What Tesla’s real-world AI could mean

Editor’s Note: Tesla AI Day is here. But many people who are very eager to see what is revealed turn to their computers for something new or stimulating as a snack. Chanan Bos wrote a tremendous piece earlier this year about Tesla and what Tesla’s “real-world AI” could mean, well beyond better autonomous driving systems and robotaxis. It seemed like a good time to relive the last part of that piece for anyone who missed it, so here it is again. Enjoy! And let us know in the comments what you have to add. —Zach Shahan

Elon’s clues about real-world AI

Now we get to the really exciting part. Elon recently posted some tweets that imply something I’ve been wondering about for a long time and that brings us back to the question of the robot that builds the robot. Here are the tweets:

AI from the real world, AI from the physical world. What Elon means is the step in which we move from object recognition to understanding the world around us. Ideally, we would like to place more robots in warehouses, factories, and many other places. Those examples are a very different real world than the streets a car drives on. The same can be said for a hypothetical robot maid / butler at home or a robot chef in a kitchen. Those are all real-world AI problems and automating the AI ​​learning process is the only realistic solution if you want to go beyond the basic functionality where a team needs to sustain AI every step of the way. towards basic functionality. In almost all of those apps, collecting a ton of data is extremely difficult because there aren’t millions of people volunteering to take care of an AI like they do on Tesla’s autopilot.

If someone could fill an empty warehouse with prototypes and chef robot arm kitchens and have everyone learn day and night automatically, continually improving and getting a salable product in a couple of years, investors would throw so much money at them that developers would be they would drown. in that. The less human involvement required, the more scalable it becomes.

Everything Tesla has learned and developed for FSD could one day be used to train all kinds of real-world AI. It’s funny because some people are already a little skeptical about whether the FSD will ever work and whether a future Tesla FSD Robotaxi is already included in the current valuation of the stock. However, what if I told you that Tesla could be the key to the real-world AI problem? If Tesla continues to automate this process, then the real-world AI market might have to prepare for a major disruption. As for the tweet above, it depends on how you interpret it. If our theory is correct, what Tesla could offer could be much more than raw processing power. Somehow this will work like Google Cloud’s AutoML (machine learning), but instead of a little machine learning operation, which is good for learning something useful from a bunch of images, it will be a set of tools to train an AI. real-world operation, be it a kitchen or warehouse delivery system.

It’s not the robot, it’s the AI

By now, we’ve all seen the amazing things that the walking robots of Boston Dynamics they are able to do either with four legs or with two legs. In fact, SpaceX recently used one to inspect the Starship SN10 crash site. The breakthrough is that Boston Dynamics managed to solve the problem of walking like animals, maintaining balance, not falling, and more recently, being able to understand where they can walk and how to overcome obstacles.

However, when it comes to truly understanding their surroundings and interacting with them, they are not that smart and desperately need training. They may also lack the necessary built-in processing power, but mostly, AI is the problem, as no AI has really solved the problem of real-world AI, which is essentially a learning problem. Solving this in the case of robots can make new and incredible things possible.

The auto body shop at the Tesla factory in Fremont. Image Credit: Chanan Bos | CleanTechnica

The robot that builds the robot.

Tesla has been working for years on the machine that builds the machine, which is in essence a highly automated factory, but if we get to a point where one robot can build another robot, then exponential growth becomes possible. Unfortunately, the direct multiplication of robots in this way is still a long way off. However, if a group of robots can be programmed to build a factory on the moon or on an asteroid from local raw materials with minimal input from Earth, a factory that can build more robots suddenly becomes possible. massive Kardashev 1 * scale projects. In some ways, it is like the von Neumann probe *, but now in the form of hardworking robots capable of converting the raw materials of the solar system into what we need.

* The Kardashev scale is a measure of how advanced a civilization is where right now we are type 0 and type 1 is technically capable of using and storing all the energy available on your planet.

* Von Neumann probe: the idea of ​​a self-replicating spacecraft that, upon reaching a new solar system, uses local raw materials to replicate itself and sends replicas to nearby stars to do the same. With conventional slow space travel, probes could fully explore our galaxy in just half a million years. When applied to construction, self-replication or even regular replication allows for exponential growth and incredible mega-projects.

For years I have wondered why Elon does not have a company like Boston Dynamics that focuses on the problem of the robot that builds the robot. However, once again, assuming you are working towards it, you may have foreseen the bigger problem, which was not the hardware but the AI ​​that will power it, in which case you found the most cost-effective and realistic way to finance the development of methodologies. needed to decipher real-world AI. However, I really wonder how much of that was actually part of the plan when he decided to go on autopilot in 2012/2013.

How real-world AI could fit into Elon’s grand plan

Elon Musk’s ambition, as he has said several times, is to preserve the existence of consciousness. This can be broken down into just 3 different subgoals and various technologies required to achieve the goal. The easiest to define is to make life multiplanetary using SpaceX’s rapidly reusable rockets to build a civilization on Mars. The second is to preserve the earth and the humanity on it. That would be done with the help of Tesla, which will help the world’s transition to renewable energy and ensure that this world remains livable. Then the last one has to do with AI taking over.

OpenAI it was originally intended to prevent the AI ​​from extinguishing consciousness. Although instead of being the savior we need, they themselves may have become the villain they were supposed to save us from, there is a lot of irony there. What does Elon keep saying? Does fate love irony? The flip side of that coin, founded around the same time, is that Elon started Neuralink so that we can join the AI ​​if we can’t beat it, something so that consciousness at least remains relevant alongside the minds of AI.

The Boring Company and Starlink can be credited with helping Tesla and SpaceX achieve their respective goals. Starlink by providing funding and The Boring Company by helping to avoid traffic on Earth and possibly by creating places for us to live underground on Mars, because all the radiation makes the surface dangerous.

However, if we can solve the problem of real-world AI, then we can make robots do much of the dirty work. Populating Mars without them would take much longer. Once again, robots and real-world AI can drive progress.

Do you appreciate the originality of CleanTechnica? Consider becoming a Member, collaborator, technician or ambassador of CleanTechnica – or a patron in Patreon.


Have a tip for CleanTechnica, want to advertise, or want to suggest a guest for our CleanTech Talk podcast? Contact us here.

Leave a Comment