Musk thinks that is it acceptable to ship features when there is a safety risk, Waymo doesn't
There is also a bit of ego in there as well. Tesla have made a business decision not to use lidar. Its nothing to do with the maturity of the tech. Tesla have spent Billions of dollars trying to re-create a lidar like depth sensor using radar and monocular slam.
Its just not going to cut it for a number of years. CNNs are just not fast or robust enough to replace a lidar. Especially with the quality and placement of the sensors in a tesla.
That is in fact, the top comment. Respectfully, I don't think that this is a reduction though. It's a sidestep. I don't mean that the safety (and other ethical) concerns aren't real, and I accept that there are legitimate criticisms of Tesla/Elon... both philosophically and practically.
That is not the actual debate in this article though. It's not the debate between Tesla and Waymo's CEOs. That debate is about the path to a technological goal: actually autonomous vehicles.
The lidar question is closer to the core of the debate, but IMO it's a detail. Here the disagreement might have gotten unproductive. As you say, egos (famously big & nerdy ones) have have been dragged into it.
>> Tesla have made a business decision not to use lidar. It's nothing to do with the maturity of the tech.
I agree with the first and disagree with the second statement. This, IMO, is actually at the core of the debate. In (my assumed version of) Musk's Modus Operandi, technology strategy and business strategy converge much more than in the Waymo/Krafcik MO... They're at opposite ends of the poll.
Business decisions are R&D decisions. Putting things into production is R&D, affects R&D, determines R&D.
In a caricature/model version of the Waymo MO, technology is invented in a research lab. Once it's ready, it can be productized outside of the lab. "Technology" is defined as the core "proof of concept." In the Musk MO, technology is produced in an industry/company. A lab cannot reduce the price of computing by half every 18 months for decades. A lab can't improve the price/weight/performance of batteries to the point of EV viability. It takes an active industry/market to do these things. You need customers. You need commercially scale factories, increasing annual unit production. Etc. Large and increasing numbers of people working on it.
A lot of technology never leaves R&D. Lunar Travel didn't becomes cheaper over time. This confuses people because we expect technology to improve. Doing something people did 50 years ago should be cheap and easy, intuitively. In reality, that momentum is not a guarantee. If no one works on it, it doesn't improve. Software, hardware and such improve at a monstrous pace. But, every year there are way more people working on hardware and software. If the space "industry" had 100X more people working on it compared to 60s, it would have improved to.
I'm not calling Musk's approach superior, but Waymo's is certainly more common. Novelty itself is valuable. It's interesting that we have both approaches at play. n=2, but at least we'll get to see how this plays out.
Musk thinks that is it acceptable to ship features when there is a safety risk, Waymo doesn't
There is also a bit of ego in there as well. Tesla have made a business decision not to use lidar. Its nothing to do with the maturity of the tech. Tesla have spent Billions of dollars trying to re-create a lidar like depth sensor using radar and monocular slam.
Its just not going to cut it for a number of years. CNNs are just not fast or robust enough to replace a lidar. Especially with the quality and placement of the sensors in a tesla.