What is happening USA. Please stop.

"It Hurt Itself in Its Confusion!"

I saw a comment in a "I moved from Windows to Linux" thread implying Windows has more configuration potential than Linux. I wonder what that commenter would make of Gentoo.

I wish I had more time I could dedicate to maintaining my system, I'm marooned on Arch due to lack of time, such a shame.

Wait Corey Quinn is on El Reg now? That's awesome

Vibe coding the tool to remove the vibes from text is kind of beautiful really

Move fast and break things and duct tape it if the center cannot hold.

ProtonVPN free tier is great

On size limited platforms like steam deck and friends this is a huge W

Remember this not a billion dollar company we are talking about here, this is volunteer OSS project. If no one wants to maintain the X11 support then that's up to them.

Its more the case that certain people decide X11 will no longer be supported and block anyone from trying to support it in "their" project.

Click the export tab

My issue is that the quality of the macos UI is degrading over time. They can't even get rounding consistent, not quite at windows levels of mismatching yet though.

Also no one bothers making the beautiful native apps now, everything is electron, which is equally inconsistent everywhere.

So I think the advantage over time Vs a Linux system is diminishing... Slowly.

That is pretty impressive.

So impressive it makes you wonder if someone has noticed it being used a benchmark prompt.

Simon says if he gets a suspiciously good result he'll just try a bunch of other absurd animal/vehicle combinations to see if they trained a special case: https://simonwillison.net/2025/Nov/13/training-for-pelicans-...

https://www.svgviewer.dev/s/TVk9pqGE giraffe in a ferrari

"Pelican on bicycle" is one special case, but the problem (and the interesting point) is that with LLMs, they are always generalising. If a lab focussed specially on pelicans on bicycles, they would as a by-product improve performance on, say, tigers on rollercoasters. This is new and counter-intuitive to most ML/AI people.

The gold standard for cheating on a benchmark is SFT and ignoring memorization. That's why the standard for quickly testing for benchmark contamination has always been to switch out specifics of the task.

Like replacing named concepts with nonsense words in reasoning benchmarks.

Yes. But "the gold standard" just means "the most natural, easy and dumb way".

I have tried combinations of hard to draw vehicle and animals (crocodile, frog, pterodactly, riding a hand glider, tricycle, skydiving), and it did a rather good job in every cases (compared to previous tests). Whatever they have done to improve on that point, they did it in a way that generalise.