2 ms·
In the present, sentience is a philosophy question, nothing stops neuroscientists from practicing a little philosophy, but they (as in the author) should be a l
by colechristensen 4y ago
In the present, sentience is a philosophy question, nothing stops neuroscientists from practicing a little philosophy, but they (as in the author) should be a little more cautious when asserting that ML models being somewhat similar to human brains means they can be sentient in their current state. There is no science that can currently be done with sentience beyond basic psychological games, there's no objective testable theories where you can point to something and convincingly say "this is sentience" analytically.
Plenty of current neuroscience fields were once only a question for philosophy but have graduated into matters of science (Kant has plenty of opinions about things which can easily be replaced with actual neuro- knowledge, for example)
- mannykannot 4y agoI am completely in agreement with your second paragraph - and with your first as far as it goes, though for balance, I think we should cast an equally skeptical eye over what philosophy has achieved during its tenure over this topic. Personally, I feel it spends way too much time worrying about zombies. There's a slightly ironic moment where the author mentions the pushback Ilya Sutskever got for tweeting "It may be that today's large neural networks are slightly conscious." This is not a view to which I would assign much credence, but it is more or less axiomatic (though in a different sense, perhaps) in some forms of panpsychism that I am also rather skeptical of (what on Earth does Goff mean when he says that electric charge is a form of consciousness?)