It sadly doesn’t get in to too much technical detail, but it’s still somewhat interesting to listen to:
Some of the broad strokes are about creating a system (in this case software/hardware) that drills down into the specifics of a certain musical problem or idea, as opposed to a more general thing. And then optimizing that.
The founder of the company did his postgrad at NYU with Juan Bello, which I guess is a dude in this field. They wrote a paper together on genre classification.
Some interesting stuff about the hole in the market, or I guess in musician stuff in general, where there isn’t very much on using ML or deep learning stuff in a musical way (e.g. not curating playlist for Spotify). And how this was their attempt at resolving some of that.
Sadly towards the end he talks about some of what they are working on going forward and it is some generic “reinventing the DAW” kind of pablum. That’s, by far, my least favorite aspect of their software is that it’s a “cohesive” sampler/tool thing, and I find that shit super useless/limiting. So will see. I also hope they have some new hardware coming out too.