It’s High Time Apple Bought Stability AI

Big techs working on generative AI have their own research labs doing all the work for them. Most recently, Oracle revealed that it would be working on generative AI services based on Cohere’s technology. At the same time, Amazon has announced its partnership with Anthropic, and an investment of $4 billion for competing against Microsoft, Google, Meta, and NVIDIA. 

But what about Apple?

Let’s not forget Apple’s affinity for Stability AI. Since December last year, Apple has been optimising its silicon chip to run Stable Diffusion and other Stability AI products. After releasing Transformer for Apple Silicon, the company also open sourced Stable Diffusion XL on Core ML. 

The blog reads, “One of the key questions for Stable Diffusion in any app is where the model is running. There are a number of reasons why on-edge deployment of Stable Diffusion in an app is preferable to a server-based approach. First, the privacy of the end-user is protected because any data the user provided as input to the model stays on the user’s device. Second, after initial download, users don’t require an internet connection to use the model. Finally, locally deploying this model enables developers to reduce or eliminate their server-related costs.”

Since Apple has been all about privacy, and it bets on the fact that Stability AI’s offering are best for on-edge computing, it seems like Stability AI would be an ideal choice for Apple for building privacy-focused and on-device generative AI products. It seems like the partnership is already in the works in the backdrops.

Last year, Apple introduced OctaneX, a GPU rendering feature that can be found through the Mac Store, allowing rendering on Apple M1 and M2. Octane has reportedly used the Render (RNDR) token, which was built by OTOY, a company that provides decentralised GPU-based rendering solutions. There is a speculation that Apple might actually partner with OTOY.

This matters because in April, Stability AI launched StableLM, which was touted as an open source rival to ChatGPT. Jules Urbach, the founder of OTOY, posted on X saying that RNDR has a large GPU for powering AI inference, which is going to be available on Apple Neural Engine Soon. To this Emad Mostaque, founder of Stability.AI, replied that they “should probably chat”.

If Apple partners with OTOY, which seems to be in talks with Stability.AI, it is quite likely that Apple might be interested in talking to Emad Mostaque himself. In an interview, Mostaque said that he wants to run Stable Diffusion on mobile devices. 

Why is it high time for Apple?

Apple seems to care less about building upon the generative AI miracle. The company built an “Apple GPT” for its employees on the Ajax foundation. But nothing happened after that. Even at the recent Apple Event, the company was not gung-ho about generative AI, and very subtly integrated them in its products.

It may seem like Apple is staying away from generative AI labs. But though it doesn’t like to say the word “AI”, it definitely is leveraging machine learning within its products. The company that was the leading buyer of AI companies a few years back, now seems to be staying away from investing in them.

Undoubtedly, Microsoft’s investment in OpenAI, though not a surprise, was definitely a strategic move that helped the big-tech leverage generative AI within its products. This led others in the field to adopt the same strategy. Google’s got DeepMind, Meta has Meta AI, now even Tesla has xAI. Now it is time for Apple to have its own AI research lab which would do all the work for it.

Steve Wozniak and Mostaque were one of the signatories of the pause giant AI experiments letter. But Apple’s approach towards its products of keeping it simple and user-centric might hinder its step towards partnering with the open source champ Stability.AI. Still, they both seem to be the best partners for each other. 

Meanwhile, OpenAI is partnering WHOOP to compete with Apple hardware products by roping in Jony Ive, one of the main designers of Apple iPhone.