You should take a look at the games support for Valve’s Proton project. Games in Steam now mostly work perfectly in Linux because of it.

My god, the accountants who eventually have to clean up this mess will make so goddamn much money.

Oh, but they provide hallucinations and bug-ridden, incomprehensible code faster than anyone! That’s progress, right? Right?

AI expected to create efficiencies

Couldn’t help but chuckle at that bullet point. We should know by now that AI creates lots of cognitive debt and unstable, vibe-coded software, as well as making people stupid. But one thing it absolutely doesn’t seem to do is make things more efficient long term.

To me this translates to, “America’s military is being led incompetently and we don’t want anyone to see the real damage being done to it.”

Jesus fucking Christ, this is the stupidest possible way to start WWIII.

No they didn’t. CloudFlare created a vibe-coded knockoff of Next.js that will be a nightmare to maintain and no one with half a brain would touch, so they could get a few headlines from credulous outlets like The Pulse.

In reality, this garbage will wither away quickly, because it’s a security and stability nightmare created by one guy who hasn’t read his own code and doesn’t actually know what’s in it.

Every time I have to use my work laptop (with Windows) for anything, it feels like a giant step back. Lately it’s even worse; it feels like that step is right into some dog shit.

This might legitimately be the year of the Linux desktop, not because Linux suddenly got better, but because Windows finally got unacceptably bad.

A very cynical part of me wants to know who their parents voted for.

One possibility is that with Israel focused on Iran and Gaza, other countries take it as an opportunity to attack Israel. The U.S. escalates in defense of Israel, other countries escalate and start including U.S. bases as targets (becoming defacto allies of Iran in the process), and it all spirals from there.

How about a $10 billion fine for OpenAI for every mistake? Make it hurt. Make them pull the plug on this travesty.

Thanks!

It goes beyond the problems introduced by the model router, though. I have to work with GPT 5.2 for my job (along with Claude, Gemini, and a few others), and we have enterprise API access to it. So when I select GPT 5.2 as the model to use, it’s spending tokens to actually use it.

And it’s pretty bad. It’s noticeably worse than the 4.x series. I find myself having to fix its mistakes far more often.

I’ve struggled to reason out an explanation, and model collapse really seems like a contender, especially if you follow information theory and why training these things is so hard.

As it happens, there’s a new talk about exactly this from George D. Montañez. You might find it interesting: https://youtu.be/ShusuVq32hc

Then why are newer versions of the major models performing so poorly? For instance, GPT 5.2 is definitely not an improvement over 4.5. What’s the root cause?

And that is why I no longer buy anything from them. I’m just embarrassed it took me as long as it did to realize what they were really doing.

I mean, we’re watching it happen. I don’t think it’s hypothetical anymore.

It’s already happening. GPT 5.2 is noticeably worse than previous versions.

It’s called model collapse.

Oh, the guy from Hermit Tech! He’s great, and his blog is hilarious and poignant in turns (though sometimes both at the same time).

You wanna know who really bags on LLMs? Actual AI developers. I work with some, and you’ve never heard someone shit all over this garbage like someone who works with neural networks for a living.

They’re also not providing a large language model, so they actually did have a path to profitability. It’s keeping LLMs updated and running that costs so much money that companies trying to do so are losing billions, and Midjourney doesn’t have that problem.

It’s just that their path to profitability was built on plagiarism on an astonishing scale. You’re spot on, they should have been utterly destroyed right at the start.