Links, Feb 27, 2026

§Jimi Hendrix Was a Systems Engineer

From IEEE Spectrum, a wonderful piece which touches on a lot of my interests at once:

Hendrix’s mission was to reshape both the electric guitar’s envelope and its tone until it could feel like a human voice. He tackled the guitar’s constraints by augmenting it. His solution was essentially a modular analog signal chain driven not by knobs but by hands, feet, gain staging, and physical movement in a feedback field.

Coincidentally, the Fuzz Face is a very simple circuit, and if you want to get into DIY guitar pedals or music circuitry, you can learn a lot by building one and playing with its schematic.

I hadn’t heard of this online magazine before, and they also have a great piece on computer music pioneer Miller Puckette, creator of Max and PureData, whose “patching” paradigm of programming is something I still think is under-appreciated.

§Snakes.run: Rendering 100M pixels a second over SSH

Nolen Royalty, who did the One Million Checkboxes thing, is back with a thing you can check out from the terminal with ssh snakes.run, and again provides a wonderful writeup:

There were 3 core challenges I ran into building snakes.run.

  • Display: Making snake look nice in the terminal

  • Bandwidth: My early code used a shocking amount of bandwidth

  • Performance: Supporting thousands of concurrent players was hard

§Fix Your Hearts or Die

A meditation on toxic masculinity with some advice I wish I had taken to heart a lot sooner in life:

A man free of patriarchy is a man who has found not only every woman’s humanity, but one who has at last discovered his own.

§On Joy and Resistance

There is a reason why the administration has cracked down on the arts and humanities alongside its brutal assault on migrants. It knows that art is dangerous, that knowledge leads to asking questions, and that those questions don’t always lead where they want you to go.

§I verified my LinedIn Identity. Here’s What I Actually Handed Over.

A detailed piece on the terms-of-service behind online identity verification:

The whole thing took three minutes. Scan, selfie, done.

Understanding what I actually agreed to took me an entire weekend reading 34 pages of legal documents.

Garbage Day on generational media trends:

Culture right now is determined not by human teams of editors and producers picking and choosing what youth culture gets the spotlight, but, instead, by the unthinking algorithms that power YouTube and TikTok. Which means the only things that have the level of scarcity and danger required to be seen as cool by young people will, slowly, but surely, be whatever is unacceptable on those platforms.

§What is a Token

Jennifer Moore dispelling the illusion of LLM text generation with a cogent explanation of how the technology works, and touching on the cause of “LLM Burnout”:

Vigilance means the same thing in this context as in colloquial speech. It’s remaining alert for potential hazards. And it’s really, really hard. It’s why you feel exhausted after a day of driving. Humans are just bad at this. So bad, in fact, that early humans may have domesticated wolves because it was easier than guarding against them.

§Google API Keys Weren’t Secrets. But then Gemini Changed the Rules.

This was shared with the commentary “I bet someone got a promotion for this” and I said something very similar when google let my kid sign me in via the youtube app on my AppleTV. Does anyone at Google actually respect people not employed by the company?

You created a Maps key three years ago and embedded it in your website’s source code, exactly as Google instructed. Last month, a developer on your team enabled the Gemini API for an internal prototype. Your public Maps key is now a Gemini credential. Anyone who scrapes it can access your uploaded files, cached content, and rack up your AI bill.  Nobody told you.

AI will destroy the world, but in the most offensively boring way possible

I really wish I didn’t feel like I had to post about this, but as I continue to see people post things like “Claude can make binaries now! in a year we won’t even have programming languages!” and “Just keep your passwords in ChatGPT. Heck, in six months AI will just solve authentication” I kinda feel like I should. I’m selective with these – for every link I end up sharing about genAI stuff, there are typically three I don’t.

§How did we end up threatening our kids’ lives with AI

From Anil Dash,

It used to be that encouraging children to self-harm, or producing sexualized imagery of children, were universally agreed upon as being amongst the worst things one could do in society. These were among the rare truly non-partisan, unifying moral agreements that transcended all social and cultural barriers. And now, some of the world’s biggest and most powerful companies, led by a few of the wealthiest and most powerful men who have ever lived, are violating these rules, for profit, and not only is there little public uproar, it seems as if very few have even noticed.

How did we get here?

Anil follows up with Taking action against AI harms.

§Acting ethically in an imperfect world

If you missed it (which, I’m jealous), Cory Doctorow posted a screed decrying people who are against LLM usage as “purity culture” while defending his use of Ollama for local LLM spelling and grammar correction. The argument is full of holes, ultimately against a strawman, and I think it’s a perfect encapsulation of why I’m not a fan of his. This piece is a great rebuttal:

Cory shows his libertarian leanings here: If everything is somehow “free and open” then we have won. But “free and open” in this context usually means that “certain privileged groups have easy access to it and are not limited in what to do with it”. That’s one of the core problems with the whole “open Source” movement: That it reduces all struggle to if one can get their hands on the tools and has any restrictions to using them.