Gary Zhou 
A new Bloomberg article popped up in recent days with a headline that sort of irked me. The title was: “Elon Musk Looks Increasingly Isolated As Automakers Embrace Lidar.” I don’t know what is in the article since I’m not a Bloomberg subscriber, but I could guess, since we cover this space quite obsessively. Most automakers are using NVIDIA or Intel/Mobileye tech to improve their driver-assist features and move toward fully autonomous driving, and those hardware packages include lidar. Tesla CEO Elon Musk, meanwhile, has never thought lidar was needed for eventual full self-driving tech and doesn’t even think radar is needed.
That’s all fine and good. Correct differentiation. What irked me is that this isn’t new, so Elon Musk and Tesla are not being “increasingly isolated.” In fact, we wrote about the differences in the approaches in detail in 2015 (7 years ago), and it wasn’t really news then. Back then, as today, almost no one in the industry had the opinion that lidar wasn’t needed. There was Tesla (Elon Musk, most notably) and later on there was comma.ai (George Hotz). I published an exclusive of Hotz discussing this in 2017. The title of the article was “Geohot: Tesla Autopilot = Apple IOS, Comma.Ai = Android.” The argument Hotz made was that only Tesla and comma.ai had the right approach, Tesla would be first to the finish line, and then comma.ai would be a couple years after Tesla but well ahead of the rest.
The point is that Musk’s team at Tesla has always been isolated in their view that lidar isn’t needed. Yes, way back when, the debate was about Tesla versus Waymo, but then there was GM’s Cruise taking essentially the same approach as Waymo. Then there were plenty of other similar startups and big automaker announcements about R&D and contracts in that field. As NVIDIA and Intel/Mobileye developed their driver-assist tech and long-term robotaxi plans, they also shunned the Tesla approach for a broader hardware suite. This is not new, and nothing has fundamentally changed in that regard in, like, forever.
Which Approach Is Right — Tesla’s or Everyone Else’s?
This is what everyone wants to know. Like you, I don’t have a crystal ball. There are argument’s for Tesla’s approach that we made in 2015, 2018, 2019, etc. However, since 2015, Tesla has changed its specific coding/AI approach — and hardware approach — significantly a few times. It has also massively missed some of Elon Musk’s predictions or forecasts. At the moment, Tesla “Full Self Driving” (a software package, not actual 100% self-driving capability) can do things no other consumer product can do. However, as someone who has the package, my perspective is that Tesla is still far away from robotaxi capability. I won’t go into it in detail, but the Washington Post just had a superb story come out explaining the issues many drivers are having, and they also published a great one almost two months ago that was basically a shallower precursor to this new one.
The fundamental questions of 1) whether Tesla will achieve robotaxi capability and 2) whether it will do so far in advance of others are still up in the air, but what is also up in the air is whether or not an affordable, scalable robotaxi rollout is possible that relies on cameras, radar, and lidar. None of these questions are going to be answered this year, in my opinion. I’d love to be proven wrong, and I do leave the door open to that possibility. I think we will see great progress in self-driving this year. However, the fundamental debate of the last 6+ years will probably be the same in a year from now. I don’t believe I will be able to get into my Tesla, tell it where to drive, and then lay back and go to sleep or start working on the computer. It can hardly get out of my neighborhood today and has more glitches than a 1990 Nintendo game. At the same time, economically viable mass-market robotaxis using lidar, radar, and cameras aren’t coming in 2022 or 2023 either. We’re still at early high-cost, low-scale trial stages and tech development.
Ending back where we started, though: Elon Musk is as isolated on this issue as he was in 2015. He’s not budging (well, as I noted, he even ditched radar since then) and none of the competition is budging on their take on the tech. Lidar or no lidar? Make your own guess and pick a side — or just wait and see what happens.
