There is no magic tech Tesla has that Waymo doesn't or did not think about.
Tesla will have to face the same uphill battle to gain experience and win trust, and it is starting from a lot worse position when it comes to reputation for engineering and safety.
Not using lidar is by now foolish, as it is cheap and adds an independent datastream that is critical in borderline situations when cameras degrade.
As to "scaling", one can note that Waymo also went through periods when we thought fast sustained scaling was inevitable, but Waymo chose instead a more drawn-out process.
If Tesla scales too fast, it will run into issues like Cruise.
What you say makes sense, but I feel a fair amount of uncertainty about it. Waymo has been cautious so far, and it has served them well. But it's possible that Waymo has been too cautious, and that Tesla will be able to scale up more quickly without serious safety incidents. We'll find out in the next couple of years!
Tesla and Elon's entire history has been built on being reckless and technically unserious.
My best guess is that they will try to scale faster with much less preparation, which will end with disaster, and for Tesla that won't be the first or second time people ended up dead but this time they can't blame the driver.
I dunno man, SpaceX has accomplished some incredible stuff, and building a car company from scratch isn't easy either. I agree that there has been a lot of over-promising and smoke-and-mirrors with FSD, but Elon Musk is not a technological lightweight, and I think skeptics should be careful not to confuse "it's going to take longer than he says" (probably true) with "it's never goingto happen" (not a good bet to make).
With that said, I'm definitely worried about him going too hard and it ending in a disaster. As you note, this is what killed Cruise, and I think there's a non-trivial chance something similar happens to Tesla. I'm just not sure.
We know SpaceX has done incredible stuff. This was Elon at his finest in 2015 or so. Young, energic, focused, humble, with a mission.
Now he is taking on the whole world, with plans for humanoids, semiconductor fabs, AI, datacenters in space, unsupervized FSD.
StarLink is actually going well. Starship is very risky, but Bezos has his own problems.
He looks to me like the classical case of hubris. Self-confident, ketamene-fueled, unconstrained by serious advisors, thinks he can bulldoze over everything.
Wishing him well. We need people like him. Hoping for best.
I really wish that Zoox would deploy in some small towns. The conditions are so easy out here. And there's lots of seniors and poor people who could use them. They could in fact operate on a fixed route and do a good service for people, though I suppose it would work a lot better if they had a bigger version.
I guess what I'm really saying is that I wish Olli hadn't gone under and it was deployed in small towns, actually.
There is also May Mobility which seems to be doing something like this in small-sized cities.
Though I'm not sure the conditions are all that much easier than the big cities? There's an irreducible challenge to self-driving that doesn't go away. And with Zoox in particular, having the density of larger cities probably makes it operationally easier when the vehicles inevitably have to be towed from time to time?
I think Texas, Alabama, Mississippi would be great states to test market Tesla cabs. That said, I have a Tesla even though I despise Musk (I know). Autopilot (FSD) has worked for me, but I would not get in a cab with no access to the wheel. At the same time, my bet would be that FSD would womp human drivers for safety. It doesn’t get distracted, fall asleep, or get drunk.
It is nuanced. And if I am right about the safety factor, insurance companies should have to cut the cost of premiums for those who use the self drive feature. My understanding is that Mercedes has a level 3 self driving vehicle being test marketed in Vegas. Mercedes takes on the liability so no insurance necessary for folks who own those cars.
If we can save 10-20k lives a year, it should be mandatory. But, I can hear the right wing cries about liberty as I type that sentence.
Thanks for sharing your post. It’s very interesting and informative.
I've ridden in Waymo, CyberCab, and Robotaxi in Austin and they were all equally flawless experiences. Musk is correct that lidar is unnecessary with a good enough visual system processed by enough compute, and apparently other EV companies with high level AV offerings (like Chinese BYD - illegal in the US) have come to similar conclusions. The FSD in my "old" tesla with older cameras and much less compute is not quite level 4, but it's still level 3+, worth every penny even to a cheapskate like me, and has been life-changing and I would never buy another vehicle without at least as good FSD than that. Older people who find driving increasingly stressful or difficult should be jumping on this tech as quick as they can. I'm not old, but even for me it makes getting around so much more enjoyable and longer road trips much more tolerable.
There is no magic tech Tesla has that Waymo doesn't or did not think about.
