> I've also heard Cory Doctorow recently offer a similarly dismissive view, describing AI as "just statistics".

Well, AI partisans have applied grandiose terms like "thinking," "intelligence," and "soul" to these machines. It's not wrong to push back and remind people what they really are.

Well, if you look into the current scientific consensus on human cognition, our intelligence is also "just statistics."

So I guess we all agree, except that some people think "just statistics" is derogatory phrasing!

What a software engineer comment.

That is absolutely not the consensus in neuroscience.

Was thinking more about cognition than neuroscience, but sure, even there. Think of any counter-examples -- I'm guessing things like symbolic reasoning or pre-existing structures in the brain -- and then consider the origins of those concepts.

Admittedly I'm an absolute layman just dabbling in this, but I could find not a single concept that cannot be traced to statistical origins. Like, symbols are essentially "vectors" of thousands of neurons firing simultaneously, and the pre-existing structures in the nervous system originate from evolution, which is a statistical process.

"Just statistics", as far as I can tell! ;-)

To a hammer, everything looks like a nail.