Tesla is now the most valuable automaker in the wold by far with a market capitalization of over $800 billion. Investors view at as much more than just an automobile manufacturer. One of the key bull cases for Tesla is their full self diving technology which they plan to eventually use to develop a robotaxi network. Unfortunately, developing the FSD technology is much easier said than done. It now looks like we may never see a fleet of autonomous robotaxis in our lifetimes.
Email us: Wallstreetmillennial @gmail.com
Check out our new website: Wallstreetmillennial.com
Support us on Patreon: https://www.patreon.com/WallStreetMillennial?fan_landing=true
#Wallstreetmillennial #Tesla #FSD #Musk

––––––––––––––––––––––––––––––
Buddha by Kontekst https://soundcloud.com/kontekstmusic
Creative Commons — Attribution-ShareAlike 3.0 Unported — CC BY-SA 3.0
Free Download / Stream: http://bit.ly/2Pe7mBN
Music promoted by Audio Library https://youtu.be/b6jK2t3lcRs
––––––––––––––––––––––––––––––

What's up guys and welcome back to wall street millennial on this channel, we cover everything related to stocks and investing with the meteoric rise of tesla's share price over the past two years it has become the most valuable automaker in the world by far with a market Cap of more than 800 billion dollars for comparison, this makes it almost three times more valuable than toyota, which has a market cap of 300 billion dollars. This is despite the fact that toyota delivered 10 million vehicles in 2021, which is roughly 10 times tesla's deliveries in the same period. Clearly, investors don't view tesla as just a car company. They view it as an innovative technology company which will eventually have revenue streams.

Far beyond just selling automobiles, one of the key bold cases for tesla that could justify its current valuation is its full self-driving technology and their eventual plan to make a robo-taxi network. If they're successful in building a robo-taxi network, it's hard to overstate how valuable this could be as a business, they could dominate the ride-hailing market by undercutting uber and lyft on price because they don't have to pay a human driver. The robo taxi network alone can be worth trillions of dollars and propel tesla to become the most valuable company in the history of the world. In a 2019 investor event, elon musk said that he was very confident that they would have a robo-taxi network ready for operation.

In 2020., i feel very confident predicting uh autonomous rover taxis for tesla next year, not in all jurisdict, not all jurisdictions, because we won't have regulatory approval everywhere, but i i'm confident we'll have at least regulatory approval somewhere literally next year. Unfortunately, building a trillion dollar technology is easier said than done. It is now 2022, two years after musk's original target, and we still don't see any tesla robo taxis on the road. The hold up has come from problems with their full self-driving or fsd technology, which is currently being beta tested by tens of thousands of tesla owners.

While musk says that self-driving cars will be far safer than human drivers, it appears that we're still years away from this coming to fruition. Since the release of fsd beta, there have been numerous cases of autonomous, teslas crashing running red lights and making wrong turns where an experienced human driver wouldn't have had any problems. In many complex urban driving environments, it looks like fsd may be far less safe than a human driver in this video. We'll look at the current problems with full self-driving and whether we will see the robo-taxi network within our lifetimes.

Tesla has been planning fsd for a long time. Since 2016, every tesla car sold is equipped with eight cameras and the necessary computational hardware to be fully autonomous. The last missing piece is perfecting the software to recreate the brain of a human driver since 2016. They have offered autopilot features on their cars autopilot dynamically adjusts the driving speed to match other cars on the highway keeps you within your lane and changes lanes for you autonomously, when you tell it to while this is a step function, improvement over traditional cruise control systems, It still requires the driver to monitor the situation on the road at all times.
It also only works on highways. Highways are the easiest places for autonomous driving because they are long, simple roads with predictable movements from other cars. Urban environments are much more difficult because there are an almost infinite number of intersections pedestrians. Road signs and reactions from other drivers.

Tesla's plan for self-driving technology is very similar to their strategy for vehicle manufacturing. They start off with the easiest track, producing the low volume high price point roadster. They reinvest the revenue from the roadster and build off of their manufacturing expertise to build the cheaper model; s and eventually the mass market model 3, which was their ultimate goal. Similarly, with self-driving, they started off with highway autopilot, which is the easiest challenge.

Tesla collects data from all autopilot, enabled vehicles over the internet and uses these as real-world simulations to train their artificial intelligence. They developed their own proprietary supercomputer, which they call dojo so far, they've collected 1.3 billion miles of real-world driving data, which they use to refine their neural networks and get closer to full autonomy. Tesla is the only car company that collects real world data from cars on the road at scale, so their data is orders of magnitude bigger than waymo gm cruise or intel's mobile eye. Musk says that this will be key to tesla's success at building the robo-taxi network.

