Keep pulling the thread on Mikey Shulman.
Suno's AI models are designed to generate music by treating it as a raw sound wave sampled at 48,000 times per second and represented as a continuous float32 number.
Suno's founders intentionally avoided providing their AI models with explicit musical theory, such as the 12 tones of Western music, to prevent the models from being creatively constrained.
CEO Mikey Shulman argues that improving AI music generation is primarily a research and data problem, not a scaling problem, because music lacks the objective benchmarks that exist for LLMs.
On a typical day, 90% of Suno's users actively create music on the platform.
Songs created entirely on the Suno platform have achieved positions on official music charts.
Musicians using Suno to create their work have successfully secured record deals.
Suno has entered into a partnership with the record label Warner Music Group.
Mikey Shulman believes that a sustainable competitive moat in the AI industry cannot be based solely on model superiority, as large companies like Google can likely close any model quality gap through spending.
Suno was originally founded to be an audio analysis company using AI to understand sound, not to generate it.
Suno's initial go-to-market strategy of launching as a Discord bot was inspired by the success of Midjourney.
Suno's platform utilizes Large Language Models (LLMs) to generate song lyrics based on a user's initial prompt.
According to CEO Mikey Shulman, Suno's AI model currently excels at generating country and pop music, genres he characterizes as being more formulaic.