Re: Clever stuff
True - neither LLMs nor any other form of ML is "intelligent" in any way that we like to ascribe to ourselves.
BUT the large part of that line of argument comes from the usual misrepresentation of a field of investigation and the name it chose; misrepresentation that is inevitable given the behaviour of humans in certain fields[1].
So we have people who misrepresent "the Big Bang"[2], "Evolution" - and the modern favourite, "AI". People who work on the Big Bang are not looking at/for any kind of detonation[3], evolution does not come in flavours that you can pick and choose[4].
The field of AI (as it was described[5] when I started taking the classes) is about looking for ways to create useful, or at least interesting, processes that can be applied to perform, or aid in performance of, behaviours that appear to require intelligence. That is, to find any ways we can MIMIC those behaviours; there is good reason that the Turing Test is also referred to as the IMITATION Game. When we have some machinery that does a decent job at something it is made glaringly obvious that it is ARTIFICIAL.
Ok, there are confounding factors, such as "if it works, it's not AI (any longer)"; i.e. the tendency for AI Lab creations to be applied out in the Real World and become mundane and (shrug) "well, yes, so what?". Remember, Back In The Day, playing chess was something only humans did (it was one of the "open challenges" in my book of "100 programming exercises", way after "print the Fibonacci sequence, with and without recursion"[7]) and now we have Toledo microchess born from an IOCC entry, a far different goal. It would be nice if people gave the attribution ("I'd like to thank the Academy for this award, and the tireless AI boffins, without whose heuristic search my runtimes would have been unfeasibly long"), hey ho.
Other confusions: people just throwing around "smart" and "intelligent" around willy-nilly (home automation, 'phones). And a scarily large number of people who seem to be really, really angry at finding out that the human system is "imperfect" (i.e. has limitations) and get upset at finding out there are mechanical methods of doing things they find hard or impossible (thankfully, most now seem ok with the pocket calculator) and/or that they can get so easily tangled up by interactions with simple mechanisms (oh, ELIZA, why, why?). Or, so much worse, "simple": "I am better than 200 lines of code!". Um, have you considered all the human intelligence applied to allow those 200 lines to exist and execute?
My conclusion:
Trying to split up or do away with the phrase "artificial intelligence" is as sensible as trying to be rid of the phrase "artificial hip": no surgeon or biomechanic is trying to claim that that lump of titanium is a "real hip", any more than any (non-misrepresentationalist) AI bod is trying to claim[8] that software is "real intelligence". But both AH and AI are doing a sterling job helping someone/something to run.
[1] basically, anyone who believes they can make money by misrepresentation: salesmen, hucksters, fraudsters. Oh, and philosophers - and even more "philosophers" - who keep trying to muddle things up by trying to imply they have anything useful to say about either physical reality or mathematics; for money.
[2] all in quotes to indicate their misrepresentation; without quotes, these are all perfectly respectable fields.
[3] there are a *lot* of those to study, but they all came a lot later! And we should all be very glad that they did.
[4] it doesn't cover abiogenesis at all and the attempts to create a split between "micro-evolution" and "macro-evolution" are complete twaddle.
[5] probably in a lot better fashion than I'm going to manage here, but also over a longer period of time: like any introductory class, we started with flashy[6] demonstrations before getting on with it
[6] sadly not as flashy as the rocket scientists get; but then they do have to take their maths more seriously, as getting it wrong tends to do more than waste a chunk of runtime, so let them have their fun.
[7] so, did he buy that as a first edition the day it came out or pick it up years later? Ah, the mystery deepens.
[8] yes, yes, you get them all "tired and emotional" and they'll probably wax philosophical; which is as useful to reference as the fact that you're my best mate, you know that, I love you, I really do (clunk).
