That’s a brilliant idea, I wonder why more places don’t have them?
Today, Waymo and Uber are excited to announce a new, multi-year strategic partnership to make the Waymo Driver available to more people via the Uber platform starting in Phoenix.
Since both companies were founded in 2009, Waymo and Uber have in their own ways each revolutionized access to mobility. Now, we’re partnering to bring together Waymo’s world-leading autonomous driving technology with the massive scale of Uber’s ridesharing and delivery networks.
https://blog.waymo.com/2023/05/waymo-and-uber-partner-to-bring-waymos.html
This looks like it might be very big.
Autonomous vehicle announcements just feel like a continuous chain of hype to keep expectations and capital elevated. Uber lost $9B last year, and this in some sense is where that money comes from.
Is this big? I don’t know. I know if I used Uber eats, I would prefer to have the food delivered to my door rather than having to go down to the street to retrieve it.
On the other hand, Waymo is having real difficulty scaling, so maybe they can scale faster from this.
Recalling that the original “vision” of Uber was self-driving cars from the git-go. With human drivers as their interim, known to be money-losing, stop-gap until self-driving cars came about in, oh, 2013 or so. And Uber in fact had their own major R&D effort towards self-driving cars then. Which they shuttered later.
So now, 14 years after Uber burst on the scene we’ve proven the gig-work app-to-call-a-ride biz is a) a money loser for most drivers, b) a money loser for most app providers, c) enormously popular with users, d) would be even more so if cheaper, and e) self-driving cars are still decades away.
But now somebody other than Uber’s investors are footing the R&D bill.
Move along; nothing to see here.
It’s these things that I have such trouble understanding. It seems to me, there must be a place for an app to make money, and not screw either the drivers or the customers. I know, the enshitification of business means we can’t have nice things, but looking at these numbers, it seems possible.
At a recent event, people were paying about $100-120 for an Uber from the airport to the hotel. That is 43 miles, and about 50 minutes. Using the IRS $0.655/mile rate as the cost of running the car, that leaves $71.83 left for the driver and app provider to split as revenue (the mileage expense will come out of the driver’s gross.)
It seems the driver could, after car expenses, be getting a quite reasonable $25-50/hour, and still leave lots of money for the app to collect, even after credit card fees, and other expenses. Maybe airport runs are easy and profitable trips, but that’s the example I have at hand. I’m sure the 1am, 6 mile home-from-the-bar ride has a different financial breakdown.
I know that the problem is an app that connects passengers and free-lance drivers will eventually reach a saturation point, and even if profitable, won’t be growing the way the investors demand. Enshitification.
Of course, that doesn’t count the time and miles to get drive to the airport to pick people up, or drive from the hotel after dropping them off.
The problem is that Lyft, Uber, Waymo, and others are trying to stand up something that’s incredibly expensive to produce and still unprofitable (AND not functional), while in a sea of competitors and substitutes, both autonomous or not…they’re spending their way into a commodity market. It’s nuts.
This sounds very exciting. Let’s look at the announcement.
Today, Waymo and Uber are excited to announce a new, multi-year strategic partnership to make the Waymo Driver available to more people via the Uber platform starting in Phoenix.
Multiple years could mean decades. When are they starting this thing?
This integration will launch publicly later this year with a set number of Waymo vehicles across Waymo’s newly expanded operating territory in Phoenix, and will include local deliveries and ride-hailing trips.
One is a “set number.” if it’s already set, why don’t they tell us how many?
Waymo and Uber look forward to providing a safe, enjoyable – fully electric and fully autonomous – experience.
I look forward to it too but I’m not holding my breath.
I’m leery of any announcements of things people plan to do in the future. I would be impressed if they said, “Starting today, 100 Waymo cars will be providing services on Uber throughout the Phoenix area. Book now through the Uber app. Here are the limitations…”
Five years. It’s always five years.
That’ll be a big year. Flying cars are coming out at the same time.
A very big year: that is also when Helion will start providing Microsoft with electricity from nuclear fusion [see our nuclear fusion thread].
I recently received the update to Tesla’s Full Self Driving 11.4.2. software, and it is a noticeable improvement over even the 11.4.0 software.
They’ve added training data to help with misjudging the speed of distant objects in order to reduce phantom braking. So far it seems to have helped, as phantom braking incidents seem lower.
I’m not sure what change was made exactly, but the car can now use continuous right turn lanes without coming to a complete stop (this makes Tesla FSD a better driver than 60% of the people in Boulder).
Acceleration is also more confident. FSD doesn’t accelerate very hard, but it used to be slow to about 30, which upset drivers behind. Now it seems to be slow to about 20, and then goes harder to the set speed. The new behavior means keeping up with or leading traffic, rather than holding it up.
I’m finding in the last few days that the main reason I cancel FSD is to handle road hazards—a pothole I need to avoid, or dips and humps in the road that require slowing down.
There are still lots of problems, like signaling for not going into the left turn lane. It also waits much to long too signal for actual turns.
Surely everyone agrees it is very difficult to see a firetruck?
https://jalopnik.com/tesla-confirm-automated-driving-engaged-fatal-crash-185034791
That was a 2014 car, using Mobileye tech. Completely unrelated to the latest versions of either Autopilot or the FSD beta.
It’s the same problem that all radar-enabled cars have: the radar has to ignore stationary objects to reduce clutter from fixed objects (like roadsigns), but a firetruck in the middle of the road is also a stationary object.
It’s always been made very clear that Autopilot requires the driver to pay attention. The car makes a bare-minimum effort to ensure the driver is still holding the wheel, but no more. It’s the driver’s responsibility to look out for things like firetrucks.
If only it had been a tow truck instead of a firetruck, we’d be praising Autopilot for the sweet jump:
It’s early yet on details of this incident, but dogs (and kids) can run out from behind obstacles such that there’s just no preventing a collision and perhaps a fatality.
Considering the location & environs I have to say a dog appearing from behind a parked car there would be very surprising to me as a human driver. Surprising enough that I may not have processed what I saw until it was toast. Or more like creamed chipped dog on toast. 
Unlike in, say, a suburban street with the same speed limit and the same arrangement of parallel parked cars where I might well be actively anticipating kids, dogs, etc., and be totally spring-loaded for an emergency maneuver.
In that sense, the Waymo AV might have done better than a human in the same spot. Sometimes better can’t be good enough.
[This is a non-paywalled msn.com repost]
The big question is how many with an equal amount of usage would have died or crashed if the vehicles had not been using Autopilot. I would guess significantly more.
Reading the links from the WaPo story, I lean a bit the other way — although I haven’t seen the detailed stats needed to have a strong opinion.
Here is a non-paywalled summary of some of what seems wrong:
My wild guess is that a small number of additional people are dying now, but what is learned will save far more lives in the not-too-distant future. Whether the learning curve would be shorter without Elon Musk is another question.
The radar they’re putting in is not the same one that they took out. The traditional radars were totally incapable of dealing with the “firetruck” situation, since due to their low spatial resolution they had to ignore stationary objects. They were arguably worse than useless for self-driving. It remains to be seen if the new radars are better. They do have better spatial resolution, but radar is inherently low-resolution (compared to vision).