“nvidia’s confidential compute” had me choke when reading it. Sure bro, sure.

Unless everyone you communicate with have agreed to use the same standard as you, no, it is not.

Eh. I’m mostly a power user, all day at work in terminals and keyboard shortcut galore.

It doesn’t prevent me from laying back and running a “filthy casual” kubuntu with little to no setup at all. At one point you reach the state where you just want to use your computer, not tinker with it all the time.

I don’t know how it works in the US, and it’s certainly not perfect, but if you bring “compromised” evidences of actual child abuse, you can be sure it’ll trigger some extensive investigation. Sometimes, this leads to nothing. Sometimes, the “evidences” were not really evidence, but more of a hunch. Sometimes, it leads to actual consequences.

One issue with such a vigilante system is that sometimes, you’re wrong. That’s why due process is a thing. But the “this evidence was collected illegally, so we won’t even look up the guy” thing isn’t really a thing, as long as the evidence is better than “I’m sure of it”.

Great. Just dump anyone you dislike as a “pedophile” and then it’s a free for all.

It will obviously look different, but with binfmt, wine, and a sane initial setup, you could get a lot of .exe works from a click in the UI (or the CLI, after all, CLI apps exists on windows too).

You could improve your reading comprehension.

but ignore the long history of vulnerabilities, bugs, and cursed workarounds present in X.Org

You’re not wrong on the other points, but that one… you’d also have to ignore the things that got fixed in X.Org, and the things that will show up in the various wayland implementations that were fixed previously. That’s the thing when doing things from scratch, old issues shows up sometimes.

Those figures are larger than the total storage usage on my work computer, with every tools installed and repositories cloned locally. I know that large storage are way more accessible, but it still sounds crazy to take so much space.

The only way I can go over that is by installing npm dependencies in every source tree, which is also a thing that really should be improved.

Until it doesn’t work. There’s a lot of subtlety, and at some point you’ll have to match what the OS provide. Even containers are not “run absolutely anywhere” but “run mostly anywhere”.

That doesn’t change the point, of course; software that are dependent on the actual kernel/low level library to provide something will be hard to get working in unexpected situations anyway, but the “silver bullet” argument irks me.

No, they hate flatpak, one of the many option to distribute software, which is not the only one even if you consider the “must run on many distro” restriction (which isn’t 100% true, kinda like the Java write once run anywhere). There are other options, some more involved, some simpler, to do so.

They didn’t say they hate devs, that’s on you, grabbing a febble occasion to tell someone that voiced his opinion to “fuck off”.

Yes. It’s flooding places, and suddenly people decided that “smooth looking” was the absolute end goal of any drawing/music/creation/etc. It’s not. Some of the most famous art piece are completely wrong, some aren’t. That’s not the endgoal. Nobody’s gonna care that you can take that very simplified drawing and “generate” an extremely high-detail, fully shaded image that looks like it, as it was never the purpose.

Creative direction, intent, consistency (or absolute lack of consistency), execution, style, and a lot more goes into any creation, art or not. That’s what make a piece feel interesting. There’s a reason even now, with generated content being plausible as far as glaring mistakes go, we can still point out which image “feels” AI across a lot of different styles. At best, to remove that feeling of it being wrong, you’d have to spent a lot of time on the output of a model to touch it up everywhere and change details, which requires time and proficiency, which a lot of people jumping on that trend definitely lacks. Some of the worst results I’ve seen have been from people trying to make other “pay” for their output.

There’s also the issue of how these works. For decades, creative people (among other) have been sued by big companies, some very harshly, to protect IP from such overexploitation as “using a three second excerpt in a video” or “using the vague likeness of a character”. And now, these same targets are getting fleeced of their work by more big companies under the cheer of the people. That’s a gut feeling of disgust right there. Combined with the utter lack of creativity in these, we’re really watching the potential death of an activity (artistic creation), and that’s not a good place to be. If one wants to argue that “generated art” is also a form of creation, keep in mind that these models can’t be trained on generated pieces without extreme prejudice. Killing the very source they need to operate does not seem like a good long-term plan. But who cares about long-term when you can make a quick buck, right?

I’d also like to point out that all this rambling is about generated content that goes from “output of a model” to “final piece” with little to no afterthought. The “common” piece, where people will be happy to see twenty broken pieces because “well, there’s a lot of them, so it’s good”. AI and LLM models, as a tool, may or may not be useful in the long term, but I can see smaller applications, even for art. A lot of menial tasks can be improved, general posing, references, simple background that are marginally considered part of the product, guides, etc. Taking something you’ve drawn/created, and locally use an AI “filter” to remove an extra line cleanly or touch up a mistake you want out? Great. The tool carries the intent of the artist, the same way a pen do.

But AI generated content? Make a prompt, a stick-figure sketch, and call it a day? These, IMO, will always look and taste like garbage, no matter how pretty they look. Because it was never “pretty” we were looking for.

I don’t know how people can be so easily taken in by a system that has been proven to be wrong about so many things

Ahem. Weren’t there an election recently, in some big country, with uncanny similitude with that?

I honestly believe their self driving tech is safer than humans.

That’s how it should be. Unfortunately, one of the main decision maker on tesla’s self driving software is doing their best to make it perform worse and worse every time it gets an update.

I read this as “we don’t want you, the user, to interact with our 100% user-content driven website that depends on your presence to keep having value”.

Let’s hope they won’t cheap out on those bolts. Thankfully cheaping out on everything is not an habit they have, right?

He does a lot of things, in particular layer positioning/whatever this is called. I can’t really compare with PS though, since I don’t have it, but to open and do basic stuff on complex psd files that other software do not handle well, it’s ok.

No idea how large you can get with it though.

I don’t like the approach of piling more things on top of even more things to achieve the same goal as the base, frankly speaking. A “local” kubernetes cluster serve no purpose other than incredible complexity for little to no gain over a mere docker-compose. And a small cluster would work equally well with docker swarm.

A service, even made of multiple parts, should always be described that way. It’s easy to move “up” the stack of complexity, if you so desire. Having “have a k8s cluster with helm” working as the base requirement sounds insane to me.

Just checked the part about self-hosting. While it’s probably possible to handle things with a less heavy approach, their only “easy to use” example right now is to have a full-blown kubernetes cluster at hand or run locally in the source directory. That’s a bit much.

Locked-in platform closing the door. How surprising.

Accepting DRM in the first place is the problem. Hard to avoid, but still. I just got a boox; great value, can’t use adobe DRM. Didn’t have any problem there. Of course, money is going everywhere except big “publishers”, but that’s hardly an issue; they choose their business model, I choose my customer model.