Damus
MichaelJ · 6w
Something that grinds my gears is when people talk about the human mind as if it is like a computer. It's actually the exact opposite. The computer is like the human mind. The ancients and medieval...
Daniel Wigton profile picture
I am guilty of this, but I think there is something to be said of trying to understand something too complex in terms of something simple. I don't have much insight in to the workings of the brain without resorting to comparisons with computers or LLMs.

On recent thought experiment for me is how neuron activation works compared to an AI dot product. Turns out that it is more of time weighted average with locality effects. In other words, it is a super complicated differential equation with temporal and spatial compents that is really hard to model in a computer.

Also the brain has no trigger or drfinite output state similar to a token. Instead it is in continual motion and sense perception merely adds a perturbation that might affect the existing pattern.

It is fascinating, but I don't think I would have known what questions to ask without thinking of the digital version first.
MichaelJ · 5w
I think the fact that we've created machines that reflect the image of our own minds does indeed make those machines a useful analogy. To you point, we can use computers as an analogy to figure out what questions to ask about the human mind much the same way we can use a scale model of an aircraft ...