I’m certainly not saying that the people under attack should be obligated to prove it in any way.

Just curious if the AI is in any way accurate or if it ends up just being generic naked bodies with a recognizable face. I’m guessing it’s the latter.

Hey whatever it takes to convince people.

It’s disgusting yes but I don’t see it as any different than the magazine-style clipping cut-and-paste. More expensive and worse for the environment I suppose,…

But none of the people involved in creating it care if it’s real, just as before, and the realism isn’t even the point. So it’s functionally identical.

You don’t need to tell that it’s real vs AI. If the objective is to create a picture of the person naked all you do is go “yeah that’s not what I like like naked” and move on.

There is no way other than random chance that the AI would know about a random scar, mole, or a million other details it will get wrong. And I expect that it will have that plasticky airbrushed look, but AI is getting better about that so maybe not.

Could be.

I don’t doubt that they do really look like naked people, but i doubt they actually look like the person they are supposed to portray.

In which case… back to square one, you’re putting someone’s head on the body of a different naked person.

I doubt it.

Someone should do side-by-side AI slop fake nudes compared to the real thing. I bet they get it wrong as much as they do everything else, i.e. constantly.

Why is this any scarier than pasting a photo of their head on a clipping from a porn magazine?

What about the rest of them?

Aren’t they all linked to the Muskrat?

The logic is that the most useful app will be the most popular by virtue of its utility.

Not always true but I’m sure there’s some correlation.

It works like profiles in chrome now.

Is it gonna pop up obnoxiously every time you start the program?

Is it gonna demand that I create a new profile every time I sign in to Google?

Still gives money to Google though.

“The other side of the airtight hatchway”

Or do that and cite the decades of work and investment it takes to build that industry.

It seems only natural…

• the “grandma”/casual users never try anything complicated or different so nothing goes wrong.
• the “pro” users either know what they are doing well enough to not make a mistake or to fix it when it goes wrong.
• the middle users will always have it harder, they are trying things beyond the margins of “easy” so of course things go wrong and they don’t know how to fix it.

Anecdotal example: just yesterday I found out that I broke my file picker function in five out of six web browsers, by loading an Xcompose file with some definitions that GTK apparently doesn’t like. It took me about 5 hours of poking at things to figure out that a change I did a week ago, broke a function I hardly ever use. So I did fix it eventually but I it took me a week to notice and then hours to track down what was going on.

Is there any chance at all that the casual users would be using a compose key, let alone loading a custom definition file for it? Hell no!

But here’s the secret: there is nobody out there who is the perfect expert who never makes a mistake and knows all things. We’re all out here pushing boundaries; the only difference is where those boundaries are.

This applies outside of IT just as much, maybe more. It’s the rare person who will admit it though.

Nah, lots of train nerds doing train nerd stuff.

Surprisingly little Thomas fan-mods tho.

Railroader. Lots of heavily-modded Railroader.

Isn’t nfs pretty much completely insecure unless you turn on nfs4 with Kerberos? The fact that that is such a pain in the ass is what keeps me from it. It is fine for read-only though.

Because other people stop existing in that bubble, because they become part of the background, bubbled people stop caring about them.

See also !fuckcars@lemmy.world