There is a lot of confusion of how Tesla AI compares to xAI. Both are headed up by Elon Musk. That said, they are two distinct companies.
This much is known to most people. What gets overlooked is the fact that both are separate technologies.
Tesla AI and xAI are not the same. As such, they will not serve the same market.
For this reason, Tesla is going to need a licensing agreement with xAI for it to bring its autonomy to scale.
Image generated by Ideogram
Let us start by defining both technologies.
xAI developed Grok. This is a large language model which is focused upon, no surprise, language. Thus, we look at it as language data.
Tesla AI is visual data. Here we see things such as perception and planning implemented. Obviously, these are necessary to maneuver an automobile.
Certainly, with multi-modal learning, a LLM is something that Tesla could develop. The question would be why?
This is where xAI, with Grok, comes on. Tesla will need the technology, thus coming to an agreement with xAI.
Language is required for both interaction with the vehicles, especially the robotiaxi, and Optimus. Both of these need to interact with humans. Since we tend to operate verbally, it is a skill these devices will require.
Tesla does not have the technology for this.
Enter xAI.
We tend to use the overarching title of "AI". However, there are many different facets, along with varied areas of focus.
Autonomous entities will need to interact with humans. This is an inescapable fact of the world. As stated, a crucial part is the understanding of human language, in its native tongue. This means advanced translation.
Tesla is an example of a company not working on this. Since that is the case, we can safely say none of the other automakers are doing it either. Even in China, we see agreements being reached with the likes of Baidu.
Naturally, this applies to many areas. Tesla's AI is not going to enable one to edit images. Nor will it allow for the generation of videos.
The automotive and robotics industry will partner, in some way, with the companies that have LLMs. The cost associated with developing one in-house excludes even major companies such as Toyota or BMW.
What is an option is turning towards open source solutions. Both Meta and NVIDIA offer this. The latter has a platform dedicated to self driving technology, so that could be viable for some companies.
Like Microsoft, some companies are going to make a fortune by licensing its technology. For the last 40 years, almost every PC was running Windows initially. This allowed Microsoft to become one of the most valuable in the world.
LLMs are going to be a centerpiece of our world in a few short years. We already see remarkable progress. At this point, we are starting to move towards embodied AI. This is where AI is put into physical entities that move in the real world.
Here is where we see another layer added to the "intelligence" stack.
