Dwedit

Comedians often make jokes.

There's also "the" vs "the". As in "the boy ate the apple". First 'the' ends with a schwa sound. Second 'the' ends with 'ee' sound.

Even Qwen 3.5 can explain this disclaimer correctly.

"This is not simply a common pattern on bare metal — it was also undefined behaviour in C++."

Not just X (em dash) but also Y.

Even with a small context size (4096 at 64KB per token), that's like 256MB for the context. It's more than just the context that's eating up VRAM.

It needs more VRAM than just the model weights. With 6GB of VRAM, I got 44/65 layers loaded into VRAM. Has anyone tested 8GB yet?

I tried it on my 6GB GPU and got 0.67 tokens per second. Need more than 6GB to run it well.

You just need to split apart "Wall is stop".

This site is really hostile to people who do any kind of javascript blocking.

It was made during the brief XHTML craze. (And it's also invalid XHTML)

So this is a JPEG decoder that runs on an Apple II? Monochrome only, or can it do color?

I wish the browser would either allow local-only or internet-only, then only make up its mind after the first attempted request. Fetch an image from the internet? No script access for local files. Fetch a JS from elsewhere on the hard drive? No internet access. That could provide the sandbox while still being flexible.

Not to be confused with the actual radio station in Arizona called KOOL-FM.

You'd think so, but web browsers are downright lobotomizing the ability for local HTML files to run any meaningful javascript code. Want to run a JS file from the same directory? CORS ERROR!!!!

Hence 200MB applications that are complete copies of the Chromium browser to run under 1MB of actual web code.

Many of the projects here predate 2022, so they wouldn't be "ai-generated" in that sense.

Much like WEBP, JPEG-XL is two different codecs in one. There's VARDCT mode, and Modular Mode. Modular Mode is normally used for lossless, but it can also be used for lossy as well. When you use Lossy Modular Mode, you get different kinds of compression artifacts than what you'd see in a JPEG file.

For Lossless: JPEG-XL wins, then WEBP, then PNG. (Sometimes lossless WEBP wins)

For Lossy: AVIF wins, then JPEG-XL, then WEBP, then JPEG.

I looked it up. The one place where UTF-16 validation takes place is when using WideCharToMultiByte to convert to UTF-8 text. Before Vista, that was not validated.

Even just "-" in a filename is problematic, since you can name actual files with "--" at the beginning, and the name could match a switch.

Meanwhile, Windows uses UTF-16 everywhere internally, so "a" and "あ" are both 2 bytes large. You'd still have to exclude surrogate pairs and the Unicode control characters from a filename.

I use Google maybe 10 times a month.

Honey says Hi.

Any name that you can't type is going to be a problem.

Generated SEO slop is the misuse of AI. Very unlikely to find any guardrails to stop that kind of thing.

How about obsolete cables? IDE, RG59 coax, Cat 5, Composite video, etc...

Testing out Tongue, it detected "馬鹿外人" as Chinese.

I was one of the people who added sound into a TI83 game. It ended up in my IceClimb game. Not particularly well done or anything, it's just there. I also didn't properly adjust the timing as the notes got higher, so it plays faster when you get higher in the level.

Turns out that it's actually 2-bit and not 1-bit. It uses a combination of the EAR and MIC jacks, and when you output through the MIC jack, it slightly boosts the EAR jack. That gives you 4 possible levels, but they are not evenly balanced.

How much data does it download off the server to run the challenge? I've made some tiny webassembly binaries before, but I've never seen Rust generate something tiny.

Nah, it's not piracy, it's AI model training.

What does "Browser's Main Thread" even mean when there are over 30 processes running?