is what neural nets (at least try to) do.
Gotcha. I get the massive practical difference between neural networks and logic, but not the philosophical difference. Numbers, the things that neural networks operate on, aren't out there in the world, so there's still some modeling.
But that's quite similar to how our own sense organs work, so there's that.