7 hours without takeover: Tesla releases its longest-ever FSD demonstration. Musk promises to upgrade automatic charging, making driving as easy as playing a game.
How far has FSD evolved?
Tesla recently released its longest-ever FSD demonstration video on social media:
The journey from San Francisco to Los Angeles covered a total distance of 580 kilometers in 7 hours without any human intervention.
Some people exclaimed "crazy," but others thought it was not perfect. This fully automatic driving was not "fully" automatic. The operator had to get out of the car for manual charging, and the success rate of automatically driving into the Supercharger station was not high enough.
Elon Musk quickly stepped forward and promised that the new charging function of FSD would be realized soon.
FSD with 7 hours of zero takeover, new charging function is coming
Tesla just released a new FSD demonstration video, showing a journey from San Francisco to Los Angeles, covering a total distance of 362 miles (about 583 kilometers).
In the video, the operator in the driver's seat never touched the steering wheel. The entire journey lasted nearly 7 hours, including the time spent charging at the Supercharger station.
It can be seen that when two-thirds of the journey was completed, the operator got out of the car to charge.
So far, this is the longest FSD demonstration video publicly released by Tesla, and this version is likely to be the one launched for private car owners later.
In the comment section below the video, many people were amazed at the rapid development of autonomous driving, and some even exclaimed "crazy."
However, some people also pointed out the problem, which was charging.
A car owner said that Tesla still needs to solve the problem of automatically driving into the Supercharger station. According to his experience, the success rate was not high.
His comment was quickly replied to by Elon Musk: This function will be realized after a software update in the future.
Elon Musk promised that before car owners go to the Supercharger station, they can check how many parking spaces there are and which specific parking spaces are available. In this way, Tesla can identify these parking spaces and park automatically, making the parking ability more reliable.
This seemingly insignificant function will be much more convenient for car owners with FSD, and it is even more significant for Robotaxi.
According to Tesla's plan, car owners will be able to join its Robotaxi fleet next year. When car owners don't use their cars, they can let their vehicles go out to work as taxis to earn extra money.
However, for this model to work, solving automatic charging is the primary threshold to cross.
This automatic charging process can be divided into two parts:
One part is to accurately drive into the charging point automatically, which is the function that Elon Musk promised to upgrade.
The other part is to let the Supercharger and the vehicle automatically complete the connection and charging without human intervention.
Previously, Tesla's proposed solution was "snake-arm charging." The charger, shaped like a snake, could sense the car's charging port and then automatically plug in for charging.
However, it is quite difficult to implement this solution. For example, Elon Musk once revealed that making a metal gear snake-shaped automatic coupler was an important problem.
So Tesla changed its approach and is currently trying wireless charging.
Tesla acquired a wireless charging company called Wiferion in 2023. This company won orders from hundreds of enterprises in five years and sold more than 8,000 sets of equipment with its low-loss wireless charging technology.
Of course, adopting this charging method is not easy either. There are problems such as energy loss, short charging distance, and difficulty in increasing power. It may still take time to see Tesla's research results.
If all the problems of automatic charging are solved, Tesla's Robotaxi can not only attract more car owners to join and expand the fleet size but also operate around the clock without considering charging time.
Although it has not been truly launched yet, it can already be seen from the operating scope of Robotaxi that Tesla is preparing for that day.
Tesla's Robotaxi fleet is currently operating in Austin, Texas, and the Bay Area of California. It is also applying for approval in other states such as Nevada, Arizona, and Florida, aiming to achieve the goal of "half of Americans taking a Robotaxi by the end of this year."
The latest development is that the next city to enter is New York. One characteristic of New York is that many people don't have cars, so there is a high demand for taxis.
Tesla now needs to wait for approval to test autonomous vehicles. However, it has already "prepared the supplies" and started recruiting operators for autonomous vehicles in New York, requiring drivers of autonomous vehicles to have "excellent driving skills."
It is unknown when it will be launched in New York, but the huge changes that Tesla's Robotaxi will bring to New York can already be foreseen.
Some foreign media analyzed that after Tesla's Robotaxi enters New York, it may "crush" New York's iconic yellow taxis and quickly gain the upper hand.
There are roughly three reasons:
Firstly, the operating cost and fare are lower. New York implements a taxi franchise system, and taxi drivers need to pay high license fees. It is much more cost-effective to buy a Tesla Robotaxi for operation, and passengers don't need to give tips and other benefits when choosing Robotaxi.
Secondly, Robotaxi can operate continuously throughout the day, ensuring the number of vehicles in the fleet even at night.
Once private car owners can also join this Robotaxi platform, the waiting time for a taxi will be much shorter than that of traditional taxis.
Thirdly, Tesla's Robotaxi will provide a relatively better riding experience. For example, passengers can choose the cabin temperature and music by themselves, and the additional rear screen can also be used to select various games or YouTube videos.
△
Of course, these imaginings are still based on Robotaxi getting the license and the charging problem being solved.
In addition to the charging function, there is another interesting upgrade of FSD that has been exposed.
Unreal Engine comes to the car, driving will soon be like playing a game
A Tesla hacker named "greentheonly" recently discovered through the code in Tesla's 2025.20 firmware that the company is planning a major upgrade to the driver visualization function of Autopilot.
Tesla's current visualization function shows the vehicle and its surrounding environment on the central screen and displays the driving route in autopilot or full self-driving mode.
Although Tesla's visualization function has been gradually improving over the years, with information such as the shapes of pedestrians and other vehicles added, such a significant improvement is the first in recent years.
The hacker said that Tesla plans to use UE (Unreal Engine) to achieve autonomous driving visualization, while the current engine is "based on Godot."
Moreover, according to his disclosure, this function has been installed on Model S/X models based on AMD processors but has not been activated yet.
To briefly introduce, UE is a 3D computer graphics game engine developed by Epic Games, the developer of the popular game "Fortnite." The latest version is currently UE5, and UE6 is expected to be released in 2027 or 2028.
Godot is also a game engine, suitable for 2D or small and medium-sized 3D projects.
The difference between the two is that Godot is a lightweight open-source engine with an installation package of only 8GB, while UE is a commercial engine with a more powerful resource library and graphics rendering ability.
By switching to UE, Tesla can present the vehicle's surrounding environment more realistically, making driving feel like playing a game and improving the driver's visual experience.
It may be something like this:
This kind of experience of driving like playing a game has already had small attempts in China:
When XPeng released its new P7, it launched a "anger relief" function, which is realized by combining HUD and launching props.
After car owners set the custom steering wheel button, they can click to trigger the "monster - hitting" function and throw toy bags at surrounding cars.
It seems that making driving more interesting is a new field.