In october of 2021, tesla released their full self-driving beta program to tesla owners with high driving scores. It is installed on the car as an over-the-air software update and originally cost. Eight thousand dollars with the price recently increased to 12 000, despite the name of full self driving, it is technically only level 2 autonomy. That means the human driver is still required to sit in the front seat and be ready to take control of the vehicle at any time.

Since then, tens of thousands of tesla owners have bought the package and have posted youtube videos showing their experiences. While fsd can't handle most simple driving scenarios, it often has trouble with complex intersections and non-standard street signs. For example, the youtuber ai addict posted a video of him driving through the streets of san jose california, using full self-driving. He had to take manual control of the car multiple times as it ran red lights, attempted to drive into the tram lane and even crashed into a construction pylon.

Obviously, it's not fair to judge fsd based on a single video. The safety engineer, dan o'donne combed through 21 youtube videos of tesla drivers showing their tesla fsd experience, and the results were pretty bad. 15 of the videos were using fsd version 10, which is the most up-to-date. Every single video had at least one scoring maneuver error with one of them having 17.
all, but four of them had at least one critical driving error and five of them had at least one likely collision. A likely collision is a situation where the car could have crashed on its own. Had the driver not taken over a critical driving error is defined as unintentionally, making contact with an object, ignoring stop signs or pedestrians or making dangerous maneuvers requiring other vehicles to take evasive action. A scoring maneuver error is a more minor violation, such as failing to keep an adequate distance from other cars or causing confusion in the street with unusual driving patterns according to dan o'donne.

If the tesla fsd tried to take a driving test without human intervention, it would be denied a driver's license in the majority of cases in november of 2021. Only a month after fsd beta was made widely available. A driver experienced a severe crash in california, while fsd was activated. The car went into the wrong lane and crashed into another car while taking a left turn left turns are generally considered more dangerous than right turns, as you have to navigate through incoming traffic.

On the other side of the intersection, luckily, nobody was killed in the accident, but the car suffered severe damage. Despite musk's claims, it looks like we're still many years away from the robo-taxi network. The national highway traffic safety administration is currently investigating tesla for 11 crashes of tesla autopilot cars involving first responder vehicles such as fire trucks and ambulances. But it's also important to put tesla's problems into perspective.

It's impossible for any autonomous driving system to be completely infallible to accidents, no matter how good it is. There can always be situations where other cars do something crazy and the tesla gets into a crash at no fault of its own. In 2020, 42 000 people died in car accidents and it is the leading cause of death among young people. From a public policy perspective, we don't need autonomous vehicles to be completely safe.

We just need them to be safer than humans. Tesla publishes crashed data for all tesla cars with and without autopilot it is measured in millions of miles driven per accident, so the higher the number, the better on the surface it looks like autopilot, is a huge winner in the fourth quarter of 2021 teslas with autopilot Engage had one crash: every 4.31 million miles teslas without autopilot engage experienced one crash, every 1.59 million miles, and even that is way better than the one crash. Every 484 thousand miles of the national highway traffic safety, administration or nhtsa reports for all vehicles in the us. Based on these numbers, it looks like everyone should buy a tesla and use autopilot as it reduces your chances of a crash by tenfold.
However, the numbers are not as good as they first appear: the nhtsa reports, every incident that results in a police report. The vast majority of these are minor, fender benders, which do not result in bodily harm and do not trigger the airbag tesla has not disclosed what it considers to be a crash. It's possible that they only record a crash when the airbag is activated, which would not include the fender benders. But despite that, it's at least good to see that the chances of a crash when autopilot is activated is less than half of that when it is not activated.

These are both tesla's numbers, so they should be comparable as far as what they consider to be a crash. However, autopilot is usually engaged on long straight highways. Crashes are far less likely on highways because there are no intersections pedestrians or complicated road signs. It's hard to tell how much of the safety improvement can be attributed to the autopilot technology and how much can be attributed to the fact that highways are just safer and remember.

This is only autopilot. Tesla has not yet released comprehensive crash test data for the full self-driving beta, but even if they do, it will be difficult to interpret to qualify for the beta. You need to have a high safety score. The people who have a high safety score will probably be less likely to get into a crash, even if the fsd software is faulty because they can take back control of the car to prevent the crashes.

The fsd system has an incredibly difficult task. Pretty much. All it has is the eight cameras on the car. It has to view video in real time and make decisions about what to do.

