When you say (paraphrase)"it doesn't matter when you adjust the weights" you're clearly correct, but when you adjust the weights by training the effect of the training depends on the current state of the system. AFAICT, nobody goes around directly adjusting the weights.

The opportunity cost depends on your situation. I really should have devoted a bit of time to learning Japanese while I lived there, but instead I got along with a bare few phrases in Japanese, and usually using English. The opportunity cost would have been minimal.

On what basis do you say "That's provably not the case." to my claim that the training needs to be applied while it's developing? If you've got a good reason, I'd like to hear it. My opinion is based on things like "The chain of logic" not reliably representing the eventual conclusion. It's far from rigorous, but I think it's strongly the "best guess".

It's not clear to me, though, that GPT5 is good enough. OTOH, I'm going off second and third (or more) hand reports.

That said, there are already definite use cases, and those will only increase. The question is "how quickly?". I wouldn't want to play John Henry vs. the steam drill, but many seem anxious to do just that.

There ARE decent arguments. Different languages facilitate different ways of thinking. (The weak form of the Whorf hypothesis. Pretty clearly true, just as the strong form is probably false.)
OTOH, you won't get that benefit from a quick course. It probably requires years of immersion.

That depends on what you mean by "machine". It is perfectly reasonable to have a meaning of machine that includes these effects. And you're right about the difference between utility and intelligence. A screwdriver may have very high utility, but has essentially no intelligence. OTOH, slime molds *are* intelligent. Not *very* intelligent, but still, intelligent. More than that, they're goal-seeking intelligences. It's not clear to me that pure LLMs are goal-seeking except in a very limited way. But they do have a bit of intelligence. It's different than the intelligence of a human or a slime mold, but it's there. (Understanding, perhaps not, or perhaps one should rather say "only in a linguistic context".)

Except that alignment training is the WRONG approach. Well, unless the training is applied while it's developing. And tested in adversarial environments. (Yes, it would be nice to have that candy, but it's wrong to take it without permission.)

There's no plausible reason to claim that it doesn't involve "some kind of irrational quantum mechanical voodoo", There's just no reason to claim that it does. FWIW, photosynthesis involves "some kind of irrational quantum mechanical voodoo" that allows multiple photons to energize the same reaction.

OTOH, involving quantum mechanics isn't a claim to non-locality, except at a REALLY sub-microscopic level.

Personally, I doubt that non-classical mechanics is required for consciousness, but quantum mechanics might be used to make it a lot more efficient. To me it appears that you get consciousness when you have self-monitoring goal-seekiing.

It would be surprising if it wasn't shedding more heat than it receives from the sun:
Uranus is pretty distant from the sun, so it doesn't get much solar heating. Large blobs of gas tend to collapse, emitting heat in the process. Pieces of rock tend to have radioactive components. Etc.

That said, 12% seems a bit higher than I would expect.

I guess it did, and your response didn't clarify things. Suppose I listed the choices as
"Apple, Debian, Red Hat, Suse, Ubuntu, Windows"? Putting them in nice alphabetical order, and leaving out the choices that require a bit of tech knowledge. (I'm not sure about SUSE, but I suspect it's just as easy to use or install.)

The real limit is that most people don't install an OS, so they just use whatever their computer came with. The last time I installed windows it was harder to install than Debian...of course that was back in the 1990's. I haven't touched their updated TOS, and even in the 1990's I had someone else click through the requirements. (OTOH, Windows 95 and 98 were relatively easy to install...but then you needed to get all the drivers installed. So Debian was still easier...but that's not fair, because the Debian I'm comparing against is at least the 1998 version, and often one slightly later.)

So use Red Hat, Debian, Ubuntu, or (possibly) SUSE.

I don't like Gnome, but if you want what somebody else chose for you, just pick the defaults. (Personally , I use either Mate or xfce.)

Claiming that it's important to tight Elon Musk is not a sign that the results are untrustworthy, but it may well be taking as being extremely partisan. OTOH, anyone who trusts Musk is as intelligent as someone who trusts Zuckerberg.

Which "The Guardian"? Certainly the one I kept getting in my mailbox was NOT reputable journalism. Perhaps "The Manchester Guardian" is reputable. But I've never seen a copy. And if the US edition of "The Guardian" (without adjectival mods) is the same company as the British edition, then neither one is trustworthy, as they just say what they think will sell. And they occasionally directly contradict each other.

IIUC (dubious) the 'next version of Siri" is actually an entirely different program, just one that fits the same advertising slot. So you really can't judge it based on the current siri...OTOH, judging it based or rumors and PR is, at best, dubious.

That's not plausible. There are too many choices...or do you mean "just let them decide what fits on the first page"?