Spent last week in Phoenix, rode Waymo a dozen times. Autonomous taxis are the future. Don't have to tip, don't have to worry about pissing off the driver if I'm only going a few blocks. Price is reasonable, seems less than Uber or a standard taxi.
Question is how many humans will forgo owning a car altogether once autonomous vehicles are ubiquitous.
Create a market segment where everything costs more for everyone, "employ" countless people -- usually on restrictive work visas and with a limited understanding of labour laws, rights, and protections -- to be the boots on the ground of the operation, pay those people so little that they drive and ride dangerously in traffic, bike lanes, and on sidewalks to eke out more money out of the system, get people used to paying $40 for a burger, and then just... automate the whole thing away?
This is an ethical no-win scenario for companies like Doordash in my mind, but it's one of their own making. Food delivery as a business catering to the general public needs to go away (with exceptions for meals on wheels-type operations serving the sick and the old who may otherwise not be able to get food on their own).
I live in a city that has had Waymo's (via Uber) for a while now and I have done a complete 180 on them. Not only are they usually cheaper than a traditional Uber, but they drive far more defensively, and don't come with the social baggage associated with a traditional Uber either (tipping, small talk).
Or just live in a non retarded country where neither tipping nor small talk is expected when you take a fucking taxi. Plus it will probably just cost half or less by itself
Interesting trade-offs for a customer. No more expectation to tip or dealing with drivers potentially running scams. On the other side, I assume you now need to go unload the delivery from the car yourself, a much worse experience for apartment dwellers or the disabled.
Either way, we're going to see a lot more of this. More and more of the gig economy being automated away.
Not surprised. Phoenix is one of those rare and contrived physical locations that doesn't get weather or seasons. No snow, little rain. The road surface and road edges are always visible and never change. People always stay in the clearly visible lanes. It's the perfect place to field semi-autonomous vehicles that can't hack it in normal regions so it looks like they're more capable than they are.
If Waymo were launching in Minneapolis I'd be surprised and delighted. But this is just more of the same.
It's definitely easier to drive than in many places, but "doesn't get weather" is not quite right when there are significant dust storms every year. But few people even know what haboob means, so I guess they don't know how bad the loss of visibility is during one.
DoorDash appears very inexpensive when you have their DashPass product, however ive notice that basically every food service business will raise their menu prices, and grocery stores will restrict which items that allow you to buy.
This is really interesting because if you have autonmous drivers, DoorDash doesnt really have a lever to lower prices except removing tipping.
It would be interesting to see if this could be combined with those little sidewalk bots to do the last mile, effectively having the Waymos act like buses for the bots.
Food delivery is something I truly have never understood. I have very rarely been in a situation where I was thinking about food and couldn't think of any nearby restaurants within walking distance (~30 minutes on foot). Why would I order if I could just walk, which is also more healthy anyway? Even if I was extremely busy, if I have time to eat I also have the time to get the food.
This is a great step in increasing the utilization of Waymo vehicles. Ideally Waymo's would be operating continuously doing useful work and reducing the number of dead-head legs.
We're still moving thousands of pounds of vehicle around a public highway to carry a 1lb burrito, obviously lightweight aerial drones are the future for food and grocery delivery.
I think we should go back to Pneumatic tubes. I wonder if justifiable size could be made. But still after infra is build we could move lot of food and package delivery to the system.
If one keeps going down the rabbit hole, one might infer that the way our cities are designed is entirely wrong but probably with 60-70 year of invested capital difficult to change fast enough
Or small autonomous carts that drive on the bicycle and pedestrian roads. Safest, simplest, easiest to implement. The biggest hurdle is not technology, but politics and infrastructure.
Are aerial drones really any better in terms of energy expended? Not to mention the massive amount of noise, which has already shut down several attempts at drone delivery.
Currently it cost DoorDash an Ubereats 2 to 3 dollars to send food to the front door of someone’s house. It takes about 20 to 30 minutes for that driver to form all the steps to do the delivery. This works out to $6 dollars per hour of driver time. The driver is paying insurance, gas, time, depreciation insurance, accident repairs, maintenance, and everything. DoorDash pays $144 per day for approximately 48 deliveries with a human driver.
Now let’s try to assess the automated car situation. Assuming each delivery is about 5 miles that works out to 240 miles per day for 48 deliveries. Most cars are useful for 100,000 miles so the vehicle should be able to deliver for about 416 days. Assuming it gets gas mileage of 30 miles per gallon and gas costs $4 per gallon (using gas for simplicity though these are probably electric) for $13,333 in fuel over the life of the vehicle. Maintenance for these vehicles will vary of course, but a reasonable estimate is $1000 per year for brakes, oil changes, etc. adding up to $1140 total over the life of the vehicle. There are other costs that will be required as well like parking for the car when it’s not in use, cleaning of the car outside and inside, software maintenance, etc which I am unable to estimate, but it won’t matter as you’ll see below.
Automated cars are likely to cost at least 60k each (being really generous here … see below) given current prices on cars.
Cost of vehicle - $60,000 / $200,000
Vehicle Maintenance – $1140
Fuel – $13,333
Insurance - $1000
Other costs???
Total automated driver - $75,473 / $215,473
* Found article that states Waymo vehicles cost $200,000 as of June 2025, but included the scenario where the cost of the car is $60,000 and human drivers are still less expensive. So even if the Waymo vehicles dropped to 1/3 the current cost which is not likely, they are still more expensive than human drivers.
“Waymo vehicles are equipped with numerous expensive sensors and can cost roughly $200,000, enough to buy five or six regular cars. As of May, there were just 1,500 Waymos operating in all its markets.”
