Hey Everyone,
Stable Audio 2.0, an audio generation model for Stability AI, now (April, 2024) lets users upload their own audio samples that they can then transform using prompts and create AI-generated songs.
It’s been really painful to try to follow what’s going on at Stability AI in recent weeks. The company says that CEO Emad Mostaque has stepped down to "pursue decentralized AI." This is unlikely out of his own volition, as their main investors had asked for him to leave many months ago amid very questionable conduct, media reports, employee accounts, former-employee accounts, etc…
On the Podcast of Peter Diamandis, Moonshots, unfortunately Emad was not very transparent about what has actually been going on in his case:
Since stepping down, Stability AI's former CEO has been vocal about not liking the job. Agreeing with Elon Musk about the role and about how corrupt centralized AI leadership in the space is. Everything but, commenting on his own conduct and personal downfall. The way Forbes have portrayed him in a series of articles, is very damming for the credibility of Stability AI as a whole. So it’s rather impressive that they keep launching things with the talent that they do have remaining there.
Peer Recommendation of a Newsletter on Music:
Can’t Get Much Higher by
Stable Audio 2.0
Around three hours ago, on April 3rd, 2024 they announced Stable Audio 2.0.
Introducing Stable Audio 2.0 – a new model capable of producing high-quality, full tracks with coherent musical structure up to three minutes long at 44.1 kHz stereo from a single prompt.
Stable Audio 2.0 goes beyond text-to-audio to include audio-to-audio capabilities. Users can now upload audio samples and, through natural language prompts, transform these samples into a wide array of sounds.
This model was exclusively trained on a licensed dataset from the AudioSparx music library, honoring opt-out requests and ensuring fair compensation for creators.
Explore the model and start creating for free at stableaudio.com
You can read the full blogpost here: https://stability.ai/news/stable-audio-2-0
The latest version of Stability AI’s Stable Audio audio generator now lets users create three-minute-long songs. Some reporters and analysts were debating its quality (relative to Suno AI for instance).
Listen to the Demo
Stable 2.0 goes beyond text-to-audio to include audio-to-audio capabilities. Users can now upload audio samples and, through natural language prompts, transform these samples into a wide array of sounds.
Stable 2.0 was exclusively trained on a licensed dataset from the
@AudioSparx music library, honoring opt-out requests and ensuring fair compensation for creators.
This is what they said in the blog about the research behind this:
Research
The architecture of the Stable Audio 2.0 latent diffusion model is specifically designed to enable the generation of full tracks with coherent structures. To achieve this, we have adapted all components of the system for improved performance over long time scales. A new, highly compressed autoencoder compresses raw audio waveforms into much shorter representations. For the diffusion model, we employ a diffusion transformer (DiT), akin to that used in Stable Diffusion 3, in place of the previous U-Net, as it is more adept at manipulating data over long sequences. The combination of these two elements results in a model capable of recognizing and reproducing the large-scale structures that are essential for high-quality musical compositions.
Stable Radio
Stable Radio, a 24/7 live stream that features tracks exclusively generated by Stable Audio, is now streaming on the Stable Audio YouTube channel.
Explore the model and start creating for free on the Stable Audio website now.
When asked about his departure, Mostaque told New York Times reporter Kevin Roose that "being a CEO sucks." So does defrauding your cofounders or facing lawsuits I guess. Or wasting investor money with very publicized inept leadership skills. Suffice to say, I’ve read a great deal about Stability AI’s leadership in recent days.
Addendum on Stability AI’s Post Mortem
Emad Mostaque’s talks with Peter Diamandis on his podcast Moonshots, is some of the wildest hyping of Generative AI I’ve ever seen or heard. Given how little money Stability, Cohere or even Anthropic makes compared to their funding, or how the makers of the AI agent Devin thinks they are worth, there’s a pretty brutal disconnect obviously in the space. As someone who covers Venture Capital trends around AI and stock market valuations of startups, I find it pretty outrageous to say the least.
Silicon Valley's AI gold rush fervor (of 2023) clearly made some people pretty mad. As the dust starts to settle a little bit in 2024 we have to wonder about funding vs. revenue discrepancies. I mean Stability AI reportedly ran out of cash to pay its bills for rented cloudy GPUs.
According to Futurism (ironically once owned by Singularity University related to Diamandis), in June 2023, Mostaque's relationship with his funders took another blow when a Forbes report alleged that the founder had a history of exaggerating his qualifications. Forbes, Fortune and others have done Op-Eds on Emad in excruciating detail that makes Sam Altman look like a rockstar in comparison.
Formerly involved in Crypto, Emad departing told the media: "I look forward to moving onto the next problem to handle," the statement continued, "and hopefully move the needle."
I’m not sure Stable Audio 2.0 Model will move the needle. But it’s still pretty cool.
Copyright Free
The first version of Stable Audio was released in September 2023 and only offered up to 90 seconds for some paying users, which meant they could only make short sound clips to experiment with. Stable Audio 2.0 offers a full three-minute sound clip — the length of most radio-friendly songs.
All uploaded audio must be copyright-free of course.
"starting a company is like staring into the abyss and eating glass."
Yes we know Emad, thanks for letting us know.
Emad Mostaque resigned from his role as CEO of Stability AI on March 23, 2024. Read the official announcement here. In the meantime, Stability has launched many interesting products over the past few months.