- Turing Post
- Posts
- Stability AI Chronicle: Why It Can Fail in 2024
Stability AI Chronicle: Why It Can Fail in 2024
Investigating the Thin Line Between Genuine Innovation and Strategic Exaggeration in AI
When you ask three language models – ChatGPT, Claude, and Bing – about 'when and how Stability AI started,' you will receive three different answers with three different dates and a lot of loud words about democratization, open source, human potential, and extensive work in AI. That's what characterizes Stability AI the most – inaccuracy, vagueness, and, most of the time, severe exaggeration. It's an example of how a good talker can dominate the narrative of AI models. It is also an example of how the embezzlement of VCs by buzzwords stops them from careful due diligence.
Stability AI is still in the Unicorn Family with a valuation of around $1 billion. But there are opinions that 2024 will be Stability's last year of existence. In any case, the company and its CEO are very interesting cases to study. Let’s explore the world of Emad Mostaque, the way to build a unicorn company in six days, and how it all can fail.
Starting point of Stability AI
Financial situation – to a $1 billion valuation in six days
The very important question: How does Stability AI make money?
Flagship product and problems of appropriation
The tech behind their products
The latest issues regarding the use of child sexual abuse material (CSAM)
The mission of Stability AI
Approach to AI Governance
Conclusion
Starting point of Stability AI
The Forbes correspondents did a really good job, discovering that Stability AI's journey started in 2019 by Emad Mostaque as 'an AI-powered data hub that global agencies would use to make decisions about Covid-19.' It launched with a virtual event hosted by Stanford HAI in July 2020, with AI expert Fei-Fei Li and representatives from UNESCO, WHO, and the World Bank as speakers. Emad was also a speaker, presenting ‘Knowledge Graph Architecture.’
The project didn’t take off. After that, Mostaque was all over the place: vending machine network, emotional support dog NFTs, etc.
Mostaque certainly deserves credit for seeing the opportunity in Generative AI that, at that moment, wasn't yet a huge thing. He also deserves credit for noticing the potential behind the Latent Diffusion model and lending the computational power he had from AWS to let researchers check and improve the model’s performance.
We don’t know how he presented that help to the researchers, but they did launch an improved model in August 2022 under the name Stability Diffusion. This naming led to another appropriation by Stability AI: it simply made everyone believe that Stable Diffusion equals Stability AI and that Mostaque was the main evangelist behind the model and its open-source nature.
Which was at least an exaggeration but a very beneficial one. It led to ->
Financial situation – to a $1 billion valuation in six days
The rest of this fascinating story is available to our Premium users only. Please →
How was it? |
You are currently on the free list. Join Premium members from top companies like Datadog, FICO, and OpenAI; AI labs such as MIT, and Berkeley; and .gov agencies; as well as many VCs, to start understanding AI and its people deeper→
Thank you for reading, please feel free to share with your friends and colleagues. In the next couple of weeks, we are announcing our referral program 🤍
Reply