Rough napkin calculations show that it’s not cheaper for the company to buy some brand new, super high-tech automobile that is unproven and requires tons of research and development to refine it to the point that it can’t even complete the complete task (pick up food at counter and delivering food to the door of the customer).
I live in a big city. DoorDash always delivers to an address 10 minutes away walking from my building. It is quite inconvenient in the winter. With human drivers, you can at least try to convince them to use Google Maps, but with an AI?
On second thought, prompt injection via delivery instructions?
Interesting that there's no mention in TFA or the comments about DoorDash's ground-based delivery "drones": https://about.doordash.com/en-us/news/doordash-unveils-dot -- I've seen them testing these things in parts of the Bay Area for more than a year now.
I think they're intended to use sidewalks and bike lanes, so should address concerns about cars clogging up streets.
My hunch is that just because they're smaller and lighter does not mean it's an easier problem to solve than a self-driving car. A more interesting partnership between Waymo and DoorDash would be licensing a scaled-down version of the Waymo tech for these things.
64 comments
[ 4.3 ms ] story [ 64.7 ms ] threadQuestion is how many humans will forgo owning a car altogether once autonomous vehicles are ubiquitous.
> Me, in 2015: [Uber's price] is reasonable, seems less than a standard taxi.
We already know how this story progresses.
Obviously you can already get delivery from Whole Foods, FreshDirect, etc., but it's expensive due to the drivers.
And public transport and bikeshares are great for transporting you, but not for trying to carry four or six bags of groceries along with you.
Assuming it's those little cooler-sized ones
Damn it's an entire car for a package? hmm maybe they combine people and food (points to head)
That'll be a new traveling salesman algorithm, the waymo doordash problem
Create a market segment where everything costs more for everyone, "employ" countless people -- usually on restrictive work visas and with a limited understanding of labour laws, rights, and protections -- to be the boots on the ground of the operation, pay those people so little that they drive and ride dangerously in traffic, bike lanes, and on sidewalks to eke out more money out of the system, get people used to paying $40 for a burger, and then just... automate the whole thing away?
This is an ethical no-win scenario for companies like Doordash in my mind, but it's one of their own making. Food delivery as a business catering to the general public needs to go away (with exceptions for meals on wheels-type operations serving the sick and the old who may otherwise not be able to get food on their own).
Either way, we're going to see a lot more of this. More and more of the gig economy being automated away.
If Waymo were launching in Minneapolis I'd be surprised and delighted. But this is just more of the same.
Obviously the cars can drive themselves on public streets, but how do you go up to someone's house and put a burger on their doorstep?
This is really interesting because if you have autonmous drivers, DoorDash doesnt really have a lever to lower prices except removing tipping.
Food delivery is something I truly have never understood. I have very rarely been in a situation where I was thinking about food and couldn't think of any nearby restaurants within walking distance (~30 minutes on foot). Why would I order if I could just walk, which is also more healthy anyway? Even if I was extremely busy, if I have time to eat I also have the time to get the food.
Absolutely untrue.
People should be given enough time to make and eat healthy food, or dine-in at a restaurant if they want to eat restaurant food.
Boxed up food tastes bad, and is largely a solution for overworking people.
cheap enough to deployed in the thousands. small enough to take up less than 1 sq / ft of sidewalk space.
We live in a very dumb era.
Now let’s try to assess the automated car situation. Assuming each delivery is about 5 miles that works out to 240 miles per day for 48 deliveries. Most cars are useful for 100,000 miles so the vehicle should be able to deliver for about 416 days. Assuming it gets gas mileage of 30 miles per gallon and gas costs $4 per gallon (using gas for simplicity though these are probably electric) for $13,333 in fuel over the life of the vehicle. Maintenance for these vehicles will vary of course, but a reasonable estimate is $1000 per year for brakes, oil changes, etc. adding up to $1140 total over the life of the vehicle. There are other costs that will be required as well like parking for the car when it’s not in use, cleaning of the car outside and inside, software maintenance, etc which I am unable to estimate, but it won’t matter as you’ll see below.
Automated cars are likely to cost at least 60k each (being really generous here … see below) given current prices on cars.
Cost of vehicle - $60,000 / $200,000 Vehicle Maintenance – $1140 Fuel – $13,333 Insurance - $1000 Other costs??? Total automated driver - $75,473 / $215,473
* Found article that states Waymo vehicles cost $200,000 as of June 2025, but included the scenario where the cost of the car is $60,000 and human drivers are still less expensive. So even if the Waymo vehicles dropped to 1/3 the current cost which is not likely, they are still more expensive than human drivers.
“Waymo vehicles are equipped with numerous expensive sensors and can cost roughly $200,000, enough to buy five or six regular cars. As of May, there were just 1,500 Waymos operating in all its markets.”
https://sherwood.news/tech/as-the-race-for-autonomy-heats-up...
Total Human driver - 416 days x $144 = $59,904
Rough napkin calculations show that it’s not cheaper for the company to buy some brand new, super high-tech automobile that is unproven and requires tons of research and development to refine it to the point that it can’t even complete the complete task (pick up food at counter and delivering food to the door of the customer).
On second thought, prompt injection via delivery instructions?
https://youtu.be/040ejWnFkj0?si=buVxsUBX4tIRmxDS&t=858
I think they're intended to use sidewalks and bike lanes, so should address concerns about cars clogging up streets.
My hunch is that just because they're smaller and lighter does not mean it's an easier problem to solve than a self-driving car. A more interesting partnership between Waymo and DoorDash would be licensing a scaled-down version of the Waymo tech for these things.
If I walked or biked in these places, I'd be worried. Sharing the sidewalk with robotic wheelie-bins seems like a bad idea.