Elon Musk addressed the upcoming debut of the Tesla Fully Automated Beta, claiming that the technology is "superhuman" when it comes to highway driving and "quick to reach in urban areas."

Tesla already has a robust suite of driving aids and driverless features that cannot be legally used without someone in the driver's seat. However, FSD Beta V9.0 definitely shows that we are getting closer to fully autonomous vehicles that can legally drive around on the highway without anyone in the driver's seat.

Not shy, Mask tweeted: "FSD Display V9.0 shows the actual probability distribution of objects. This is so cool!"

Somewhat simplistically, this means that the AI technology basically predicts the road ahead and any obstacles that might get in the way of an autonomous Tesla electric car. Basically, it uses complex mathematics and computer vision to determine the probability of an object getting in the way of the car.

Musk's claim means that the advanced nature of fully automated driving can depend entirely on "vision," which is what the car's camera sees, rather than on a combination of radar and camera; this is a bold statement, as LiDAR technology has been at the heart of many self-driving car systems

Also, the "vision" of the car is not only the "vision.

It also raises the question of what happens if dirt, mud, rain, or other muck kicked up while driving covers these cameras. But if the FSD system does not use radar, that could be a blind spot. Certainly, such things may not be a problem on the dry, sunny roads of California, but on the rainy roads of England, filthy cameras could be very problematic.

And again, Mask states that the technology is so advanced that such things may not be a hindrance. It should also be noted that FSDs are not expected to achieve Level 5 autonomy. This is the same as Tesla's Autopilot system.

Musk also tweeted that AI technology has not yet reached full self-driving systems because roads are designed for humans and their gray matter neural networks, not advanced AI systems.

"In order to achieve unsupervised, generalized, fully automated driving, we have to solve key pieces of real-world AI. [That's why Musk and Tesla are concentrating on autonomous driving systems that prioritize vision over radar.

And Musk noted on Tesla's Q1 2021 earnings call that when a vision-based AI system works, it will outperform humans.

"And when vision works, it's like having eight cameras, it's like having eyes in the back of your head in addition to your head, and it works better than the best humans because it has three eyes with different focal distances looking forward," Musk said.

The Tesla boss noted that it is a tricky technical problem to solve and requires "real-world artificial intelligence."

One of the main reasons for this is the need to figure out that it is not easy for AI to predict and analyze obstacles and road features. For example, a car being carried by a truck or other vehicle, or an object hanging from it. The goal would be for computer vision to have a positive view of the road ahead. Apparently, more data is needed, which would require more testing of driverless cars in real situations.

However, as technology continues to advance and self-driving cars continue to be tested, Level 5 autonomy may find its way into future EVs and mass-produced vehicles. There are regulatory issues to overcome first, but self-driving cars are coming anyway.