Fusion has not been "around the corner" except in the fantasy of bad journalists. It has only bceome clear in the last few years that it will be possible with regards to plasma physics and material sciences without extreme effort. Commercial viability is likely but will take aditional decades. This is what actual experts have been saying all along.

And no, these are not "yahoos" or "no-name". These are actual scientists. Your disdain for them lets me suspect you are a no-clue MAGA.

I see the denial is strong with some of the moderators. Come on, waste some more of your precious mod points on me! Not even the religious fuckups ever managed to decrease my karma and you for sure will not.

Well, in a commercial setting, several concerns apply. First, it may be illegal for you to actually have the code you work on leave your enterprise. This can be soft (breach of contract) or hard (criminal act), or in between. Second, that backend may give your code to others. That may cause a lot of issues. If you did this voluntarily, trade-secret protection may be gone. If there were vulnerabilities in there, attackers may get access to and then craft exploits. I am sure other problems may exist. And third, what if you get code that somebody else had leaking this way? Suddently, you may be using code under copyright by somebody else. Or take, for example, that you bublish the code, attackers find a vulnerability and the original source gets attacked and finds out what you did. You may be legally in the clear, but years and years of litigation could result before that is established.

So, yes, the model is a massive problem. But so is sending your code to it or getting code from it.

Why can't it be both?

Obviously, this is not "transparency" in any meaningful sense. But the LLM cheerleaders are not very smart, so it probably will convince some of them.

Not quite. There is and had been a lot of plasma physics and material science research going on, all targetted at making commercially viable fusion a reality. It has been slow going since so much of this stuff is new and requires real research not just engineering. That "detail" is an end-goal to be attained when it realistically can be. Which is not yet, but it is getting closer.

That said, this is another hurdle taken and things moved a step. It also means the understanding of things has moved a significant step forward.

I know assholes like you like to ridicule fusion research. But all it shows is that you have no insight.

Yep, keep dreaming. In a few years this hallucination will be over.

That is not what I said and you know it. Why are all you nuclear fanbois soft in the head? Oh, well, that is really redundant at this time.

Being a nuclear tech does not qualify you to judge economic aspects. Seriously.

You are just trying to confuse the issue now. Pathetic.

Bla, bla, bla, you have any facts? No, apparently not.

The whole idea is deeply flawed and apparently cannot even be implemented securely.

Every pen-testing situation is different and a proper test starts with a risk analysis and interviews for that. Sure, AI can replace already automated ElCheapo simple pen-tests, but that is it. And that will likely not last.

Hence just another meaningless AI stunt.

I can't comment on where sudo itself lives on the spectrum from aggressively solid implementation to really-dodgy-smell-around-the-edges; but it seems like its purpose is a fundamentally tricky problem even if its execution were impeccable.

The basic "user is authorized for root; but we'd prefer he be thinking and logged when he uses that authorization" is reasonably cogent use case; but it's more of a reminder than a security barrier. Then you get into the actually-interesting attempts at limited delegation and determine that you'd basically need a different userland for a lot of purposes: aside from the modest number of things(often with setuid already in place) built specifically to carefully do a very particular delegated function on your behalf and provide you with nothing else if they can help it; very little aside from garbage kiosk UIs or web or database-backed applications with user and permission structures mostly orthogonal to those of the underlying OS actually tries to constrain the user's use of the application(within whatever context that user is operating; generally having a privilege escalation is considered bad).

Half of what you run considers having an embedded shell to be a design feature; so including any of that on the sudoers list essentially means being able to chain arbitrary commands from that sudoers entry; and the other half doesn't outright intend to include a shell but would require some really brutal pruning, likely of important features, to prevent being able to chain a couple of interactions into having the ability to run whatever. And that is assuming that sudo itself is working entirely correctly.

Nope. That hallucination was already pushed in children's books 50 years ago. Never panned out and will not magically do so now.

I think your outlook is to positive. My take is it already has become mostly impossible to get new general training data sets. (Specialized ones you can, but at huge cost.) Hence the training data is now slowly aging into irrelevance.

Naa, don't worry about it. Like _all_ previously AI hypes, this one will fizzle out and only smaller things will remain. Most jobs _cannot_ be replaced by AI. Where it is possible, you typeically need higher qualification people to guide and monitor that AI and these people are rare. The only reason why the hype is even going is greed, ego and stupidity. As in all the previous AI hypes.