They are essentially recreating the brain of a human driver using computers. This is something that has never been done before. It may not even be possible with current technologies. Tesla already has 1.3 billion miles of driving data and has been using it to train their neural networks for years and even now, they're nowhere close to a working robo taxi in 2018 apple, co-founder, steve wozniak, said that he doesn't think the technology is possible and that Tesla will never have a robo-taxi network.

Tesla has no no um, no autopilot, they call it beta. Wait, i'm sorry what car company puts out a feature and calls it beta that doesn't count. Tesla makes so many mistakes. It really convinced me that that auto piloting and auto steering car driving itself is not going to happen, and i believed it and i kept upgrading my tesla to one that would have a camera and radar and then one that would have eight cameras radar, because the First, one would never do it and then i gave up - and i said it's really not going to happen.
You said quote: i've just been fed too many hopeful wishes and lies about the future and i've given up on tesla and elon musk, believing anything they say. That's pretty that's pretty harsh, i don't like it when i feel duped, and now, almost four years later it looks like he's been right so far, alright, guys that wraps it up for this video. What do you think about tesla's, full self-driving technology? Do you agree with elon musk, that's right around the corner, or do you agree with steve wozniak? That will never happen. Let us know in the comment section below as always.

Thank you so much for watching and we'll see you in the next one wall, street millennial, signing out.

By Stock Chat

where the coffee is hot and so is the chat

17 thoughts on “Tesla has a big autonomy problem”
  1. Avataaar/Circle Created with python_avatars Jay Tang says:

    their self driving will never fully work at least not within out life times
    we will need 4g if not 5g embedded roads for "full self driving" capable of being a robo taxi
    basically a completely new generation of infrastructure

  2. Avataaar/Circle Created with python_avatars BradK says:

    Isn't Apple planning on making an electric car and competing with Tesla?
    If so Wozniak's comments mean nothing.

  3. Avataaar/Circle Created with python_avatars MrLynchInRealLife says:

    Here's a challenge, see how long you can go without making a video about or referencing Elon Musk or Tesla

  4. Avataaar/Circle Created with python_avatars ScrewCollege says:

    Wrong they will have robotaxis by this month

  5. Avataaar/Circle Created with python_avatars SomeGuy says:

    With an appearance by the Great and Powerful Woz!

  6. Avataaar/Circle Created with python_avatars XantheFIN says:

    Its not possible as they need be evolving constantly faster than world changes around them. Computers are never as well learning or awake as humans can as they are limited with sensors (calibration, sensor limits etc) and if any sensor fails the whole car fails.. human can adapt environment but computer can act faster on clear thing like crash warning systems can make emergency breaking in clear situation.

  7. Avataaar/Circle Created with python_avatars john patrick says:

    There are driverless/ no humans aboard, non-remote controlled autonomous semi-trucks driving in AZ on Interstate 10 during the day, night, and in fog.

  8. Avataaar/Circle Created with python_avatars plague180 says:

    I feel like you don’t understand the technology you are attempting to talk down on. I normally love your videos but I see at least 10 factual mistakes, unfair framing, or worse quoting people who clearly have it out for Tesla. I wouldn’t call that a good source of information.

  9. Avataaar/Circle Created with python_avatars Moon Woolf says:

    People actually paid $8-12,000 to potentially crash test dummies and have their insurance rates increased smh

  10. Avataaar/Circle Created with python_avatars Samson Soturian says:

    Robo taxi profit potential is wildly overstated since it is based on current uber prices. If robo taxis were widespread they'd be dirt cheap as taxis available would exceed demand.

  11. Avataaar/Circle Created with python_avatars logicxhardcorex says:

    fuck tesla they are against the right to repair and want to monopolize working on a car lol what a fucking joke

  12. Avataaar/Circle Created with python_avatars Insidious DrNine says:

    Someday it will be possible, the average person on the road isn't that fantastic at driving. An autonomous racing series would help. I don't like Tesla's odds, when they eschew LIDAR. I would want cameras, lidar, back up lidar, backup to the backup….

  13. Avataaar/Circle Created with python_avatars FriedChicken says:

    That's why it is call FSD beta

  14. Avataaar/Circle Created with python_avatars SomeGuy says:

    #4 ! Yea

  15. Avataaar/Circle Created with python_avatars Alavi Mirza says:

    first

  16. Avataaar/Circle Created with python_avatars Basketball Extremez says:

    Great content !

  17. Avataaar/Circle Created with python_avatars Basketball Extremez says:

    So early

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.