@xeow And even those 0s and 1s are really just another abstraction for the presence or absence of electrical current.
Stimulus->response
No stimulus->inert
@xeow I have a feeling you're right. And I guess that—apropos of your original post—is what we're afraid of.
("He has too much of his father in him.")
—using em-dashes since before they were cool!
@xeow For that matter, is biological sentience any more "true"? Doesn't it derive just as much from its underlying electro-chemical infrastructure, however incomplete our knowledge of same? Or is there some extra "spark" that robots can never have?
(And how would we know, if so?)
@xeow I'm even less an expert—I have about a late-90s grasp of computer science. But AI/LLM technology—albeit massively abstracted and miniaturized—still all boils down to basic binary digits, right?
@xeow And as follow-up, is there a difference? Is sentience distinguishable from a perfect facsimile of sentience? Or are they as much the same thang as 0.999… is the same thing as 1?
@xeow I guess I mean, is the asking itself—not merely the content of the questions—somehow more spontaneous? LLMs are designed to converse, but can computers become inquisitive other than as the (intended) result of their programming?
Does that, like the Kaylon, approach sentience?
@xeow Are the questions asked by LLMs somehow more…organic…than those asked by basic computer programs according to their instructions?
Are these the computer's "own" questions, as opposed to those asked as agent of its programmers?