Even better

I’m not so sure that the laypeople will, but I do expect a shift. Personally I’m still running Windows 10 next to Linux currently. Most of my time is still spent on Windows, because it’s generally a bit more stable and hassle free due to the Windows monopoly. Software is written for Windows, so sadly it’s usually just a better experience.

But so many things I read about Win 11 (and beyond) piss me off. It’s my computer, I don’t want them to decide things for me or farm my data. I’m mentally preparing for the transition to Linux-only. 90% of the software I use will work out of the box, and I think with some effort I can get like 8% of the rest to work. It’ll be a lot of effort, but Micro$oft has pushed so far that I’m really starting to consider.

Multiple friends and colleagues (all programmers) I spoke are feeling the same way. I think Linux may double in full-time desktop users in a few years of this goes on.

Maybe you should learn how to hold a normal discussion without attacking the other party. Cursing and telling the other party to fuck off just because they disagree with you will not convinced anyone of anything and will only make you look bad. It’s not like the person you’re responding to is advocating for school shootings or anything else immoral.

:3

Kinda makes sense though. I’d expect images where it’s actually labelled as “an Indian person” to actually over represent people wearing this kind of clothing. An image of an Indian person doing something mundane in more generic clothing is probably more often than not going to be labelled as “a person doing X” rather than “An Indian person doing X”. Not sure why these authors are so surprised by this

Haha, when I heard about it I was expecting as much. It’d be pretty impressive if it went smoothly with the amount of testing they seemed to have done. Still an interesting project, curious how it’ll evolve.

When I was a student I never understood how something like this could happen. “Just rewrite it” I thought, how hard can that be. But working in a corporate environment I now totally understand it. Everything you write will at some point become part of code that, to the fast majority of colleagues, will just be a black box that they’ve never touched and don’t intend to. Despite my urge to test and document everything, I’ve already more than once complained about my own code, only realizing later that I wrote it myself. Often you can still find out what it does, but the “why” gets lost and because of that people are afraid to change it.

I like how specifically this relates to my experience with the discount factor gamma in Reinforcement Learning. Like, pretty close to the exact numbers (though missing 0.99 and 0.999)

Ngl, it’d solve a lot of bugs

Bullshit. Even if AI were to fully replace is software developers (which I highly doubt), programming is still a very useful skill to learn just for the problem solving skills.

Haha thanks

I looked at these videos with very mixed emotions. On the one hand, I marveled at how far we’ve gotten. In a few years we went from generating sort of okay images in a very confined domain and essentially uncontrollable, to generating high resolution video that on first glance looks real.

But then the sadness struck me. I think we’re entering the post-truth era, where the truth is harder and harder to find because all the fake stuff looks so real. We can generate text, images, sound, and now also video of whatever we want in the blink of an eye. Combine this with the tendency of people to accept any “information” that fits their view, and the filter bubbles that already exist, and we can see that humanity will start living in separate bubbles. Every bubble will have their own truth, and even if someone proves that a video or image is fake, that information will probably not even reach them because the truth doesn’t generate enough clicks.

I want to stay optimistic, we’ve overcome so much stuff as a species, maybe we’ll right the ship at some point. But with all the shit that is already going on in the world, the last thing we need is the ability to fake videos like this in no time at all. At some point the separate filter bubbles will tear our stable western world as we knew it apart, and we’ll see shit like WW II again. The situation is already heating up.

Opera is also Chromium though, so that wouldn’t really change anything

I’ve done a lot of stuff with generative models called GANs, like StyleGAN (2?) which I believe these pictures to be from. My main focus is the “hair bubble effect”. This works best for people with longer hair, which is why I had 30% wrong in this test. Basically, by starting at images generated by these models for a long time, I started noticing that it is bad at creating the few loose hairs that stand out from the main pack. These plucks of hair often seem to go around some invisible “bubble” or weirdly flow together with the background. So my main point of focus is often the transition between hair and background, or just the hair in general, since that’s where it’s most likely to mess up. But the images picked here were also intentionally picked to be the most confusing according to the rest of the article, so it’s not that weird that these are hard to classify. Some of the real ones looked extremely AI to me, and it was only after the first false positive that I got a lot more careful with labeling some as “AI” than I normally would.

Example, the strands of hair here (though admittedly the effect is not very convincing here):

This is probably a difference between countries, but personally I love it here in the Netherlands. I go to the store after work multiple times a week and I have yet to encounter a queue or problem that stalled me longer than 1-2 minutes. Usually I can just directly walk to a self-checkout machine, check out my stuff, pay by holding my debit card (or phone) against the payment terminal, and be on my way. I like it way more than the old way of doing things, because I now have time to properly pack my bag and I don’t have to talk to anyone. It’s also way more space efficient. There’s even the option to take a scanner with you so you can scan while shopping, though I have yet to try that.

I’m very interested in machine learning, or AI as it’s usually referred to in hyped-up content, and have also done research in the topic of generative AI. It hurts me so much that such interesting and useful technology has become another empty hype word. Nobody is waiting for this kind of gimmicky shit. It does nothing but hurt the public’s view of this technology, similar to all the copyright issues and privacy concerns. Why must companies take things that can ultimately do something good in some cases and apply it to absolutely everything with no regards for the impact on their customers.

Most companies don’t need “AI” and would be much better of spending those resources on actually improving their products. Useful places for generative AI would be something like Photoshop incorporating it, or a video game with better AI or more natural NPC conversations. Not some gimmicky AI assistant in a car, or some annoying AI chat window bothering you when visiting a news site. Soon people will be completely done with all this gimmicky “AI” bullshit and the useful applications will constantly have to deal with the stigma created by these greedy mindless corporations forcing “AI” on everyone in places where it adds nothing.

Even if it wasn’t about the money, I’d personally still be very pissed if someone would claim my work as their own and there’s no law to prevent it. I’m not saying that today’s system is great, because it sucks in many ways, but some copyright system needs to be in place for the good of everyone.