Tesla will have to face the same uphill battle to gain experience and win trust, and it is starting from a lot worse position when it comes to reputation for engineering and safety.
Not using lidar is by now foolish, as it is cheap and adds an independent datastream that is critical in borderline situations when cameras degrade.
As to "scaling", one can note that Waymo also went through periods when we thought fast sustained scaling was inevitable, but Waymo chose instead a more drawn-out process.
If Tesla scales too fast, it will run into issues like Cruise.
What you say makes sense, but I feel a fair amount of uncertainty about it. Waymo has been cautious so far, and it has served them well. But it's possible that Waymo has been too cautious, and that Tesla will be able to scale up more quickly without serious safety incidents. We'll find out in the next couple of years!
Tesla and Elon's entire history has been built on being reckless and technically unserious.
My best guess is that they will try to scale faster with much less preparation, which will end with disaster, and for Tesla that won't be the first or second time people ended up dead but this time they can't blame the driver.
Wishing them well but hoping they go slow.
I dunno man, SpaceX has accomplished some incredible stuff, and building a car company from scratch isn't easy either. I agree that there has been a lot of over-promising and smoke-and-mirrors with FSD, but Elon Musk is not a technological lightweight, and I think skeptics should be careful not to confuse "it's going to take longer than he says" (probably true) with "it's never goingto happen" (not a good bet to make).
With that said, I'm definitely worried about him going too hard and it ending in a disaster. As you note, this is what killed Cruise, and I think there's a non-trivial chance something similar happens to Tesla. I'm just not sure.
We know SpaceX has done incredible stuff. This was Elon at his finest in 2015 or so. Young, energic, focused, humble, with a mission.
Now he is taking on the whole world, with plans for humanoids, semiconductor fabs, AI, datacenters in space, unsupervized FSD.
StarLink is actually going well. Starship is very risky, but Bezos has his own problems.
He looks to me like the classical case of hubris. Self-confident, ketamene-fueled, unconstrained by serious advisors, thinks he can bulldoze over everything.
Wishing him well. We need people like him. Hoping for best.
I really wish that Zoox would deploy in some small towns. The conditions are so easy out here. And there's lots of seniors and poor people who could use them. They could in fact operate on a fixed route and do a good service for people, though I suppose it would work a lot better if they had a bigger version.
I guess what I'm really saying is that I wish Olli hadn't gone under and it was deployed in small towns, actually.
There is also May Mobility which seems to be doing something like this in small-sized cities.
Though I'm not sure the conditions are all that much easier than the big cities? There's an irreducible challenge to self-driving that doesn't go away. And with Zoox in particular, having the density of larger cities probably makes it operationally easier when the vehicles inevitably have to be towed from time to time?
I saw a Waymo in Saint Petersburg a few days ago for the first time. Had to do a double take seeing it drive through my parent's neighborhood.
I think Texas, Alabama, Mississippi would be great states to test market Tesla cabs. That said, I have a Tesla even though I despise Musk (I know). Autopilot (FSD) has worked for me, but I would not get in a cab with no access to the wheel. At the same time, my bet would be that FSD would womp human drivers for safety. It doesn’t get distracted, fall asleep, or get drunk.
It is nuanced. And if I am right about the safety factor, insurance companies should have to cut the cost of premiums for those who use the self drive feature. My understanding is that Mercedes has a level 3 self driving vehicle being test marketed in Vegas. Mercedes takes on the liability so no insurance necessary for folks who own those cars.
If we can save 10-20k lives a year, it should be mandatory. But, I can hear the right wing cries about liberty as I type that sentence.
Thanks for sharing your post. It’s very interesting and informative.
Safe travels!
I've ridden in Waymo, CyberCab, and Robotaxi in Austin and they were all equally flawless experiences. Musk is correct that lidar is unnecessary with a good enough visual system processed by enough compute, and apparently other EV companies with high level AV offerings (like Chinese BYD - illegal in the US) have come to similar conclusions. The FSD in my "old" tesla with older cameras and much less compute is not quite level 4, but it's still level 3+, worth every penny even to a cheapskate like me, and has been life-changing and I would never buy another vehicle without at least as good FSD than that. Older people who find driving increasingly stressful or difficult should be jumping on this tech as quick as they can. I'm not old, but even for me it makes getting around so much more enjoyable and longer road trips much more tolerable.