After sifting through millions of sampled sounds, including typical instruments from snare drums and synthesizers to more exotic fare like kopuz (a Turkish string lute) and yanggeum (a Korean dulcimer), it’s easy to see how the sample library Splice has become a cornerstone of modern music creation. At a time when almost all producers record on laptops, Splice functions as a cache of sounds that can be dragged and dropped into a recording.
Since its 2013 launch, the music tech company has amassed millions of users, and it’s not uncommon to discover a Splice sample in a song from Billboard Hot 100 stars like Justin Bieber, Bad Bunny and Travis Scott, among many others.
One consequence of the company’s catalog size is the difficulties it can create for musicians searching for the perfect sound. Peering through a pair of chic, angular eyeglasses, Splice CEO Kakul Srivastava acknowledges that getting started is one of the “hurdles” that may discourage creators from using the platform, one she hopes to overcome with its new artificial intelligence-powered product called Create.
Unlike what Srivastava describes as “push button” song-generation platforms — she doesn’t offer specifics, but Boomy or MusicLM come to mind — Create is designed to encourage users to discover and incorporate new sounds into their songs, she says. Users can pick a genre they’re interested in, and then its proprietary AI will stack up to five recorded loops to construct a new arrangement.
Splice uses a subscription model, starting at $12.99 each month. Srivastava says the company — whose staff comprises almost entirely musicians, even in marketing and engineering roles — had a surge in subscribers during the pandemic, and since 2021, has more than doubled its number of subscribers from 4 million to 8.1 million.
Srivastava, who took over Splice’s top job in spring 2022 — replacing founder Steve Martocci, now the company’s executive chairman and chief strategy officer — is a newcomer to the music business, but she has been building technology to aid creatives for decades, including high-ranking gigs at GitHub, photo-sharing app Flickr and Adobe. She says the biggest difference between her work with musicians and other types of creatives is that “musicians are weirdly undervalued in this industry.”
She likens music creation to A Tale of Two Cities, although her metaphor strays from the plot of the Charles Dickens novel. There is one city with “glittering lights and disco balls, with beautiful people in gorgeous clothing” — artists and music executives chasing fame and money.
Although she recognizes that there will always be those who aspire to become the “next Drake or Rihanna,” Srivastava says Splice also serves a second, far less glamorous subculture of creators. “It’s a place with dull, poor lighting; where creators work in the basements of their homes,” she says. “That is the place of creating music as its own joy. It’s purposeful in and of itself.”
You’ve been CEO for a little over a year. What is your proudest accomplishment?
When I took the role of CEO, I focused on bringing in some tech-industry sensibility to Splice without losing the music culture at its core. How do you build an organization that is releasing new things to customers every week? This idea was very different from how things were before. We are constantly listening to user feedback and using that to drive our next product. Since launching Create, I think we have added 60 new features to it. I’m really proud that we’ve been able to retain both things: the science and the art.
There are a lot of AI-based music tools today. What do you think makes Create stand out in this crowded market?
One of the things that we find disturbing about a lot of the work that is happening in generative AI is that it is “push button” creativity. You push a button and poof, a song comes out. For us, that’s eliminating the creative as opposed to putting tools into the creatives’ hands. Our tool has been built with that mindset at the core of it: It should make more people curious about creating as opposed to less. I think that’s a big difference in what we offer.
I do think there is a playfulness to the tool. I love watching some of our users posting videos of them using Create — because they’re playing, not just in the sense of playing an instrument, but it’s almost like a live performance or experience in action.
An AI executive said he sees the future of music creation as becoming fully democratized, more similar to how we can all take photos and videos and share them online in seconds. What’s your perspective?
I really hope that AI and other tools — like our sample marketplace — help make music creation more accessible to people because we don’t want to just support people who want to become famous. We want to also support people who also just make songs for the intrinsic value and joy of music creation. The healing power of music is undervalued.
You once called sampling “a hyperlink” to the past. How do you feel about the increasing number of songs that reference other songs, such as Jack Harlow’s “First Class” and Latto’s “Big Energy”?
I think it’s beautiful. I see sound as a conversation. Music creators are having such a fun conversation with history. They are taking a song or a sound that might have meant something to them when they were a kid or a period [in their lives] and transforming it into something new.
Recently, some employees talked about how they approached building a new sample pack based on the sounds of ’70s funk. They were talking about the artifacts and details of the samples that came from a certain tape machine that we actually got from a historical archive. I love that we care that much about making the sounds we offer as authentic as possible.
Create uses AI to form unique arrangements of human-made samples and loops, but would you ever be open to offering AI-generated samples to users?
There’s so much conversation about it in the market. I think we’re certainly listening to the conversation, we’re talking about it internally. We are going to be customer-led in our approach to generative AI. Right now, our customers tell us that they love the sounds we have. They love the sonic complexity of our sounds. They love that it’s human-created. If it feels like this is something that people are curious about, we’ll certainly explore it. Because our training data is so good, we could probably do a pretty good job of it, but it’s not a focus for us right now.
Do you collect data on what users like at any given time? If so, I imagine you must have a lot of insight into music trends on a granular level.
It’s super fascinating. We have a global footprint in terms of our user base and in the sounds we offer, too. It’s also cool to look at what sounds matter in which geographies. We are going to spend more time thinking about our data and letting it inform our governance strategy.
What is music’s role in the creator economy?
I think social media is so electric and energizing to many people because it captures a moment, an emotion — something very real very quickly and in an easily consumable way. Music is the emotional layer of pretty much all social media. It is essential to the creator economy, and I think social media companies recognize that. It’s why we have partnerships with many of them and why companies like TikTok are investing in music production tools as well.
