Waymo now open to everyone in San Francisco | Hacker News Reader

Waymo now open to everyone in San Francisco

twitter.com

106 pointsalphabetting33 comments

I have to admit I was really doubtful we'd have fully self driving cars within the next 50 years in any major city, but here we are. I guess I watched too many Tesla fail videos. Bravo Waymo!

Very interested to see how this goes, especially in SF. SF is a city filled to the brim with crazy people, homeless, criminals, drug addicts, and a hell of a lot of people who hate tech for what it has done to the city's cost of living. And these groups have lots of overlap. My intuition is that giving everyone in the city access to highly expensive toys from the tech industry with no human oversight is a recipe for disgusting, broken vehicles that nobody will want to ride in, but I'd love to be proven wrong. Hopefully those geniuses over at Waymo have already thought of a solution to this.

Unfortunately it will take a lot more than Waymo to solve the problems you mentioned.

It will take enforcing laws, arresting criminals, and offering drug/mental health treatment as an alternative to jail, while at the same time providing housing conditional on employment and good behavior. Nothing short of this will solve our crime, addiction, and homelessness problems in a humane way.

Cruise has been running this for a while now, and nothing like that has happened. SF craziness is mostly limited to Tenderloin

They have a bunch of driverless cars driving around at night, but they've so far only been serving a handful of rides per day, and there's a waitlist to get access. So it's not really open to crazy drugged out criminal anti-tech folks OP was talking about yet. That said, I don't expect this to be a big problem given the cars are loaded with cameras. There will be people who create mischief but they'll quickly be banned from the service.

Question for those familiar: what do these driverless cabs do between rides? Do they just park at the closest free/low cost spot or do they keep driving around? Or are there dedicated spots rented out by Waymo where the cars dock?

They meet in secret locations and do illegal races and burnouts at intersections.

No source, but I was walking in SF a few months ago and passed a small parking lot with only Waymo cars parked, so likely they have locations scattered around.

The app is still showing me a "sign up to be part of our testing program" screen, and I'm currently in the city of San Francisco.

The HN headline is slightly hyperbolic. The tweet says they are approved for service but will work through the waitlist in coming weeks before they make it fully public.

It doesn't seem like the situation is any different than with Cruise. Cruise received this same license June 2 (https://www.cpuc.ca.gov/news-and-updates/all-news/cpuc-issue...), and I'm still waitlisted. Maybe Waymo will expand their operations faster, but I'm skeptical.

Só finally we get bloody self driving cars!

Glad to see Waymo pulled it off after so many years.

Hopefully if this SF test works they will expand rapidly.

It's very difficult for me to see how Waymo could actually become a profitable company. Trying to be an actual taxi company (using custom low volume, high cost cars) seems capital intensive and low profit-margin (you're competing against every other form of transportation, along with uber and lyft with their low-paid employees that provide their own cars), and it seems like the number of major car companies that could license their software is rapidly falling (Tesla and GM definitely have their own in-house competitors, and other major car companies probably do as well). So how do they actually make money? They're billions of dollars into this project, and it just seems like they've missed their market window.

> It's very difficult for me to see how Waymo could actually become a profitable company.

Patent licensing or military contracts. Almost certainly not from running a taxi service.

maybe the taxi service is just a way to gather more data, a stepping stone to a bigger market? I agree though, licensing it doesn't seem like it'll be enough, and idk what else they could be going for. Maybe last-mile delivery? Maybe they think conquering self-driving will help unlock AGI? or maybe they don't really have a plan for monetization yet

In the short to medium term they will keep bleeding cash, but in the long term as they scale, hardware and other costs will go down considerably. They will not only replace taxis but increase them by some factor, 10x? 100x? Who knows where the limit is when hardware, electricity and cleaning are the only marginal costs.

That said I think self-driving ADAS subscriptions for personal vehicles is and will generate most of the profit in this space. By any measure Tesla's FSD is already profitable to the tune of $1+ billion a year, and it's not even widely released yet.

> and it seems like the number of major car companies that could license their software is rapidly falling (Tesla and GM definitely have their own in-house competitors, and other major car companies probably do as well). So how do they actually make money? They're billions of dollars into this project, and it just seems like they've missed their market window.

I think this argument answers itself.

