I’m sorry you feel that way. You started off the thread expressing your desire, to GoogleAI, to understand why some people disagree with your opinions about Tesla’s future prospects. I thought you genuinely wanted to understand the contrary view. If you’d rather not be told what the actual issues are, and instead believe that Teslas cars are $150K cheaper than Waymo’s, that’s your prerogative- but then participating on a message board is going to be very frustrating for you, I think.
1 Like
Was posted 8 days ago:
2 Likes
I went through this on the @Home board many years ago. The True Believers® wouldn’t hear of any possibility that the company wouldn’t be an astonishing success, after all, the stock had gone from zero to 60 in just a year or two, and Milo Medin was an internet genius!
When it was pointed out that the cable MSOs have never let someone come between them and the customer, or that the overbuilt pipes would obviate the need for a “local internet hub” (as was the @Home architecture at the time) or any of the other many and various possible issues, it was fingers-in-the-ears “can’t hear you” over and over, until the stock collapsed and went to zero, of course. Heck TMF even hired one of the cheerleaders (TMFNole) to be the “resident expert.” OK, then.
Some people don’t want to hear what they dont want to hear, and that’s that.
Meanwhile: Robotaxi can’t drive at night (I believe the taxicab business does a fair amount of miles after sundown) and Elon Musk himself explains why: “No lidar”. You see, “reflected headlight” is not the same as “full daylight” or even “cloudy day”, and the Robotaxi has a hard time distinguishing things without sufficient light to provide contrast.
To lidar, of course, the visible light spectrum is irrelevant; it sees what its radar blips see, day, night, or whenever. But Robotaxi, by decree: no lidar. This puts Elon in a curious position, having admitted that “vision only” is not safe, yet running the Robotaxi until 11pm (it gets dark around 7:00 in Houston). And, of corse, the curious position of “we don’t need lidar” but we’re going to run the machine when I’ve already admitted it’s not safe at certain hours.
I would just note that I hope Tesla is successful with the Robotaxi. My animus for Musk personally does not mean I don’t hope for successful products which make the world a better place. I didn’t Iike Bill Gates, either, but I wished no ill on Microsoft (other than the penalties for foisting monopoly power on lesser competitors)
2 Likes
Hyundai is near the start of production for Waymo’s next vehicle.
In the coming weeks, Hyundai begins production of robotaxis for Alphabet’s Waymo
Regarding Lidar costs,
It turns out we do have an idea of unit manufacturing costs.
Just listen to the manufacturers.
Here’s Rivian on the cost of Lidar:
“You’re talking on the order of a few hundred dollars versus if you went back several years ago where it was thousands or even tens of thousands of dollars within the vehicle,”
1 Like
Of course if Elektrek Fred is correct, Elon has publicly said that Robotaxis are having trouble seeing pets at night (or something like that). And supposedly that is why they have been limiting drives at night even though they also supposedly have increased availibility now by 1 hour. This whole anti-LiDAR stance by Elon has probably cost them a good amount of the time they are behind Waymo and others now scaling
I dont have a link handy but it can easily be found on Elektrek. And I doubt the issue is just “pets” - it seems FSD has trouble in a lot of cases due to the passive nature of vision only.
And for anyone that pays attention to the crowd sourced FSDTracker, the versions released to the public still show very low miles between minor disconnect incidents and a very clear basis in the data to infer that they still can not get more than one or two "9"s when it comes to the more serious March of 9s.
Meanwhile, the cost of LiDAR has gotten way low for an automotive short range LiDAR and only a little more for long range. I have investments in multiple LiDAR companies and the ones doing well make up for the ones that aren’t.
2 Likes
As do I. It’s usually better if new products succeed rather than fail.
That said, I don’t think that robotaxis are going to be much of an economic success for Tesla or for Waymo or Zoox or any of the other competitors. I don’t think they’ll be able to improve much, if at all, on Uber’s economics.
Uber’s secret sauce - the magic behind the model - is that it monetizes existing but unused capital. You had all these cars, distributed throughout a metropolitan area, that were mostly being unused throughout the day. They already existed. All that needed to be done to get them generating revenue was for them to be put into service at times they were not.
That gives Uber some enormous advantages. There virtually are no marginal capital costs to have the fleet. The cars are already there. The parking is already there. People are already refueling and maintaining and cleaning the cars. Etc.
Moreover, the active fleet is always the right size (in mature metros), because cars can come in and out of service. If demand is high, you can pull more cars in; if demand is low, they just don’t run.
Waymo and Zoox and Tesla don’t have these advantages. They have capex to pay for cars to be in service. They have to pay for places for the cars to be stored. They have to pay for them to be cleaned and fueled (both overnight and during the day). And their fleet will often be mis-sized for the demand at any given time. During low points, they’ll have cars sitting idle on lots on their dime; during high points, they’ll be leaving money on the table because they don’t have enough cars.
