Hacker News Reader

> I've spent the last half hour trying to figure out what this company is, and I still have no idea. My first stop was their website and I could not find a single sentence providing concrete details regarding their value proposition.

As much as it sucks, they don't owe the general public an explanation of their business plan. They apparently have access to investors and can explain it to them directly.

> but I still am not sure what the relationship is between Stability AI and the researchers that developed the stable diffusion model.

Another post trending on HN today about data laundering says that Stable AI funds the university research behind Stable Diffusion. The data laundering comes from the university using otherwise copyrighted data for "research purposes" and Stable AI can in turn use the trained model commercially, sidestepping the fact that the underlying data is not available for commercial purposes.

Another way to read that situation is simply that it is easier to outsource your research than build your own team.

>As much as it sucks, they don't owe the general public an explanation of their business plan. They apparently have access to investors and can explain it to them directly.

Like Theranos and Magic Leap!

Their first deliverable was wildly successful, shot a hole in the business model of open ai and was praised by generally everyone.

So not quite.

I am curious what the pitch is though. As investors would want return, I assume

Reminds of of that old as rocks joke about the most powerful computer available being graduate students.

Graduate student descent.

> As far as I can tell, Stability AI simply provided the compute for model training.

As models grow complex, Compute is what's holding back AI from mass production and consumption [0]. And so whoever figures out a way to give away Compute for inference and training is likely going to be able to provide the resulting service at much competitive rates to keep the flywheel going?

Sort of similar to Google and Chrome/Search/YouTube/Android, where even if untold billions are spent to build and run the software (Compute), to the end-user (AI researchers) it costs virtually nothing, while Ads continue to bring in revenue (APIs / services).

I see StabilityAI as Yang to OpenAI's Ying. And I believe, that's where the upstart's evaluation stems from.

[0] The new business of AI (future.com), https://archive.is/g4bEe

There is an interesting interview with "the man behind Stable Diffusion" at youtube.

https://www.youtube.com/watch?v=YQ2QtKcK2dA

They wrap up studies into API calls or applications, which being sold to developers or end users.

These second-tier AI company don't need the be the best. "Good enough" with competitive pricing will do.

Thanks. So they will offer similar APIs as OpenAI and Hugging Face?

> I've spent the last half hour trying to figure out what this company is, and I still have no idea

That's not unusual.

As long as the investors understand it, that's fine.

Even though the investors have no clue what it does, it's still fine as long as the investors are willing to write the cheque.

>And if this is the case, is there any reason to expect Stability to stay on the forefront of AI research? Or are they just going to open source models in the wake of actual research done by other organizations?

Even if this was the case, turns out this is more valuable than just holding the model close to your chest and not letting people use it, build off it.

Coatue is literally the definition of an unsophisticated investor. Their AI investments and initiatives have been disasters thus far.

They're the ones that released Stable Diffusion (the ML model that can generate images from text prompts).