This one is a fun post. There is a term, “earworm” used to describe when a song gets stuck in your head. And in a world where very quickly the tagline is becoming “there’s an AI for that,” this applies equally to song creation.
To set the scene for where and when this conversation came up, I was having dinner with some friends when I was asked if I had played with “Suno” yet.
“Never heard of it.” I answered. And that friend proceeded to tell me about an app where you put in a plain text prompt and it spits out a song about whatever topic you have given it.
“Yeah, but how good is it?” I challenged.
“Pretty good,” my friend told me whipping out his phone.
“I’ll believe that if it can turn waking up at two am to feed a foster kitten into a song.” I challenged.
And apparently this is an instance of challenge accepted.
The resulting song, actually serves as a bit of an ear worm. Listen for yourself.
So what is the story behind this Suno app and what are some potential use cases?
Suno was founded in 2022 by a team of four individuals: Michael Shulman (CEO), Georg Kucsko, Martin Camacho, and Keenan Freyberg. All four founders previously worked together at Kensho, an AI startup. The word "Kensho" is term in Zen Buddhism that describes an experience of enlightenment where one is able to “see one’s true nature.” The startup focused on delivering AI solutions to financial and business sectors and was acquired by S&P Global in 2018. That technology now founds the basis of an AI innovation hub within the company.
Suno is based in Cambridge, Massachusetts with a team that is a blend of musicians and artificial intelligence experts. Their stated mission is to break down barriers to music creation, enabling anyone to make music regardless of their instrumental skills or musical background.
The initial foray into generative audio was in April 2023 when Suno released "Bark," an open-source text-to-speech and audio model, on GitHub and Hugging Face. This foundational work in audio generation paved the way for their subsequent music creation tool which became more widely available in December of 2023 through the app launch.
The company operates on a freemium model as of the time of this post. Application users can use a limited number of daily credits. Looking into my app, I have 50 credits as of today which translates to about ten songs. You can give as little or as much detail in the prompt as you like. One of my favorite results was when I asked Suno to make an upbeat workout song for me with the text of "pumping like Arnold." This song came complete with album artwork featuring a lion and a barbell. How exactly the lion is managing to lift the barbell is a but confusing but if you want to listen to the result, use the player below.
There are of course upgrade options to paid subscriptions. These start at $10 a month with yearly options also available. When you create a song, you have the option to select which version you want to use. This starts at Vintage, version V2 and goes up to a beta version 4.5 available for Pro users with a song time generation max at 8 minutes. The default for free users is version 3.5 which has a time max for song length at 4 minutes.
Like all AI, the limitation here is the extent of your creativity. Most casual use cases involve creating songs for fun or special events, like one of my friends who used Suno to write an anniversary song for her and her husband. Like a seed you plant to sprout, Suno could also be used to inspire ideas for those trying to brainstorm or play with different genres and music styles. There is also a business use case in the creation of background music for videos, jingles, or similar audio elements. Probably one of my favorite applications I stumbled across in researching this was the idea that Suno could be used for educational purposes, putting songs to concepts to help with the retention of new information. Think of the classic “Glucose, Glucose” song that so many of us used to learn the krebs cycle.