"Naturally, I anticipated controversy around this device": ChatDSP creates instruments and effects in Max for Live using text prompts, yet its developer now wonders if it should be removed.] - Guitar | Technique

Developer Dillon Bastan is responsible for some of the most creatively inspiring Max for Live devices we’ve tested in recent years: tools like Songbird, Coalescence and Natural Selection have demonstrated a vivid imagination that’s singled Bastan out as an innovative creator within the M4L community.

Bastan’s latest creation is ChatDSP, a M4L device that makes it possible to create instruments, audio effects and MIDI effects in Max for Live from text prompts. It’s really that simple: just open up ChatDSP, type in what you’re after in the Generate tab, and the device will appear in the Device tab, with up to 48 parameters available for each device.

ChatDSP comes in three different iterations: there’s a separate device for generating synths, audio effects and MIDI effects, and each one has a default template with some basic parameters and functions available. Once you’ve generated a device, you can make changes to it using additional prompts until you’ve landed on the instrument or effect that you’re after; the device can then be saved as a preset for later use.

Article continues below

ChatDSP – Device Walkthrough – YouTube

Watch On

ChatDSP works by forwarding your prompt to an external AI agent, and the user is able to choose between various models from both Anthropic and OpenAI. You’ll need to set up an account with one of these platforms in order to use ChatDSP, and each prompt requires a certain number of paid-for credits, but Bastan doesn’t get a cut.

natural selection

Play God with presets: Dillon Bastan makes sound design evolutionary with this innovative Max for Live device (Image credit: Dillon Bastan)

Bastan says “the sky is the limit” when it comes to what ChatDSP can create, and examples provided run the gamut from the basic (“a super lush reverb”) to the tongue-twistingly ambitious: “a vocaloid-style formant synthesizer with glottal pulse source 5-vowel morphing breathiness vibrato portamento formant shift nasality and per-note random vowel offset”.

The release of ChatDSP has stirred up some controversy among Bastan’s followers, provoking heated debate between those who view the device as a laudably creative use of AI that has the potential to empower music-makers without programming skills to design their own tools, and others that question the ethics of generative AI.

In a message shared on Instagram, Bastan responded to critics and said that he is open to taking the device down. “Of course I knew this device would be controversial and I understand the feedback about artists not wanting to see something like this in the world,” he says.

Of course I knew this device would be controversial

“I made it as another fun idea and a thing for people to play around with their creativity and experimentation, but I do hear many of you that are disappointed or upset by it, and those that are excited to use it. I am open to taking the device down, it’s not like I’m going seriously in an AI direction. Let me know your preferences and I can respect them as I have appreciated your support a lot over the years.”

Bastan explained further in a YouTube comment that he doesn’t view tools like ChatDSP as a significant threat to creators or developers, but that he empathizes with those who question the environmental and social impact of AI giants like OpenAI and Anthropic.

“Ironically this idea would affect myself/others like me the most, as I am a digital instrument maker,” he says. “IMO it doesn’t do as good as we do (yet at least) so if anything to me it’s saying, ‘yes have fun making the random ideas you have’, but it’s not cutting it compared to getting devices from people really crafting a wonderful audio device.