Not really. Diamorphine is a precisely described drug. Heroin might be nearly anything, down to crushed up Draino. Many reports describe it as being cut with fentanyl, which is also a highly useful drug, but the "heroin" that's been cut with it frequently kills people.

You are incorrectly generalizing. There are lots of use cases where AI improves medicine. (Possibly at excessive cost, but still, improves.) But don't expect a generalized ChatBot to provide that improvement.

Whether it is done by "the State" or not may depend on your definition of "the State". If it's done by a corporation, and the government has access, then I feel it's done by "the State". Claiming it's actually done by "a private corporation" if petty word-play, when the government controls which corporations are allowed to exist and what they can do (and how profitable they are ... see tax law).

FWIW, "The power to tax is the power to destroy!", so corporations are not independent of the government.

I strongly suspect that the criteria that you claim to be present are only in effect when the person accused has a decent lawyer. There have definitely been multiple cases where a "public defender" has acted in collusion with the police to the detriment of the person he was ostensibly defending.

What makes you believe that people aren't being thrown in jail for crimes they didn't commit? There's absolutely no way to quantify that statistic, but there are absolutely SOME people thrown in jail for a crime they didn't commit. Sometimes on the basis of faked evidence. How often? Nobody knows. It sometimes gets proven, and that kind of evidence is clearly hard to access.

FWIW, it does *seem* to be a rare occurrence. But whether it actually is rare is...unproveable.

I used to think that, but now I believe that BOTH things are going on at once, and they act in a reinforcing manner. With different people, different motives are dominant, but as an aggregate, they are mutually reinforcing.

FWIW, my profession was computer programmer. I was also an artist using various traditional media. (Not professional grade, but not bad, either.) I didn't like it for itself, but only for social reasons.

So....
Artist is an ill-defined term, but since any piece of garbage text is (automatically) copyright, I see absolutely no reason that a cleverly manipulated bunch of pigments shouldn't be copyright, no matter WHAT tool was used to create it. And no matter how *I* rate it's esthetic appeal.

OTOH, what this really means is that I think the copyright laws are a foul mess, and should be repealed. But with the laws as they are currently implemented, there's no reasonable justification for refusing to grant a copyright.

That's "would be legal under this act", but if, say, ChatGPT collected non-public data (I believe it does, but probably not about rents) then iwould it be legal? And if it merely collected "a wide span of data in various formats" it could plausibly have the same practical effect. So some would be inspired to write an additional piece of legislation.

AI will probably add some interesting twists to this. Will it be illegal to ask ChatGPT "What should I rent my house for?"?

I've been through the Lenovo series of connectors. I worked at Intel for 21 years and they were a Thinkpad/Lenovo house. I got my mother a Lenovo on the grounds they're a bit less prone to fall apart than other brands. Macs weren't in the running because she uses windows specific software. I have my yellow tipped barrel connector adaptor secreted in box for when the occasion to use it arises.

The (quite new) MacBook Air I'm typing on has 2 USB-C and one Magsafe 3. The real issue is I have only a single Magsafe 3 cable while I've got lots of USB-C chargers and cables so when the other ports are occupied, I'm schlepping off to find the one single cable. I tend to run lots of CPU heavy jobs, so it's the higher power brick for me.

But that ARM CPU. Ugh. I've hated the ARM instruction set since the Archimedes. It's not got better. They messed up the RNG instructions ( https://developer.arm.com/docu... ) stipulating 90C-RBG3(RS) structure which is the wrong choice for an instruction-as-full-entropy-source (Like RdSeed on Intel). The 90C-RBG3(XOR) is the right one since the RBG3 doesn't block the RBG2 with the XOR construction. RISC-V made the same mistake in their drafts, but they listened to my arguments and fixed it. ARM wouldn't give me the time of day and so here we are with broken specs for ARM. I wouldn't care if engineering RNG things wasn't what I did most of the time.

There's always the Framework running Linux when I want to retreat to my happy place. 6 ports, all configurable, X86 CPU and das blinken lights on the keyboard.

I don't really think you can trust everything they say to be a lie.

That's harsh. But the do need to remember what the term "bleeding edge" means.

Sorry, but you're wrong. LLMs are already quite useful in science and math...but not with the internet as a training basis. The LLMs that are useful in sciences need to be specialized for the particular science, and they need to be used as hypothesis generators rather than as answer givers. They can also be used to filter out noise (though in that case there's more need to beware of errors).

I'd call that an "if" rather than an entailed consequence. There are LOTS of ways to maintain a monopoly.

In 1960 "made in Japan" was an insult. In 1970 it was a compliment.