They’re solving different problems. Waymo works in only some areas and requires all sorts of hardware. Tesla FSD is trying to do something more generalized.
No, Tesla is trying to create human like FSD. As in without relying on pre scanned pre vetted areas, and with (ideally) just two vision sensors like a human.
No, but I don’t think its clear that Tesla deserves its valuation at its current rate. They aren’t demonstrating any competitive advantage or particular skill in the space vs Waymo
All that means is people are paying too much for their shares in the business on the market, it doesn't mean Tesla don't have the financial or technical ability to deliver self driving cars
Comma is phenomenal at highway driving. Better than Tesla.
They should keep on doing what they're doing. Make ~$1000 gadget that makes driving less of a chore. Doubtful they'll get to driverless in next 10 years.
There can be many players in the market at different offerings and price points. Comma is a sweet spot where they are handsomely profitable, don't need to raise another round and get incrementally better.
Waymo has unlimited money from Google. Cruise, Zoox, e.t.c will have to face reckoning at some point.
It was smart of Uber to get out of self-driving car billion dollar money burning pit. They can always enter the race again. They have market size advantage.
I can't fathom how "makes driving less of a chore" can work from a regulatory standpoint.
For me, that's the end-all-be-all of self driving promises. It doesn't matter whether they call it level 3, 4, 5, or 420, what the branding and promo copy says, it matters how it gets handled from a legal standpoint.
Unless you can get the vendor to accept liability, you're either hands-on-the-wheel driving, or sitting there tensing and waiting for the "you must jump back to hands-on-the-wheel driving in the next 3 seconds or we plow into the side of a lorry" alarm.
In a way, this is the same heuristic as "the company that offers a long warranty can't make crap." No company will willingly put themselves on the hook from the estate of pedestrians and other drivers until they know they've solved it to a statistically high level.
I was deeply unimpressed with Tesla’s highway driving when I test drove one. I found both Ford and Hyundai/Kia to be at or above the level of the Tesla.
Frankly, I didn’t see any benefit of Tesla over most of the lane centering and adaptive cruise solutions in the market. It felt so unpredictable that I preferred to simply drive by hand.
Waymo has been doing public rides (2015) since before either company was founded (2016). Pony.ai also got their permit suspended twice in California for different safety-related reasons, so the comparison says more about the different regulatory environment in China than it does about their relative maturity.
Yeah. Have you seen Waymo? Even with 12.5, you have to keep your eyes on the road, and at least once or twice a week, it’s going to do something that would likely cause an accident. If it does, Tesla will blame you for not paying attention.
In a Waymo, you sit in the back seat and scroll your phone and it gets you safely to your destination every time. These are not even remotely comparable products.
He obviously hasn't. The tech community is in for quite a shock with Tesla robotaxi's come out in a year or so and blow out the competition. One factor not discussed often is how much smoother FSD is compared to all other self driving. Especially noticeable when going from FSD to the free AutoSteer or GM Supercruise.
> The tech community is in for quite a shock with Tesla robotaxi's come out in a year or so
FSD has been set to come out in a year or so since 2014. At this point nobody should be foolish enough to give Tesla the benefit of the doubt; if they have a self-driving car, then they'll take legal liability for its actions, and if they won't take legal liability for its actions, they don't have a self-driving car. This is as true for any company as it is for Tesla, although Tesla is the only one who's seen fit to lie about it for a decade.
Just watched the first video that shows up on YouTube when searching "12.5.1.3" and it starts with a small compilation of the driver having to take back control before the car hits things.
> tech community is in for quite a shock with Tesla robotaxi's come out in a year or so and blow out the competition
12.5.1.3 is incredible. But it's still Level 3, attempting to do something stupid at least on every trip I've been in one.
I think Tesla will figure out Level 4 in the near future. What makes the "year or so" unrealistic is it will have to convince regulators that it's safe. For a variety of reasons, partly legitimate and partly due to Musk's antics, that might be difficult in a lot of the most lucrative markets (and markets where Teslas are).
Top comment's estimate for 2030 parity with Waymo et al is more realistic.
I test drove a Tesla last fall and was deeply unimpressed with it.
It did a few things better than my Ford and Hyundai vehicles, but it was wildly unpredictable when it’d just decide to stop working. That’s a big problem for me.
My Ubers often smell very strongly of the food the driver has been eating.
The driver is often on a call for the entire journey, which is annoying.
It’s often much colder or warmer than I’d ideally like.
An extreme case, but I once had a driver stop a ride, jump out of the car and pull the driver out of the car in front. They proceeded to have a fist fight. The other guy had honked at him for blocking the road or something.
After my first home run in business, which was widely publicized, I learned that my email and phone number were tainted. On craigslist I started getting worse deals, seems everyone seems to google everyone.
I rotated identities, just email, phone number, assumed a different last name in casual arenas, and everything went back to normal as a nobody.
