Tesla – Unfounded confidence

Tesla is set up to lose the autonomous race.

• Tesla’s removal of radar from its models shipping to Europe and the Middle East is the right choice based on its philosophical approach to autonomous driving, but there is a fundamental flaw in this strategy that is going to ensure that Tesla is an also-ran in autonomous driving.
• Model 3 and Model Y vehicles have been shipping without radar in North America for some time now but the removal of radar from vehicles that are shipping elsewhere indicates that Tesla is not turning from this path which I continue to think will be highly problematic.
• Tesla’s philosophy toward autonomous driving is correct in my opinion.
• Its view is that if humans can drive with two cameras (eyes) and two microphones (ears) then a machine should be easily able to do so if it is equipped with 8 eyes that let it see far more than a human can.
• All things being equal, there is no problem with this approach, but things are not equal as humans have an organic brain while machines have a feeble approximation of it.
• So feeble is this approximation that deep learning systems are not intelligent at all and boil down to being very sophisticated pattern matching systems and no more.
• This limitation means that deep learning is excellent for tasks where the data set is both stable and finite but when anything changes, the system will badly fail.
• The problem with the road is that the data set for it is neither stable nor finite and this is why everyone is having so much difficulty with getting autonomous driving to work properly.
• Furthermore, the problem with autonomous driving is not in what the vehicle does but in its interpretation of its surroundings.
• Almost every mistake that an autonomous vehicle makes is because it has misinterpreted its surroundings and then made the correct decision based on that interpretation.
• Hence, if the machine vision problem can be solved, then I suspect that commercial autonomous driving will very quickly follow.
• This is why Lidar is important because its impact is to make it easier for the vehicle to correctly interpret its environment and consequently, those that use it are likely to get to autonomy before those that don’t.
• Tesla thinks that Lidar is a crutch and, in this regard, it is correct but it is a crutch that Tesla is far from ready to throw away and I think that its autonomous driving solution will be much worse off without it.
• The removal of its radar from its European models is a sign that it has not changed its mind and is doubling down on its camera-only choice.
• Hence, I continue to think that Tesla is going to struggle to get its autonomous driving solution to work as well as its rivals, almost all of whom are using lidar.
• Furthermore, the cost argument not to use lidar is rapidly falling apart as there are plenty of companies all working on making high-performance lidars at a tiny fraction of the historic cost.
• So much effort is being put into this that I see lidar going the way of Wi-Fi where it becomes so cheap (in automotive and robotic terms) that everyone includes it in their vehicles.
• Consequently, I see Tesla being left behind in autonomous driving which is a major reason why I believe that the share price is very overvalued.
• The graveyard is full of traders and investors who had the temerity to short Tesla which means there is no way of knowing how or when the valuation is going to crack.
• Hence, a better strategy is to sift through the wreckage of the EV and component SPACS many of whom have fallen by 65% or more and look for good companies that have been tarred with the same brush.
• Here my top pick remains Ouster (solid-state lidar maker) which has real revenues unlike almost all of its peers.
• I bought Ouster about a month ago and remain very comfortable with it until at least $8.5 per share.
• I would not go anywhere near Tesla on either the long or short side.