I think you're wrong, though it might depend on where you live. I believe that in the US owning a CD doesn't give you the right to commercial activities involving playing that CD. You can argue that it *should*, but I don't believe that it does.

Every decent song has ALWAYS been full of plagiarism. Liszt copied music from the Gypsies. Irish Bards used to be *forbidden* (by their guild) to create new tunes. They were only allowed to set new words to old tunes. Copyright law is an abomination in the realm of music. (In some other places I'd argue that it was just too long, but in music it just shouldn't exist at all.)

I'd often prefer that they didn't have any, but that depends on how loudly the people at the next table are talking.

You're confusing motivation with intelligence. They are separate things.

OTOH, the definition given is silly.

An AGI would be something that could learn anything. Such a thing is probably impossible. Certainly people don't meet that measure.

There's probably no standard format for the input data. It's already been claimed that Google could have checked if the cases exist. (I've never tried, so that *could* be wrong.)

Your model of an AI is wrong. EVERYTHING an AI says is a "hallucination", it's just that a lot of those hallucinations match reality. Even "Stocastic Parrot" is closer to correct than "An AI just regurgitates what it finds in its data set".

(P.S.: Most of your memories are wrong. They're also "hallucinations" that sort of "match reality". One test of this is to list everything that's in a room you haven't been in for awhile (say at least 10 minutes) and where in the room it was located. Then go to the room and notice all the things you forgot, or added, or misplaced. No excuses allowed like "but that wasn't important".)

What, in your argument, is the difference between LLM copy-edited text, and for-hire human copy-edited text. The editorial services I have seen *sometimes* try to find editors that are kinda-sorta near the correct field of expertise, but there's no guarantee you'll get someone who even has a passing level of familiarity with your field, and for some services, all they have is a degree in English.

So, again, what's the difference between linguistic polishing by machine and linguistic polishing by semi-qualified human?

Following up on that idea, there are various copy-editing services that many non-native English speakers use, and are encouraged to use, to help improve their writing. The main difference from the perspective of forensic detection with AI-copy-edited text is that there are a very small number of such styles compared to the likely thousands of copy-editors' individual styles, making automated copy-editing easier to detect. I'll bet dollars to donuts that if you trained an LLM on the output of a single human copy-editor, you'd be able to identify all papers that used their services.

We use AI to help with paper writing in my lab, mostly because there are only two native English speakers, and it relieves me, the lab head (and one of the two native speakers), of having to do extensive copy-editing in order to make stilted English more readable. I still read every word that gets published from the lab, but using AI for copy-editing is no different from using a human-based writing service to fix poor language. It's just cheaper and orders of magnitude faster.

So, for us, the response would be a big, "so what?" to this report.

But, if people are starting to use AI to write entire papers, that's a different story. My experience is that current models hallucinate ideas and, especially, references, at far, far to high a rate to be seriously useful as anything other than a tool that requires full, manual verification. I half-jokingly say that if a paper is hallucinated, that means the AI was unable to find the right citation, and it represents a gap in the field's knowledge that we could address. The amazing thing about the hallucinations is how convincingly real they sound: the right authors, the right titles, the right journals. These are publications that *should* exist, but don't, at least in my experience.

As a most recent example, when writing a grant application, I tried to find citations using an LLM for an idea that is widely-held in the field. Everyone knows it to be true. It's obvious that it should be true. And, yet, there have been no publications as of yet that have actually discussed the idea, so the LLM dutifully hallucinated a citation with exactly the author list you would expect to have studied the question, a title that hits the nail on the head, and a journal exactly where you might expect the paper to appear. I've told my staff that we need to get that paper written and submitted, immediately, to fill that obvious gap, before someone else does. It will likely be cited widely.

You're talking about current AI. There's scant reason to believe it's going to stop improving. AI molecule folding just keeps getting better. Larger and larger systems are being "understood" correctly (i.e., the predictions about how they will act are borne out). In many areas AI is already better than "experts in the field".

You *could* be right, that it won't get better at your job, but you tempt me to call you "John Henry" ("John Henry drove 16 feet, and the steam drill only drove 9" and "he laid down his hammer and he died").

Well, advanced lithography equipment isn't easy to make, so it's not surprising they're having problems. If they solve those problems it will be a permanent benefit to them.

Also, there's no particular reason to believe that "the AI bubble" will pop. Certainly parts of it will, but other parts are already solid successes. The rest is "work in progress", which, of course, may fail...but the odds are that large portions will succeed. (Much of the stuff that's "not ready for prime time" is just being pushed out too quickly, before the bugs have been squashed.)

I agree entirely. I teach an intro to programming course at one of the well-known universities. It is a lab course with 2 hours of teaching contact time per week, 2 hours of reading time per week, and 8 hours of expected programming time per week. The students learn by doing.

Aw, c'mon. Basic is a horrible, horrible language that teaches poor habits and has almost no translation to serious programming.

If you want to have kids learn something easy that at least teaches good organization and thought processes, then teach them Scheme.

Micro-reactors create more nuclear waste per amount of power generated than do standard reactors. But there are special use cases where they would be preferable.

Well, the army is probably only significant to it's neighbors. Transporting lots of troops is difficult and expensive. The navy and the air force, however...not to mention the hypersonic missiles.