newsfilter.io
Interview

Mikey Shulman, CEO @Suno: The Future of Music, What is Gonna Happen? | E1244

  • Suno's core mission is to transition music consumption from passive background listening to active creation, aiming to make music as interactive and engaging as video games.
  • The company pivoted from an enterprise "sense-making" audio tool to a consumer generative music product because generative capabilities emerged faster than anticipated, rendering the initial product obsolete.
  • Unlike text-based AI where scaling laws are the primary driver, Suno argues that audio requires a strategy of "many small models" rather than massive scaling, as music involves subjective "taste" rather than objective benchmarks.
  • Suno utilizes a transformer architecture but differentiates itself through proprietary innovations in audio tokenization rather than novel model architecture.
  • The company has adopted a "charge from day one" pricing model with annual subscriptions (starting at $300) to validate product-market fit early and identify high-value user segments.
  • GPU compute costs represent the single largest expense for Suno, running "a few times payroll," and are expected to remain high as research requires high-quality experiments rather than just scale.
  • A lawsuit filed by the RIAA in June 2024 challenges Suno's use of copyrighted works in training data, a practice the CEO describes as industry standard but notes is currently a source of litigation friction.
  • The CEO expresses a strong preference for industry collaboration and "growing the pie" over the "disrupt at all costs" Silicon Valley mentality, though he acknowledges the legal risks involved.
  • Suno has secured partnerships with industry figures like Timbaland, who serves as an advisor in exchange for equity and early product access, rather than receiving cash payments.
  • Suno has rejected API requests from GenAI video companies to prevent Suno audio from becoming low-value background music, aiming instead to make music a central, valuable experience.
  • The company realized too late that relying on Discord for product distribution was a strategic error; migrating to a custom web app shifted 90% of traffic within five days, highlighting the importance of dedicated UI over chat interfaces.
  • Suno currently operates with a raised capital total of just under $125 million, using funding to deploy rapid product development and secure GPU resources.
  • The CEO predicts that future AI music interfaces will move away from "prompting" entirely, favoring more intuitive interaction methods where success is measured by output quality rather than input complexity.
  • A "good" future for AI music is defined by infinite supply enabling billions of people to participate in creation, creating a social network where users can remix and personalize content, similar to the evolution of photography on Instagram.
  • Two "dystopian" future scenarios are identified: one where artists are impersonated without consent (e.g., fake Drake/Ariana Grande songs), and another where hyper-personalized algorithms create an antisocial, isolated listening loop.
  • Suno is exploring business models that allow artists to opt-in to personalized models, enabling superfans to generate "fan fiction" style music while compensating the original artist.
  • The CEO believes that as AI lowers skill barriers, "taste" and curation will become the primary differentiator for creators, shifting the industry from technical virtuosity to editorial judgment.
  • Suno is facing a "tragedy of the commons" regarding quantum computing, arguing that venture capital is too short-sighted for the decade-long physics challenges involved, and suggesting government funding is necessary.
  • To compete for talent against giants like OpenAI, Suno leverages its unique focus on aligning AI with subjective human taste and its headquarters in Cambridge, Massachusetts, rather than competing on salary alone.
  • The company has observed unexpected use cases, including using AI music for cold sales outreach and marketing, which drives high willingness to pay among users.