Douglas Eck and the Magenta Project first came onto my radar by doing a podcast that was posted on the Web Audio API Weekly email, and I found his discussion about music and machine learning to be compelling and focused. I reached out to him, and things came together quickly: he invited me to Google for a visit with his team, and also got the planning into motion for this podcast. Once it cleared all the hurdles, we were on! We didn't have a lot of time, but it is clear that Douglas is experienced in presenting the Magenta team's vision "in the time allotted" - whatever that time might be. He certainly packed a lot of information into the small amount of time I was able to talk to him. Trying to understand how machine learning can work within a musical environment - as well as how it can draw musicians in, rather than pushing them away - is something I'd never considered, but it is clear that this is a big part of how the team is imagining their work. If you are interested in musical machine learning as a concept, you'll want to check out the Magenta project at its website, and also get some basic machine learning education wherever you can find it. Hearing about the successes that Google has had with "deep learning" and "reinforcement training" is pretty interesting, and it is exciting to watch this stuff develop from the ground floor. Want to get involved? The Magenta project is open-source, and is actively interacting with artists and art-tech folks as they are able. Start by reading their site, test-driving their tooling from Github and learning more along the way. Enjoy!