5. Interview: Stephen Philips is a software engineer with over two decades of experience in machine learning. He co-founded the popular music analytics site WeAreHunted, which attracted millions of users worldwide before it was acquired by Twitter in 2012. Stephen worked at Twitter in San Francisco in Product and Engineering, and Mergers & Acquisitions before returning to Australia to start Mawson— an AI venture studio that partners with technical entrepreneurs to build, validate, and scale companies at the frontier of deep learning. Splash is one of the latest AI music apps to come out of Philips' Popgun company.
How do you personally describe Splash and what it's capable of to someone not in the music industry or who hasn’t really dabbled in making music?
Splash is a new app that makes it easy for everyone to experience the joy of making cool music. Splash features a huge variety of sound packs that you play using a simple beat pad interface. It really is the easiest way to make music and it sounds so much better than anything else. It's a lot of fun to play.
We have been delighted with the response from users so far. People are having so much fun and are blown away by how great the music sounds. For us, helping people make good music for the very first time is a delightful experience. We are really proud to work on this and bring joy to so many people.
(Splash is 100% free and available on iOS and Android.)
What do you see for AI music creation and what comes next for it?
When musicians think about AI they often think about robots and the threat of AI replacing musicians. The reality is that current technology is nowhere near that sophisticated and won't be for a very, very long time. While the technology is still developing, AI can already be useful in ways that do not threaten anyone.
Hundreds of millions of people worldwide love listening to music every day. And as every musician knows, making music is just as much fun as listening to it. Yet millions of people still can't make music and miss out on one of the great joys of life. The problem is that making music is still too complicated. We believe AI is the answer to this problem. AI will make it easier for everyone to make music and open up a whole new product category for the music industry. I predict millions more people will learn to love making music and it will rival music listening as a super fun thing to do.
What comes next for AI in music-making? AI can already compose midi for a variety of instruments, it can master audio like a human, and even generate synthetic singing. Yet very little progress has been made on the really hard problems of sound design and determining musical quality. How do we use AI to make new sounds that no one has ever heard before yet still sound amazing? How do we train an AI to tell the difference between good, bad and amazing music? These problems are extremely difficult. We still have a long way to go.
What’s the monetization method for AI music creation?
People who love making music are always looking for ways to improve. They are willing to spend money to make better music. AI is simply another tool that can help people make music. We believe that musicians, from beginners to pros, will pay for AI products if they can help them make better music.
We have no interest in royalties or taking a share of someone's publishing or anything like that. We plan to make music-making experiences so good that people will pay to use them. We want the user to 100% own the music they make and use it in any way they like.
What’s one new music app and one new band you’re digging at the moment?
Bay Faction are my favorite band right now. Their sound really hits my indie jangly guitar sweet spot.
My most used app is Twitter which I think remains an incredible product. The world's most interesting people all opening their minds.
My favorite new music app is YouTube Red. Watching reaction videos to my favorite songs is my guilty pleasure. I am always amazed how awesome music transcends time and culture and shakes people to their core.