Sean Parker is steering Stability AI toward music

From image generation to music

The software can create full instrumental tracks or short clips from text prompts. Parker said a future update will let users guide generation by humming a melody or beatboxing a drum rhythm.

That move follows a financing deal built around access to recorded music: Sony, Warner and Universal are both investors and catalog licensors. The announcement pairs capital with material the company can use to build its audio models.

A different approach after the rescue

Two years ago, Parker took part in an $80 million effort to save Stability AI. The image-generation startup had been close to collapse after excessive spending and internal disagreements; founder Emad Mostaque left the company.

Parker acknowledged that the company’s previous strategy of “asking for forgiveness rather than permission” had not worked out well.

I think the music push is also a test of whether Stability can build a business around licensed material, rather than repeat the approach Parker now says failed. The announcement says little about how the company plans to turn its new models and editing software into a sustainable operation. Music catalogs may give Stability a more defensible starting point, but permission to train is not the same as proof that the product will find lasting demand.