We have reached another AI tool inflection point: guitar youtubers and other musicians are using Reverse Engineer Anythying (opens in a new tab) to clone plugins, and hacking firmware in their guitar pedals to let them do what they want. Without any prior knowledge of software development.
This might mean that people will buy less music hardware as they can make missing features themselves, instead of buying the new and improved version. It might also mean that Jevons’ paradox applies to some manufacturers, but not others…
Note on the use of AI: No AI was used in the writing of the text or the creation of the unicorn. I commissioned the unicorn from an artist who prefers to remain anonymous. It was made using a spreadsheet.
It is not all unicorns and rainbows. It is exciting, and terrifying. We can give old hardware a new lease of life and prevent e-waste. We can burn tokens in data-centers to do it, or do it more efficiently and locally with open-weight models. Or use other peoples' solutions and burn nothing. We can also buy much more hardware now there are more interesting experiences to be used.
I make music as a hobby, mostly with electronics, very slowly with classical guitar. In the past weeks my youtube feed has blown up with musicians hacking firmware (the code that runs inside synthesizers, drum computers, groove boxes and effects pedals. ) and letting devices run on the web, or replacing the firmware you have with something that offers a different user experience and sound. Below are a couple and some musings on their impacts. Buckle up!
Guitar youtuber without programming knowledge reverse engineers a pedal. Almost.
I know Rhett Shull as a guitar and music industry youtuber. He recently put out a video in which he clones an audio effects plugin (opens in a new tab) and gets quite far in reverse engineering an expensive guitar effects pedal using Claude Code. As in the CAD video from earlier this week, he mentions being both excited and terrified. Old synthesizers, drum computers and effects pedals get a new lease of life.
This also means, and I think Rhett mentions this, that hardware platforms that had entry level, premium and pro products built around the same internal hardware, it is now easier to buy the entry level product and make it premium. There was already alternative software available for e.g. digital cameras. Doing this becomes more accessible.
As he mentions in this video, he has no software development experience what so ever. For the effects pedal he found someone elses’ project while doing his own attempt. I don’t doubt he would have succeeded.
Last november we already had modular synth youtubers with little dev experience creating music plugins using claude code. This is somewhat less surprising when you consider software modular synthesizers as a kind of visual Domain Specific Language. They were already programming, just in a different language.
Groovebox gets hacked, many firmwares sprout, company open sources development kit
I follow quite a few synthesizer youtubers. I dabble with electronic music in my spare time. Since a month or so, the ’M Vave FM-1’ groovebox is all over my timeline. It is a $69 groovebox, with the original firmware built around the open source software synthesizer Dexed. (opens in a new tab) Dexed itself is a clone of a very popular 1980s synthesizer by Yamaha.
Floyd Steinberg explains the why, shows the explosion of firmwares (opens in a new tab) and explains how to install alternative firmware on the FM-1. First it was a hack, then the company started supporting open source. Now Akai, Behringer and others are following suit. All in the space of a month or so.
A complaint I read about the many different firmwares for the FM-1 are that “they are just UX”. UX in music is very important. It can be the difference between finishing a track in an hour, or wanting to throw your musical device out of the window.
It is the end of planned obsolescence in synthesizers and digital effects. It is unclear what the business models are for hardware manufacturers and plug-in developers going forwards.
There are now also emulators, software that looks and sounds like hardware, of most popular synthesizers from the 1990s and early 2000s. That happened in only a few months. Most of these synthesizers and drum computers used the same processors on the inside. Then it is a matter of trial and error to get it to work, after re-creating the UI. Coding harnesses are great at performing informed trial and error when they have a goal.
On the one hand, new versions of the same hardware with minor improvements, mostly in the software, now become a hard sell. Many of the more expensive new hardware synthesizers are a raspberry pi with a case, some knobs and keys and a display. On the other hand…
Jevons paradox: people now buy multiple $69 grooveboxes to run all the software
You can run the FM-1 with the UX of your choice on your phone or tablet (opens in a new tab) or browse firmwares and play music in your browser (opens in a new tab). Easy as Pi. All open source. I like Choral Root (opens in a new tab) , which lets you experiment with different combinations of chords, and GIO (opens in a new tab), which let’s you play with musical patterns in the style of Giorgio Moroder. All free.
I saw several reports of youtubers and commenters buying multiple FM-1’s, so they could play with different experiences at the same time.
Established manufacturers like Behringer and Akai are now following suit, at least announcing devices with open source firmware. Releasing them, we will have to wait and see. But this all happened in the space of a month or so.
What do you think? Less hardware by fewer manufacturers? Or an explosion of new manufacturers, new synthesizers, effect pedals and drum computers? People experimenting with software because the barrier is now lower, and then building hardware versions? Something else?
Further reading and viewing
Jevons Paradox (Wikipedia) (opens in a new tab) "Greater efficiency reduces the amount of the resource needed per application, lowering its effective cost; if demand is sufficiently price elastic, this induces demand, frequently resulting in a net increase of total resource consumption". Simon Wardley was the first to make me aware of this, in relation to conversational programming as he calls it. We seem to be getting an explosion in the amount of code, and maybe people buying more of a particular kind of music hardware. Thinking of the evolution of hardware, we now need much less energy and hardware spend to make electronic music than in the past, even with the rise in memory and CPU prices caused by the AI data centre boom.
On monday I wrote about Printing 3D objects from photos with a couple of prompts. People like me can now do with small plastic objects what the star trek replicator can do. A mechanical engineer, Michael Lydic, shows how to replicate a fairly complicated hook, armed with a simple camera, a micrometer, ChatGPT and a 3D printer.