A few weeks ago I took an Uber to a regional airport and was picked up by a Tesla. The driver used FSD, so-called Full Self-Driving, never touching the steering wheel during the entire trip. Should you tip a driver who just sits there? In the end I gave my usual tip and five-star rating. I figured the driver had to be there or the car wouldn't drive and he probably ponied up his own money for the FSD. I give five stars to any driver who gets me from point A to point B without major incident. Remember when five stars meant "above and beyond". Grade inflation gets us all.
But you can't tip Waymos. Last fall during a trip to San Francisco, I took my first Waymo ride. At one point a couple of people jaywalked in front of the Waymo and the Waymo promptly stopped. The car behind us started honking. Who are you honking at? It's a Waymo. It did drop me off at the wrong hotel but that's on me, who thought there would be two Hyatt Regency's in San Francisco.
In both cases the driving felt safer than the average Uber driver. It really is safer. About 40,000 people die in human-driven traffic crashes per year in the United States. But even a single accident could be devasting to a self-driving car company so they are designed to drive on the safer side. I have friends who see self-driving as the ultimate safety upgrade, especially as they lose fast reflexes as they age.
With new technologies (to paraphrase Hemingway), things generally move gradually than suddenly. Waymo and its rivals are expanding into several cities in the US (but not Chicago 😞) and robotaxis are all over China and across Asia. Cities like Shenzhen and Beijing have thousands of robotaxis operating daily, is the US falling behind? And why are we not seeing plans for self-driving cars in the midwest? Is it the weather, or just the usual blind spot technology companies have to this part of the country?
Nevertheless, I suspect we'll see the vast majority of ride-share as robotaxis by the end of the decade and cars without a steering wheel soon after that.
How long before "honking at a Waymo" produces an effect? Just saying.
ReplyDeleteIn my country most rent-an-electric-scooter companies pulled out after years of systematic scooter theft (that's what I've heard at least). I wonder if it will be the same for robotaxis.
ReplyDeleteTesla FSD has had some horrible accidents, but Waymo seems to have had a better safety record overall; I am really concerned about the pace with which Tesla is trying to roll out its robotaxi service.
ReplyDeleteI live in SF, and the most amazing thing about being in a Waymo is how unremarkable it feels -- like being in a car with a friend who is very good at driving: confident, assertive, yet polite. By contrast, my Tesla doesn't even get parking-spot identification reliably right.
Your larger point is spot-on, though: at least in countries with well-established driving rules and norms, we probably won't have human drivers a decade from now. Very, very, unclear when (if?) self-driving will work in India, for example. That might be a PSPACE-complete problem!
@lance, I'd be intrigued to hear which wishful playbook propagates that beijing and shenzhen have robotaxis driving around. Ah, I see, you are probably referring to a company where our fellow friend, Andy Yao is an advisor. But I'm afraid, so far, this is mostly hype. No such thing has rolled out yet -- maybe in the planning, or maybe currently very limited in scope (perhaps around the company's park? or some such limited geographic location. Even though Shenzhen and Beijing are geographically rather large, I've not come across any such robotaxis).
ReplyDeleteActual taxis that are driving around in shenzhen and beijing that millions of people are using daily, though, those are called DiDis part of DiDi, china's response to ueber.
I might have gotten a little ahead of myself but there are reports of a significant robotaxi rollout across China and it's growing quickly.
Delete@lance; thanks for sharing that link. Granted that some people with obvious incentives have hyped about it — a roll out has not happened yet.
ReplyDeleteAnd,yes, even Bloomberg recently featured an article about robotaxis in China — something’s brewing, who profits from such PR?
Nobody I know (unless select people like Andy Yao and colleagues of such companies test drove them in the past) has experienced such a robotaxi in SZ.
Actually Shanghai probably more likely to happen than Beijing for a roll out.
I suspect once green light is given, a burst phenomenon might take changing the current status. And yes about accidents in Didi, lots have been happening (of the smaller kind), particularly during holidays season.
Feeling safe in a robotaxi needs to be re-evaluated in the light of events outlined in the movie “Leave the World Behind” … did Tesla pay to appear in it or did Tesla Sue afterwards ….
But the question is, in a sense city like Shanghai, why should one even use a car?
ReplyDeleteWhy shouldn't almost everyone take high quality public transport?
US is kind of special from cats perspective for transportation, there is no deal alternative. But if I am in Paris or Berlin or Tokyo, for transportation within the city, I would not use a car at all.
And in poor countries, you have so much human cheap labor that economically would it make even sense to pay for much more expensive cars and a self driving service fee that an American tech company would be happy with from profit perspective?
The same way we have robots that can make clothes in textile sector, but it is much much cheaper to have them made by human labor in Vietnam, the cost of automation needs to come down massively for it to make sense in many countries. Don't just look at Uber drivers in SF who probably make above average income in the US, and even there Uber has been essentially subsidizing the rides by massive investor money. They only started making profit as a result of food delivery and significant percentage on the food price.
@lastanon: good question. And agreed on the US status; it makes more sense to take taxis in most cities.
ReplyDeletelooking at ex-US cities, some of us prefer the convenience, comfort and “safety” of taxis vs public transport.
This is all relative of course.
To give an example, the average person wouldn’t want to take public transport during rush hours in Tokyo — begging to be squeezed into the train — often dangerously crowded. And sometimes the metro is so huge u have to walk like 1500 meters to actual station (while you are already underground). But not much else to complain about the impeccable infrastructure.
You wouldn’t want to take public transport in China during hot summer days… AC in metros is not usually working to the fullest, and well, things have not yet reached a public awareness status for other matters.
In Paris the metro system is very archaic compared to Tokyo’s — unsure if y’d want to rely on it. Berlin has undergone too many changes and unsure about it.
So overall it kinda makes sense to have taxis operate; there is obvious demand by the demographics.
But the real question is, in China particularly where the economy is not doing so well and fresh grads can’t find jobs and are resorting to driving Didi … how are robotaxis going to impact or decimate the current sector. Maybe a very good reason to keep postponing the “green light”.
There are approximately 1.27 deaths per 100 million miles driven in 2023 according to the NHTSA, but there is also evidence that the rate is falling to 1.05 in 2025. It's noteworthy that not all miles driven have the same risk, and waymo cars typically drive around at very low speeds in quite good conditions (there is relatively little ice in San Francisco). Tesla reported 2.16 billion miles driven under FSD in 2024. Apparently there have been 59 deaths from FSD so far according to tesladeaths.com. We should beware of the law of small numbers when we assert that self driving cars are safer than human-driven cars.
ReplyDelete