Bite Code!

Bite Code!

@bitecode.dev

Been using Python since 2.4 and still can't shut up about it. Have a fantastic blog on bitecode.dev. Yeah, it's about Python.

1.5K followers231 following1.3K posts

Bite Code! reposted

I once inherited a python system that handled about $100m in ARR. It was a single Django class that was 25k lines long and had zero tests or docs that always called out to MySQL and postgres via fork+exec rather than actually using those client libraries

This post is unavailable.

This is really nasty. People talking about being listened to by their phone because of targeted apps have been gaslighted for years. But we know Siri and Alexa both got caught on listening scandal, and now this. At this stage, it is naive to think it's not happening on a regular basis.

Bite Code! reposted

Yesterday, the CSSWG accepted my proposal for mesh gradients and diffusion curve gradients! Want to make gradients like this natively in CSS? Well, we've started the work on it. It's coming! Here's the Explainer: github.com/w3c/csswg-dr...

Another grave and dangerous incident in AI's activity: I made it code a YouTube client, and it uses "never give you up" as test data. Claude Rick-rolled me :D

Catastrophic failures are rare with frontier models now. People will get used to this and take fewer precautions. But I'd argue that, as these systems become ubiquitous and more powerful, we should actually take more. This morning, Claude happily deleted all the SSH keys on one machine.

View full thread

I wiped an entire production server with a PHP 3 script at my first job. My mentor was really nice about it. The price of learning. And I learned. But consider that these AI systems can perform a thousand actions per second, with an error rate of just 0.2%. That's two fuck-ups per second.

They make fewer mistakes than humans now, and most of those mistakes are benign. But they make them so much faster, and they're becoming involved in so many things, that guardrails and contingency plans are more important than ever.

Catastrophic failures are rare with frontier models now. People will get used to this and take fewer precautions. But I'd argue that, as these systems become ubiquitous and more powerful, we should actually take more. This morning, Claude happily deleted all the SSH keys on one machine.

Of course, it was a machine dedicated to Claude, I had a full backup, and the keys were passphrase-protected, so it wasn't a big deal. I've had interns do way worse. I've done way worse myself.

Remember the FOSS AI clone of the Adobe suite? They added a clone of World, Excel, Powerpoint, Audocad and Avid: https://getartcraft.com/apps And yes, they also run on Windows, Mac and Linux offline. And yes, they are faster and free. Crazy times.

getartcraft.comCrafting Apps: open-source creative toolsImage editing, vector illustration, video, photography, PDFs, motion graphics and page layout: seven native, open-source apps from the ArtCraft team, built in Rust and free to use.

Bite Code! reposted

Just found out that a lot of people don‘t know that you can use a ”C7 figure‑8 power cord“ with any Apple power adapter in case you forgot your travel adapter. Way cheaper to get one of those usually than a full travel adapter. #travelhacks

Bite Code! reposted

Ah, among all the spam I receive for the blog, I just got my new favorite: it invites me to be part of the scam! #thuglife, am I right?

One of my oldest python pet peeves, fixes.

Python Hub@pythonhub.dev

Soft-deprecating re.match() The article explains when to use re.prefixmatch(), re.search(), and re.fullmatch(), what soft deprecation means for existing code, and how to flag re.match() usage with Ruff. https://hugovk.dev/blog/2026/soft-deprecating-re.match/

Making architectural mistakes is so cheap to fix nowadays; I'm amazed. This is the next step of agile: the short feedback loop continues after putting stuff in prod, because you can so quickly see when you didn't account for something and fix it.

View full thread

Situations that were scary just 2 years ago because it would take months to fix are now just a blip. Data migration, schema changes, cache settings, offloading work, it's just one wish away. Of course, the downside of this is that it's tempting to just try something in prod and see how it goes.

It feels like the 2000' again where I used to see people send PHP spaghetti through FTP directly to prod as root. It did bring innovation at a crazy pace. And some disasters. But boy was it fun.

Making architectural mistakes is so cheap to fix nowadays; I'm amazed. This is the next step of agile: the short feedback loop continues after putting stuff in prod, because you can so quickly see when you didn't account for something and fix it.

Refactoring, changing the whole structure of your project and entire workflows, is so incredibly easy with AI. It can do it quickly, and with the confidence of a gigantic test suite with 100% coverage, env sim, fake data, etc.

You can't make that stuff up, they literally call their artificial SEO spam links "Natural links". I think the real bitter lesson of 2020 is that if you are doing something bad, you should call it the absolute opposite of what you are doing, with zero shame.

Unrelated, but I have a lot of calls to "unsubscribe" in emails that are not in the subscription format. I assume it's to pass spam filters, or maybe it's an artifact of LLM writing more and more emails and thinking it's what you need to do. It can give really funky vibes...

(I'm contacted for all kinds of bonkers things, so I'll add it to the pile) I'm keeping the whole thing in clear because I don't think I'm actually doxxing a real person; the chances that a human wrote this are very low.

You can't make that stuff up, they literally call their artificial SEO spam links "Natural links". I think the real bitter lesson of 2020 is that if you are doing something bad, you should call it the absolute opposite of what you are doing, with zero shame.