In past issues of this newsletter, I’ve mentioned “runtimes” as being important in the age of AI. I should explain what I mean. Luckily, an interesting example showed up that I think makes the concept very plain. It’s called Endless, from a company called Polyend, and it’s a guitar pedal. It’s not out yet, but Polyend has a good history of shipping complicated hardware. 

Here’s a video that explains what it does:

Most guitar pedals offer one effect—adding echoes, or distortion, or reverb—but this one is totally programmable. You can program it the old-fashioned way, in C++, or you can describe the kind of effect you want in plain language, and then their “Playground” uses AI to transform your prompt into software. Then it loads the new effect into the pedal via USB. If you want, you can draw on a little card that labels the knobs. 

So you might type in “add a tiny bit of reverb and distortion then chop up the sound and add echo to that,” and then you’ll be able to feed your guitar into that effect. This isn’t free! According to their FAQ:

If you want to use the Playground, it depends on the complexity of the effect and the number of iterations. A simple delay could cost $1.00 or $2.00, whereas a complex granular looper might cost up to $5.00. $20 of Tokens come with the purchase of a new Endless.

Seems high! But anyway. This is an interesting AI product—much better, I think, than an AI “life assistant” or most other things I see. Polyend builds digital synths, so they know how to assemble code to make effects. I don’t have any great insight into the product, but it’s pretty obvious that they’re taking prompts, turning them into synth specs, then compiling them into software that runs on the pedal—and the pedal is, of course, a little computer with some knobs on it. It’s actually an AI-powered effects runtime that compiles to software that runs on a pedal-shaped computer.

A “runtime” is is a nebulous tech term (see the Wikipedia page for a variety of definitions), but it refers to all the stuff a programmer takes for granted when they build software—as that Wikipedia page says, it’s a “software platform that provides an environment for executing code.” Microsoft’s .NET is a runtime, and Java has a runtime, too. The runtime provides libraries for accessing the internet, opening a window, parsing an email address, and so forth. 

“Playground” looks like a runtime to me, except instead of having to know that to use reverb you have to refer to the library com.java.ReverbFactory, you just type in “reverb” and the AI routes to all the libraries and methods behind the scenes.

Of course, there’s one very real risk: If they stop supporting it, or LLM companies stop supporting them, then the user is stuck with a pretty metal brick—while many old-school guitar pedals have been working fine for decades.

That said, I think this is progress: People taking what their business does best, and letting users orchestrate and “program” however they want. One of the biggest concerns people have about AI is that we’ll use it too much and our brains will rot, and we’ll become a bunch of bot-servient losers. But this way, you use AI a little, then you go play guitar, and you might actually get cooler as a result. As long as you actually play guitar, and don’t spend all your time dinking around and coding up the effects pedal.