It depends on how you define the term. I tend to consider any choice an act of reasoning (including a simple if test). I know that most people have a different definition, but I can rarely get them to define what they mean by the term. I tend to suspect it's an "I know it when I see it" kind of thing.

Are they not forgotten now? I never heard of them before this story.

Umnnh...but I think the programs that do that are different programs than the ones that understand source code. (Well, not *that* much different, but trained on a massively different data set.)

FWIW, I think understanding binary code is probably an easier problem for an AI than understanding source code, but it *does* require different training.

You could be right, but my take was that it was made by a manager who had no idea. Of course, they could both be true.

FWIW, the albums I play are played via computer, and I edit out any songs I don't want. If they put in some stuff I don't want, I just don't play it. I feel like I'm paying for the CD, not the songs. They never fill the CD, so if they include extra stuff, I just don't care.

FWIW, this is a lot better than LPs were, where you basically had to live with the selection they chose.

FWIW, my "legal free offerings" are purchased CDs. I don't know how many feel the way I do, but I don't buy "temporary goods" unless you count food and medicine...and they come with stated expiration dates.

If you don't know how to invest in AI, just ask a Chatbot for advice.

I hadn't realized that was an ad...actually, I hadn't noticed it until you mentioned it. At least the MongoDB as has gotten less obnoxious. (I though MongoDB was free software, but maybe all the entries in the Debian repository are for drivers or interfaces.

OTOH, that's a hot enough temperature to make Data Centers in orbit seem a lot more practical. Just because you *can* run a 700C doesn't mean you need to. I wonder how radiation hardened a chip with that technology would be.

Actually, making following orders the second law isn't that unreasonable, but perhaps it *should* have been the third law, or even the 5th. The "paperclip maximizer" is an example of a robot that ONLY worries about following orders. You can always trust that there will be at least one person who gives a stupid/dangerous order.

That, of course, is a real problem. Currently AI only knows what it is told. This is a systemic weakness that can't be solved with more words, but requires "direct experience". Robots will have that, but ChatBots, probably not. ChatBots appear mired in a nest of hallucinations. (I.e., when people write, they aren't telling their experiences, but only an abstraction from their experiences. I don't think there's any way around that.)

The problem is, the AIs don't have the same motives that people do. They don't really have access to those motives. All they have is words...which bear a relationship to those motives, but it's often a pretty abstract relationship. So protein folding is easier than personal advice.

Maybe they do streaming backups, and he just duped the stream.

It might be a hallucination, or it might be a real problem. And there are other possibilities. (E.g. earlier it was suggested that MS noticed a bad bug *somehow* and the government didn't want the bug to be fixed.)

If you want to be fair, it's been headed that way ever since the 1860's. And prior to that the individual states were headed that way.

People in power like to make their jobs easier.

"Security by obscurity" doesn't work by itself. It's a necessary component of every security policy, however. You can't just pick one. (It's called "defense in depth", but that's not really a good metaphor.)