Subsequently, the most amusing thing became reading about women googling all their potential dates. I never do that, but its funny that I’m a ghost and makes me wonder how many other ghosts they're giving the green check mark on. Seems like a waste of energy.
maybe this is different in SF, but NYC taxi/uber drivers never say anything besides confirming you are the actual person and the destination, and maybe asking where you want to be dropped off
I know it is anecdotal, but I am just glad you didn’t get to experience the Uber ride I took from JFK to Dumbo earlier this year. It was around 1.5 hours of the driver trying to talk to me about Putin actually being a 5D chess mastermind and other conspiracy tier garbage after finding out that I was russian. Yeah, thanks, I will take Waymo to avoid listening to that or deal with the awkwardness of having to ask the driver to stop talking (which I didn’t do because it felt rude).
And while I agree that drivers here seem to be less inclined to talk than on west coast, I still occasionally get drivers who facetime the entire ride on the loudspeaker or watch videos on their phone mounted on the dashboard (while barely paying attention to the road).
Not even mentioning some of the stories my friends who are women told me about their uber/lyft experiences.
That reminds me of when I used to travel alot and conspiratards would tell me about the shadow groups that ran the world, but it would always be local ones
That was hilarious. Instead of Rothschilds, or Blackrock, its a French fraternal order in Paris.
those kind of people never notice the flaws in their theory, such as there being so many shadow groups that they would have to compete for power, the same as if no shadow group existed
> those kind of people never notice the flaws in their theory, such as there being so many shadow groups that they would have to compete for power, the same as if no shadow group existed
There is a simple explanation for this within their canon, similar to how it works with religions - my conspiracy theory is the real one, the others are all wrong and/or planted by the baddies of my conspiracy theory.
There are sometimes fun crossover episodes between different conspiracy theories. But as soon as some incongruency between them forms a flaw in logic that you pointed out, it instantly goes to “yup, they are wrong, and my conspiracy is the right one.”
I like 2.) and 3.) and this was a promise of uber and lyft originally, but instead its always been a cumbersome experience that might as well not exist
Waymo’s data was derived from crashes reported under NHTSA’s Standing General Order (SGO), over 7.14 million fully autonomous miles driven 24/7 through the end of October 2023 across Phoenix, San Francisco, and Los Angeles. That data was then compared to relevant human crash rates resulting in police reports, injuries, and/or property damage.
When considering all locations together, compared to the human benchmarks, the Waymo Driver demonstrated:
An 85% reduction or 6.8 times lower crash rate involving any injury, from minor to severe and fatal cases (0.41 incidence per million miles for the Waymo Driver vs 2.78 for the human benchmark)
A 57% reduction or 2.3 times lower police-reported crash rate (2.1 incidence per million miles for the Waymo Driver vs. 4.85 for the human benchmark)
This means that over the 7.1 million miles Waymo drove, there were an estimated 17 fewer injuries and 20 fewer police-reported crashes compared to if human drivers with the benchmark crash rate would have driven the same distance in the areas we operate.
> This means that over the 7.1 million miles Waymo drove, there were an estimated 17 fewer injuries and 20 fewer police-reported crashes compared to if human drivers with the benchmark crash rate would have driven the same distance in the areas we operate.
These numbers are not as high as I thought they would be.
Haven’t jaguars always been a mixed bag? Kind of low quality luxury, so I can imagine them getting some kind of deal. The irony is that jaguar used to be know for its really bad electrical wiring back in the 70s.
Freeways are much, much less forgiving of abrupt speed changes and braking, which is something waymo (used to) have quite an issue with. Moving to freeways shows they are confident that won't be an ongoing issue.
Yes I was driving on a freeway last week and traffic went from 80mph to dead stop about as fast as I could stand on the brakes. I didn't hit the car in front of me with only a few feet to spare, and fortunately the driver behind me was also paying attention. The jam eventually cleared, and there was absolutely no indication of what caused it.
>> and there was absolutely no indication of what caused it
That's the Accordion Effect.
With enough cars on the road, one little tap of someone's brakes flashes their brake lights and it ripples upstream until the 80mph column of cars is forced to go to down to zero ASAP.
Emergency braking is much harder at freeway speeds.
At 35mph you can have something (radar/cameras) look a few meters ahead, then if there is a stationary obstacle you slam on the brakes.
At 60 that doesn’t work because braking distances are much longer. There might be an obstacle directly ahead of you on the pavement, but you won’t hit it because the car will turn with the road. This means that your emergency braking system needs to be aware of the steering, the road layout, and the expected route.
Whenever I drive on highways in heavy rain I wonder how a self-driving car would behave. Virtually all human drivers drive unsafely in these conditions by following too closely. Would a Waymo keep the distance? Seems difficult to do in heavy traffic. The alternative I guess is to drive very slowly.
I thought this as well, but I think it emerged in an Waymo interview that any weird thing that can happen on city street can also happen on a freeway and the reaction time is lower and the consequences higher on a freeway.
Freeways have some of the same challenges as streets, but not all of them. Treating the environment as if anything can happen at any time is just an "abundance of caution" thing. To use a real example, you don't get people walking up and throwing an egg at the side of your vehicle on controlled access freeways.
