As of February 16th, 2024 you can find more than 509’000 models freely available and ready to use on hugging face. The models range from Natural Language Processing to Computer Vision, Graph Machine Learning, and Audio. The vast spectrum covers all the significant advents from Question-Answering-Chatbots to Autonomous Car Simulators based on Neural Networks.

Let’s look at Tesla, in particular, the technological innovation we appreciate, but at least I do not understand completely.

With Autonomous Driving, we see that Tesla AI combines a whooping 48 Neural Networks to make the Autopilot happen. 48 artificial brains can process input simultaneously and create outputs, such as taking command over the steering wheel and the throttle.

In Elon’s words: ”Our per-camera networks analyze raw images to perform semantic segmentation, object detection, and monocular depth estimation. Our birds-eye-view networks take video from all cameras to output the road layout, static infrastructure and 3D objects directly in the top-down view.”

The 48 Neural Networks are learning from the most complicated and diverse scenarios. The 48 digital brains can create 1,000 distinct predictions at any given time. Comparing this to a human brain, tesmanian.com argues that “Tesla cars using Autopilot are over nine times less likely to get into an accident than human drivers in other manufacturers' vehicles.” Further, luxurylaunches.com says that the “new D1 microchip manages to carry out 362 trillion operations per second. In comparison, the human brain is capable of one quadrillion (1,000,000,000,000,000) operations per second.” The D1 microchip capacity is almost one-third of the human brain’s capacity.

Now let’s imagine that not only graph (network) data science models for neural networks are available, but all kinds of models are available on Hugging Face. The entry barrier for every enterprise to create or adapt existing models is lowered massively. With that in mind, it is easy to foresee that every enterprise that is developing code today is or will eventually create models for their business.

We are running “all guns blazing” towards a new world! A beautiful and colorful one, one might add.

So how do we get from RAGs to Riches? A possible answer is with Retrieval Augmented Generation (RAG), so the Riches will come from the actual RAG.

Show me the RAG

In super simple terms, Retrieval-augmented generation (RAG) means, that you can chat with your PDF! Of course, it is not only your PDFs but your company’s entire wealth of knowledge by using the exact same embedding model and language model that ChatGPT is applying. Isn’t that next level?

Alright, I did not get it the first time. I understand it isn’t easy, but rather sophisticated and more complex to explain. Let’s see how ChatGPT would explain it:

Basically, you can create any business process today and rewrite it using RAG to make it engage with your company’s wealth of data points, information, and knowledge to provide a far superior service!

The broad potential shows why companies including AWS, IBM, Glean, Google, Microsoft, NVIDIA, Oracle and Pinecone are adopting RAG… But your company does not need to wait for those tech companies to provide those new fancy Services but rather decide to become an AI Model Creator yourself. Why not?

Enterprises are AI Model Creators

Imagine the applications we will see soon for doctors, accountants, and software sales. The partner ecosystems will get access to the knowledge base of their provider. The access can even be opened to the knowledge base of an entire ecosystem, enabling their customers to chat and inquire about all aspects of the ecosystem.

Imagine a webshop where you can query the knowledge base of every supplier and sub-supplier, giving the webshop’s visitor any technical or business information on that product to the most granular information piece.

Jensen Huang, President of Nvidia with a net worth of 65.3 B USD, says, “This is the beginning of a new industrial revolution. This industrial revolution is about the production not of energy not of food BUT the production of Intelligence. And every country needs to own the production of their own Intelligence, which is why there is the idea of Sovereign AI.”

This is precisely where Phoenix’s Sovereign Cloud and the Sovereign AI Cloud come into play: