Microsoft is already releasing their next console as a full PC and it will run Steam games. Probably for less than what this will cost.

I was going to argue against needing 2.5Gbe or more but then I remembered the size of a lot of Steam games. It will be faster over Wi-Fi in some cases.

It's even in the summary:

This is something a Windows PC does not have, but Xbox would.

Sure as an argument about savings, it only works on the people who think those numbers are big. But if you realize how small those numbers are, a penny is completely worthless and unnecessary.

The half dollar coins are inconveniently large. Just get people to use the dollar coins we already have. 4:1 is an OK ratio for the next step up in coinage.

Certainly not at the five and dime store.

If the total with tax on a dollar is $1.0875 now, they charge you $1.09 at the register. What difference does it really make if it becomes $1.10? You are overestimating the value of 1 penny and the few seconds of labor it takes to generate a wage of a penny.

At least with a traditional PC, you can upgrade parts. Selling for the cost of a PC but not upgradeable sounds really bad. The Steam Deck at least offered a form factor that required a unique solution.

This is a world where even Microsoft wants out of the console game. The next generation of consoles will be Windows vs Linux desktops connected to a TV. And most of the Steam games at least run as well on Windows without Proton. And then maybe Sony decides to sell games on Steam and Xbox, so their console is essentially unnecessary. And finally, Nintendo will still be just doing their own thing.

At least they acknowledged how bad HDMI-CEC implementations are. This is probably a bigger deal than most realize considering it could wreck their perceived quality if they get it wrong.

Actually, that's probably the right idea. Not blocking paste, of course. But as it is, a file downloaded from the Internet has attributes to mark it as potentially unsafe. The clipboard could probably handle such an attribute. And when pasting into sensitive areas, that flag could prompt a warning confirmation. Only a minor UAC type annoyance to the end user but at the very least a chance to validate. And if the user has an antivirus, scanning that string and the domain names in any URLs when pasting into sensitive areas wouldn't be a bad thing either.

Yes, this is the argument the MAFIAA would use on humans who memorize song lyrics. The utility of the copy as it stands should matter. Anything that a human can reasonably be allowed to do should go for machines. A human can write a song's lyrics down verbatim except sometimes with some errors. Imagine if we got as far as AGI and we had to shield computers from ever seeing or hearing copyrighted material instead of policing the behavior.

PowerShell defaults are partly to blame on the Windows side. You can't double click a .ps1 file without editing security settings. But you can pipe an irm command into iex and run a random script from the web with no checks at all. Just a one-liner copied and pasted and you're giving complete control over.

I mean, sure. A lossy copy is still a copy. I was going to make an argument about a YouTube video shot outside where someone nearby is listening to a song over headphones where it's loud enough for a severely distorted copy make it faintly into the background but I'm pretty sure they issue takedown notices for that. These things don't get tested in court but it's not even up to the threshold of fair use because it's unintentional and not always even perceptible.

I guess what I'm trying to say is that there's probably no amount of loss to be lossy enough for them to leave alone.

We could possibly create the future of AI copyright violation using prebuilt and customizable models.

Imagine the damage of only using a GPT to find out the lyrics of a song rather than the hundreds of free ad-supported web sites that also have the lyrics posted and aren't getting constantly battled in court.

Not that I agree with the ruling, but the music mafia would likely also go after a human who happens to have memorized the lyrics.

OpenAI saying it's the user who caused the violation is equally wacky.

A proper ruling would argue that it's not a copy until the original lyrics are reproduced in its output. Technically, the encoded text made its way in some lossy form into the model, but it's not really there either. These are copyrights, not patents. Being inspired by ideas is not a copyright violation. Accidentally writing lyrics that already exist is a violation but are usually handled in individual copyright cases, not blanket ones.

Arguing that its output is a derivative work under copyright law would really only make sense if it was deriving from one or two works - not hundreds/thousands. I think we should hold the software to similar standards that we would a human. And if the law doesn't fit that, the law should get changed.