AI Sings My Song
Henry Kim ·
A message came from an acquaintance in Korea. "I'm writing my diary with SUNO these days." At first, I didn't understand. Isn't SUNO an app for making songs? I opened the link. My acquaintance's day had been turned into a song, just as it was.
Three years ago, I first tried SUNO when it was released as open source on GitHub. It was much rougher than it is now, but at that moment, I thought, "The era of composers is ending." The music industry would be shaken. I thought composers would be relegated to assistants, and human-made songs would become increasingly rare. That prediction wasn't wrong. However, I was thinking within too narrow a framework.
Seeing my acquaintance's SUNO diary, it hit me. Under the existing formula, songs could only be made with a studio, a label, and a marketing budget. We were drawn to relatable lyrics, liked the songs our peers enjoyed, and shared in the enjoyment of our times. It seems the size of that group also grew with the development of media. However, it seems AI has now removed those intermediate barriers. Now, anyone can turn their own story into a song and distribute it. Now, my story is directly distributed through my own media and delivered to the people I want to reach. By moving beyond legacy methods, we can create countless personal uses, not just traditional music distribution.
Today, my second son was listening to a song by a K-pop artist he'd never heard before. I asked, "What's the difference between an unknown song by an unfamiliar artist and a song made by AI?" My son paused for a moment. He gave me a look as if to say, 'What does it matter?' He's right. In the future, there might be no difference. The pixel and sound wave data we see and hear on screens might no longer require knowing whether a real person created it or not. After all, I've never seen most celebrities in person throughout my life. Today, I feel like I could understand even if they all said, 'We are all virtual.'