3 ms·
Quicksort is a legitimate model though; there are objective answers about what is a quicksort and what isn't. Two people will never look at the same processor r
by lmm 6y ago
Quicksort is a legitimate model though; there are objective answers about what is a quicksort and what isn't. Two people will never look at the same processor running the same algorithm and disagree about whether it's a quicksort or not (or if they do, they'll be able to localise that disagreement down to a more specific question like whether using the cache in a particular way qualifies as using extra memory). Whereas two people might look at the same processor running the same algorithm, agree about all the measurable facts and all of the lower-level definitions they were using, and still disagree about whether it was conscious.
- geofft 6y agoWe can find plenty of fuzzier abstractions. Does "machine learning" exist? Does "NoSQL"? Does "scrum"? (That one can't even be used to describe an integer, and it certainly doesn't describe a physical object.) I'd claim that if two people agree about measurable facts and disagree about whether something is conscious, they do in fact disagree on the definition of consciousness, just as if they disagreed on whether something counted as NoSQL or scrum. (Though it's been a while since I paid attention to this area of philosophy, so I might just be wrong here - is there an example of two people claiming to agree on the definition of consciousness but reaching different conclusions?) Honestly maybe this was the brilliance of the article. We don't even need to agree on the definition of "breakfast" to conclude that it's a useful and meaningful term. Nobody would claim breakfast doesn't exist on the mere grounds that it's made up. Of course it's made up, it's an abstraction.
- lmm 6y agoI see what you mean - I'm not sure if it's a question of the model itself or how well it's communicated, or whether that's a meaningful distinction. Perhaps a better model would be that there's a spectrum of how well-defined a model is: for a good model everyone agrees on what it does or doesn't cover, a mediocre model is one where people agree about some cases but disagree about a lot of edge cases, and a bad model people don't agree about at all. And then the claim is that "consciousness" doesn't actually add any understanding or communicate anything, because people don't really agree about any more cases of it (beyond being just a synonym for "human").