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"?

I'm not sure you can trust all versions to have the same flags set...in fact I'm rather sure you can't.

It's already happening. It's just that the actual uses are being drowned out by advertising/crap/PR/??.

There isn't just one flow of things happening. Most of the noisy ones are crap, but they aren't the only ones.

If that's really your definition, you're being silly. Randomness is easy to add. Even genuine quantum randomness if reasonably easy, but it requires generating and filtering noise.

OTOH, it is true that randomness is part of the human though process. I just don't think it's part of rationality.

Batteries are probably approaching their limit. I doubt that another factor of 2 reduction in size/power is possible. Super capacitors seem to have stalled out (probably on fabrication problems). I haven't heard anything about hydrogen storage in zeolites for a decade (of course I haven't been looking). Wireless power transmission pretty much needs to be point-to-point, or there's no way to bill for it.

OTOH, we really need a good power store that doesn't release carbon dioxide. All I've been able to think of is battery swapping, with the battery essentially being owned by the government. So standardized connections across multiple companies, and no problems with which battery is in the car. But that needs a reliable charge-meter. (I think fast charging is destructive of batteries, and even if not it poses huge problems to the power grid.)

it's wrong, but essentially correct. The important figure here is the distance from "room temperature", because that's what thermal dissipation depends on. You get technical when deciding on the materials to use, etc., but this isn't about that (except for the claim that they could use cheaper materials).

What is your definition of "reasoning" that you distinguish from pattern matching?

Do you have any evidence that pattern matching isn't what you do? (I really doubt it.)

That said, LLMs are, by definition, working in a restricted domain with no non-language feed-back. This creates inherent restrictions. Robotics has more promise, but it's a lot more expensive, and harder to train, so the results have been slower. But they can use a lot of the same technologies.

Also the claim that if the problem is too complex, the results are mediocre seems to fit humans, too. He's working with a toy system, but the pattern seems right.

Additionally, if the heavy compute stuff is done in a compiled language, for efficiency, it doesn't matter much that there's a slow Python layer on top of it. That's often limited by I/O speed rather than by compute speed anyway.

ISTM that you can't pick up hydrogen along the way unless you're moving as close to the same speed as the local drift.
OTOH, fusion does promise a lot more power/kg. whether it can deliver that in a controlled manner, however, may be another question.

I don't consider that a "kind and gentle" response, but if he described the code accurately, it does seem warranted.

If I parsed that correctly it was the eggs that were green, not the ham.

If it's made out of mushrooms, it's not coffee. OTOH, mushroom flavored coffee might be quite good...it would really depend on both the coffee and the mushroom.