I don’t think those costs will rise to the roughly $0.80 per mile that Uber drivers get paid to drive, as opposed to covering the operating costs of the vehicle. But they’re not nothing - probably about half of the savings from losing the driver will get eaten up by the capital costs of vehicles and storage. Then some of the rest will get eaten up by labor that Waymo and Zoox and Tesla have to provide that drivers provide as part of their service in Uber: cleaning the vehicles during the day, handling lost items, handling unusual driving circumstances (though that disappears as the AV driver gets better), etc.
Rough order of magnitude, I’d guess that a fully-functional Level 4 autonomous service can deliver rides for maybe 10-15% less than Uber. Around $2.50 to $3.00 per fare. That ain’t nothing - but it ain’t changing the world, either. Especially since the autonomous services will necessarily have longer wait times, particularly when most of the rides happen.
1 Like
The pure dollar economics don’t interest me much (of course they certainly matter).
I am interested in technology that makes driving safer.
Driving is, by far, the most dangerous activity for me and most people, certainly by (hours at risk)x(risk of damage to life and property per hour).
I would like to see this technology make driving safer (recognizing that driving already has a history of getting safer over time).
The technology can also aid less-able drivers, the elderly and the disabled.
I agree with this completely. But I am also a large shareholder. I want FSD to make money (which I believe it already is). But it is still basically way overpriced as an ADS. I want the Robotaxis to be a viable revenue source too but they seem still far from it meeting basic safety criteria.
2 Likes
They’re deeply related, of course. The technology won’t save lives unless it’s adopted, and it will only get adopted to the extent that it’s economical to adopt it.
Uber’s a case in point. It costs about $2 per mile to produce an Uber trip (ballpark). But it costs about $0.40-0.50 per mile to self-drive your own vehicle (not AAA new-vehicle numbers - the average vehicle’s a decade old). Uber has had more penetration in the market than traditional taxicabs, but only to a point, because it’s just much more expensive. Even though Uber drivers are somewhat safer than the “average” driver (they are not teens, are generally not elderly, typically have a lot of driving experience, are somewhat filtered by Uber, are less likely to be driving after having recreationally drank or taken drugs), it doesn’t have much of an effect on overall accident rates - because rideshare and taxi are only about 2.5% of total miles.
If the cost of a Level 4 robotaxi service ends up landing north of $1.50 per mile, you see some modest expansion in adoption beyond where rideshare is today. But probably not enough to materially affect driving safety.
An Uber human driver with a Tesla (and FSD) or any other car with supervised self driving would both make the car safer AND have (approx) the same economics of an Uber today. They could advertise it as safer and charge an extra $0.25 or $1 (or whatever) to cover a portion of the monthly subscription fee
Mike
The above is a low return Decreasing Returns businesses model. The leap forward is getting rid of the human driver which increases the usage of the vehicle by, at least, two or three times unless you have two or three drivers per car. Uber was a good businesses model while taxis still required a driver.
The Captain
Almost 60% of Uber/self-driving market is 18-34, where “safety” is a far lower priority than “convenience” or especially “price”. It isn’t until later in life that safety becomes a more salient want. (Michelin Tires, for instance, trading on the “safety” theme, find its customers are almost entirely middle aged homeowners who value that safety over cost.)
Except that doesn’t really change the business model much. It doesn’t make much difference if the car is used two or three times more often. Almost all of the expense of the car is operational - it costs a certain amount to drive it per mile, and the vehicle wears out over distance as it’s driven. The rate at which that happens has only a modest economic effect - if the car has a useful life of X total miles and costs $Y per mile to drive, it doesn’t matter too much whether you drive it that X miles in 20 years or 5. It matters a little, to be sure - but that advantage gets undercut by the deadheading in a taxi model.
No reason to think robotaxis will be any less of a “low return Decreasing Returns” business as Uber is because the cars are driven more often.
The other major effect of getting rid of the driver is, of course, ostensibly lowering the cost to operate the taxi. The driver is perhaps the largest cost of running Uber. But as we’ve also discussed, taking the driver out of the car doesn’t save you the full cost of the driver - you pick up labor costs for people to maintain the infrastructure, you end up having to pay the capital costs for the cars and their storage and cleaning/charging services and what not, and you lose the inherent “right-sizing” of the fleet. Running a robotaxi fleet will probably be a little cheaper than Uber - but probably not by much, especially in a higher interest rate environment.
The real game-changing business would have been if they had been able to get the Tesla Network to work. But as we’ve discussed, the very existence of the Cybercab means that even Tesla doesn’t think that’s going to happen in any relevant time frame.
3 Likes