In the past couple of months, Tesla CEO Elon Musk’s social media has been a pit of controversies. However, after a recent interview with an MIT researcher, it could be said that it’s not just his social media. During the interview, Musk claimed “game, set and match” for Tesla in the race to reach self-driving cars and “I could be wrong, but it appears to be the case that Tesla is vastly ahead of everyone.” There is a good chance that he is wrong.
A research conducted by Radio Free Mobile (RFM) found out that in a lot of cases, miles driven can be an indicator of success in autonomous driving. But this could have special exclusions too. For instance, the best two players in the autonomous scene, Waymo and Cruise have driven more than anyone else, but at the same time, Apple, having more miles covered was only slightly better than Uber, which is positioned at the bottom of the list for obvious reasons.
In the current situation, the only solution may involve a combination of traditional riles based software and deep learning models, which either way will take time.
On the same note, it would not be difficult to believe that Tesla has gone further than the aforementioned companies as its Autopilot feature is in thousands of cars which are driving thousands of miles right now. Drivers have to keep their hands on the wheel at all times, but this could all be valuable experience from which the autopilot system can learn.
What is clear is that just like Apple or Uber, Tesla’s system cannot handle infinite and unstable environments which is evident from the two fatalities that it experienced so far.
Furthermore, the American EV giant has refused to share any data since 2017 putting a giant question mark on its credibility.
What is clear is that just like Apple or Uber, Tesla’s system cannot handle infinite and unstable environments.
In the past six months, the autonomous driving tech fraternity has accepted the fact that the problem is more complex than originally anticipated, bringing back expectations of commerciality back to “sometime in the next 10 years”. This is majorly due to the fact that the deep learning models have no casual understanding of what it is they are doing. This leads to a catastrophic failure when something unexpected happens.
In such a situation, the only solution may involve a combination of traditional riles based software and deep learning models, which either way will take time. This is why everyone is pulling in the timeline and if Tesla wants users to be able to take their hands off the wheel, it will have to do the same.