Sounds a bit like those Anthropic researchers who keep finding new ways Claude did something unexpected and scary every other week.

We don’t care whether you’re scared or amazed : TALK ABOUT IT.

So… this is still a ridiculous case, but they’re wealthy enough they aren’t too worried even if they lose it? All right.

In fact they should put random odds on whether you actually get extra power, so the customers can enjoy some surprise mechanics too.

I think those games that are made to be boring and absurdly grindy and then offer you to pay to skip the boring parts are even worse. And they’re not limited to phone, too.

“Pay to not play”, when we ensure our gameplay loop is so bad that you literally think your time spent in it has negative value.

The second one is not really a way to check if it’s AI, only if it may be deceiving you, and the third one’s conclusion is not “yes” but “use responsibly”, like it’s in the power of the common person to even choose to use AI and like corporations aren’t the ones pushing it with no regard to impact anyway.

The problem is those 3 questions are very vague and would need complex answers, and maybe the guy vould have been able to give these, but in any case they’re not in the article.

Theoretically in the way some particles have been theorized to maybe exist according to physical models but have never been observed.

I’ll save you a click. that article asks 3 basic questions : is it dangerous, how to tell something is AI and is it bad for the environment.

They get only non-answers. Thanks, BBC.

Yeah, part of the usual “it’s not bad, you’re using it wrong” arsenal. Definitely not the clever hack they think it is.

This probably has as much potential to create new errors as to find old ones. LLMs are trained to be “helpful”, if you tell it with total confidence something is wrong, it will answer like there is something to correct, and anything will do.

So even if it had something about right to begin with, now it will thank you for your “insightful” question and output some bullshit to please you.

I may have been on lemmy for a while, but I only got seriously into linux for like 2 months. Part because I am fed up with the AI/advertising bullshit in recent Windows, part because of end of support for my hardware. So, that definitely happens.

My previous personal desktop linux attempts were like early 00s and didn’t last long so I don’t think I’m officially part of the cult yet. I work with linux, I tinkered a bit with RasPi debians, but it’s the first time I am really considering it for my all purpose PC.

I still have a PC on W11. I am not in a hurry to convert it, because there is still stuff stucked on windows that I don’t expect will be easy to replace. Like Virtual Desktop.

I think I’m too sober for this shit.

One of my 2 PCs already switched to linux because of mandatory TPM 2.0 for windows.

It’s like they want us to leave their ecosystem. Requiring hardware changes in the middle of a major component crisis, one that they are in no small part responsible for, is certainly a choice.

I am ready to believe those that allowed this law to pass were.

Well, another big hint is how the thing answered by addressing a username that wasn’t part of the exchange, twice. And then messed up the “@” when they pointed that to it.

If it’s even manually copy-pasted, the guy doing that didn’t allocate a single braincell to what was being discussed.

The great thing about asking gen AI to look for problems, is that it’s so helpful it will create new ones for you.

Like arguing for hours that if you were to remove safeguards from your code, it would become unsafe.

https://hackerone.com/reports/2298307

Oh cool, now there’s a new way of using AI to destroy the environment. Old one wasn’t deliberate enough.

would increase costs

For us? Microsoft licences aren’t exactly cheap, keep getting more expensive, and every single basic functionality is an extra.

cybersecurity risks,

Trust the megacorp with a huge target on its back that will deny any responsibility when it fucks up instead.

limit AI and cloud services

Sounds exclusively like a “you” problem.

expand government control in ways that can undermine civil liberties and enable censorship,

Sure, give that control to the US government instead.

Fuck that bullshit. Hopefully no clueless politician falls for it.

Not a company since I’m in public administration, but my structure has a few thousands workers, most of them having access in some form to the network.

They do filter our internet. I don’t give a fuck whether people consume porn with their own devices and connections. But if you can download porn, you can download anything, including malware. And a bad actor having access to data on our network would be disastrous.

Unfortunately, meta has that kind of data too. In fact hoarding private data is what their business is about. Not securing their network is criminal.

Absolutely. Or, as they say, “sporadic” amounts.

I don’t think I really care who wins that one, but :

Meta responded in October by filing a motion to dismiss, arguing the sporadic downloads were consistent with ordinary ‘personal use’ by employees and visitors on the corporate network.

Oh, yeah, just your ordinary downloading porn on the corporate network of a tech giant megacorp, as you do.

Either a lie or an admission of baffling incompetence.

The collective works of billions of people are up for grab to train a LLM without their consent, but a couple gigabytes of responses collected from another parrot machine is theft.

Sure, Anthropic.