In the piece — titled “Can You Fool a Self Driving Car?” — Rober found that a Tesla car on Autopilot was fooled by a Wile E. Coyote-style wall painted to look like the road ahead of it, with the electric vehicle plowing right through it instead of stopping.
The footage was damning enough, with slow-motion clips showing the car not only crashing through the styrofoam wall but also a mannequin of a child. The Tesla was also fooled by simulated rain and fog.
Read about this somewhere. Iirc, Elon felt cameras were better than LiDAR at a time when that was kinda true, but the technology improved considerably in the interim and he pridefully refuses to admit he needs to adapt.
that was never true
Found the article! I had breezed through the thing. I was incorrect about the LiDAR/camera thing. Instead it was: ‘Elon even admitted that “very high-resolution radars would be better than pure vision”, but he claimed that “such a radar does not exist”’
He, of course was incorrect and proven incorrect, but ‘the problem is that Musk has taken such a strong stance against [LiDARs] for so long that now that they have improved immensely and reduced in prices, he still can’t admit that he was wrong and use them.’
I don’t even understand that logic. Use both. Even if one is significantly better than the other, they each have different weaknesses and can mitigate for each other.
It was always just to save money and pad the profit margins
A LiDAR sensor couldn’t add more than a few hundred to a car, surely
And thats a few hundred less profit, so we cant have that.
They ditched radar at a time when radar only added probably about $50 a car according to some estimates.
It may technically get a smidge more profitable, but it almost seems like it’s more about hubris around tech shouldn’t need more than a human to do as well. Which even if it were true, is a stupid stance to take when in that scenario you could have better than human senses.
And to make him think he’s a smart boy.
He didn’t think they were better. He thought Tesla could get away without the more expensive lidar. Basically “humans can drive with just vision, that should be enough for an autonomous vehicle also.” Basically he did it because lidar is more expensive.
Even if humans can drive with just vision:
I didn’t think it was about the cost. I think he just likes to be contrarian because he thinks it makes him seem smart. He then needs to stick by his stupid decisions.
I’m assuming it’s a cost because it makes sense to me. His goal was to build full-self-driving (FSD) into ever car and sell the service as a subscription.
If you add another $500 in components then that’s a lot of cost (probably a lot cheaper today but this was 10 years ago). Cameras are cheap and can be spread around the car with additional non-FSD benefits where as lidar has much fewer uses when the cost is not covered. I think he used his “first-principles” argument as a justification to the engineers as another way for him to say “I don’t want to pay for lidar, make it work with the cheap cameras.”
Why else would management take off the table an obviously extremely useful safety tool?
What makes you think people make rational decisions? Especially sociopaths like Musk?
Couldn’t he just use both… Like LiDAR as a contingency
I added a correction in another reply. Basically he stubbornly refuses to believe a powerful enough LiDAR exists. So I suppose he is all-in on “LieDAR” technology instead (yes, I kinda feel bad about this pun too)
He could. In fact Waymos, for instance, do and are fully autonomous commercial taxis while Tesla are still 2 years out from full self driving for the tenth year in a row
Every LiDAR system must use at least both. LiDAR can’t tell you about lane markings, what’s on signs, and state of traffic lights.
But absolutely, you could have multiple sensing technologies and have access to the best of all worlds.