If you’re the only company with this technology, you get to license it to everyone. If you are the tenth company, you will have a hard time finding licensees, so why bother spending the billions?

If it really costs so much to develop, the equilibrium is for only a few entities to make that investment and sell it to everyone else.

If a car's life is 150,000 miles, averages 30 mph in the city, that's 5,000 hours of life. If the driver expects to make $20 per hour, you just have to keep the hardware to under $100,000. Even dividing that by 2 to account for unknowns and work the driver would otherwise do, there's a clear path to profitability once the volumes increase.

EVs are expected to have a longer life than ICE, upwards of a million miles. That's why robotaxis and EVs are mutually reinforcing technologies, both together has the potential to be extremely profitable.

And this isn't even considering the obvious revenue stream opportunities an Alphabet company is ideally positioned to leverage from such a captive-on-our-conveyor audience.

My regular ICE engine car has 200k miles on it and it still running fine. I don't know why this cars won't get triple that, and they don't have drivers ($20 an hour not necessary)

But take Uber and Lyft for example - their drivers probably net well under $20/hr, and more importantly, both of those companies have been bleeding money for their entire existence. Now imagine we’re in a world where lots of people are actually willing to pay for rides in self-driving cars, and there is profit to be made - Tesla and GM and all the other car companies with lots of experience making millions of cars per year reliably and cheaply, and with their own self-driving capabilities in-house, are now competing with Waymo. Waymo doesn’t seem like they have a big competitive advantage in self-driving technology, car manufacturing or general transportation services - so where are their billions in future profit going to come from?

> But take Uber and Lyft for example - their drivers probably net well under $20/hr

I agree, but this isn't sustainable. You can only depend on people screwing up the math on their net pay for so long.

> both of those companies have been bleeding money for their entire existence

They subsidized rides, so that inflated the market size. Waymo probably needs that larger market size to get enough volume, but I'm not actually sure.

> Waymo doesn’t seem like they have a big competitive advantage in self-driving technology

Hard disagree here. It feels like when Apple launched the first iPhone, and someone demoed a feature phone with a resistive touchscreen and said "no competitive advantage."

You really feel like they have a huge competitive advantage over GM/Cruise and Tesla, who can and do produce millions of cars a year, and can readily monetize their 'imperfect-but-rapidly-improving' self-driving car technology?

They do seem to have an advantage in self driving technology not manufacturing. Which is fine because they're not trying to be a mass manufacturer. Manufacturing cars is a different problem they aren't interested in, but I think self driving require a lot of tech that they have a natural advantage in as part of Alphabet (besides having more money available to them and having been working on it for much longer). For example the Dojo stuff that Tesla seems to think is useful. They claimed to be almost finished with that years ago and still don't seem to be using it. In the meantime TPUs have gone through two new generations that Waymo can use en masse.

Besides that generally Waymo is a much bigger organization with a much bigger budget than the part of Tesla that is working on FSD

> It's very difficult for me to see how Waymo could actually become a profitable company.

If the autonomy is consistently safe and reliable, profitability will be trivial. I don't think you've tried very hard.

The average car sits around doing nothing except depreciate in value > 90% of the time. We spend enormous amounts on our cars. The opportunity frontier here is huge. This doesn't solve society's demand for mobility at peak travel times, but I could see "Better Lyft / Uber without labor costs" as a viable business very easily.

i got to try it out in arizona and it worked pretty well. it was a nice day and light traffic so will be interesting to see how well it does in a congested city. but i can see waymo being able to expand to several cities a year. also checked uber and waymo was ~$1-$2 less for a ~$10 ride.

Can anyone comment on how much this costs? Is it on par with Uber / Lyft? Is it cheaper?

I've only rode two times in Arizona but it was probably 25-35% cheaper than a comparable Uber.

Yeah, Uber is toast. So is Lyft et al.

I think any prices anyone has seen so far are pretty arbitrary and have no relationship to the 'cost' of actually providing that ride. The price is that cheap because if it wasn't 25-35% cheaper than Uber, why would anyone test it for them? They've spent billions of dollars so far and probably have a burn rate of billions more each year, and that's without doing any kind of future capital-intensive build out of cars and maintenance infrastructure.

is waymo dealing with lane insertion in a consistently safe way or does it stumble sometimes like i've seen tesla car do on youtube ?

Reply on news.ycombinator.com