Waymos are one of the first things that my visitors to SF want to try. It feels like living in a science-fiction future.
The ride and navigation feels very smooth - after the novelty of having no driver in the front seat wears off, you become accustomed to the experience surprisingly quickly. In comparison, I found the Cruise driving experience pretty uncomfortable and stress-inducing.
The app and software inside the vehicle is really well-designed. Which is unusual for a google consumer product!
And my female friends, in particular, far prefer Waymo to Uber or Lyft because they don’t have to engage with a creepy driver trying to hit on them.
I, for one, welcome our new self-driving overlords.
Maybe it is the traffic that is wrong? Vehicle-induced injury and death is still a huge issue, and marginal speed increases make no difference to overall trip times.
One of the best things about Waymo vs Uber and Lyft is that they drive smooth and comfortable, they aren’t zooming around between every stoplight and making me carsick.
I wish they would test going back to point-to-point pick-ups and drop-offs. I can easily see reasons to move to a designated pick-up and drop-off location model, but I feel it weakens the value proposition for people with mobility issues if they have to walk a few blocks.
I don't think that's up to Waymo, it's a legal thing mostly. The city has to allow that.
The funny thing about self-driving cars is that they're designed to obey laws to the letter, which real drivers never do. I think we're going to see a lot of laws necessarily getting updated to allow real-life driving behavior, such as double-parking during pickup and dropoff.
> I don't think that's up to Waymo, it's a legal thing mostly. The city has to allow that.
Did the SFPUC make a change to disallow that then, because point-to-point is how Waymo worked in SF as recently as a month or so ago? Perhaps they did, as a condition of allowing expansion.
I'm happy for Waymo and hope they continue to expand. I do expect Tesla to still win out on "robotaxis" in the nearish term, e.g. I expect them to have given way way more rides than Waymo by say 2030
Tesla has yet to demonstrate autonomous driving capability. A human ready to intervene in long tail scenarios (which is the hard part of the problem) does not constitute autonomous capability, no matter what the marketing material says.
> Tesla has yet to demonstrate autonomous driving capability
I'm betting on Waymo, too, but Tesla has chutzpah in a way Google and GM do not. That might permit them to take more risks (and incur more costs on the public) earlier.
I would love it if there was a working autonomous vehicle outside of Waymo. Google's monopoly in search engines has stifled innovation in that space for long enough. The problem is that Tesla just hasn't demonstrated they are capable of doing it. My bet is that achieving autonomous driving with only RGB vision + sensors that are not LIDAR is just too hard.
they have demonstrated autonomous driving capability. the main difference between them and Waymo is miles per intervention, of which Tesla is about 10x behind where Waymo was when they started giving rides. but soon they will be equal and Tesla will have millions more cars and built out service centers ready to activate
Waymo's continuous advancement and expansion in the field of autonomous driving technology. The plan to test fully autonomous vehicles without human safety drivers on freeways in the San Francisco Bay Area is a significant milestone. By having employees as the first testers, Waymo demonstrates confidence in its system.
I don't understand why people are so eager to give up autonomy. I doubt the software is certified in any meaningful way (see DO-178B), the algorithms are unknown, and the NHTSA sits on its hands whenever these vehicles have accidents.
Here is a litmus test: would you allow Waymo to drive schoolbuses with your children? I know I certainly will never allow my family or myself to get into a metal box going 80 mph without a driver.
Show me the statistics. If there is data which shows that autonomous vehicles are significantly safer than than the 90th percentile driver, then I absolutely will put my children in it. In fact, it might be irresponsible for me not to.
School buses are an interesting case because they drive fixed routes, so you can imagine some intense mapping and route planning process which makes them even safer than a normal autonomous vehicle. They could also afford to have a larger number of higher quality sensors.
I've seen that study before and it's not exactly hard numbers. There's nothing to show the data set used ("Researchers analyzed tens of millions of QuoteWizard by LendingTree insurance quotes") is suitable for such an analysis.
Correction: Ram is at the top, while Dodge is fairly low down. Ram is exclusively trucks (formerly Dodge Ram) and was spun off in 2010, and is now owned by Stellantis.
No muscle cars, just oversized pickup trucks with a high rate of DUI (second to BMW).
One of the last times I took a taxi, the taxi driver explained at length how he worked extra shifts so he could afford to buy meth which he would take in order to work more shifts which allowed him to buy even more meth which he would take in order to work more shifts etc.
I won't be the first in line to take Waymo on a freeway but at some point the results will speak for themselves.
This is a good point. Life is full of risks and challenges, but I guess in my case I choose to trust in humanity in cases like these... god knows Ive seen enough production software in my life to be skeptical for two.
> I won't be the first in line to take Waymo on a freeway but at some point the results will speak for themselves.
Indeed - lets hope we do get to actually see the numbers in due time.
I had a Lyft driver who nodded off several times during a ride to the point we had him leave us at the nearest corner as soon as it was safe. It was genuinely scary.
I messaged Lyft and they really didn't seem to take it seriously or care. They of course worded it in a way to avoid any potential legal problems, and I doubt anything came of it. That driver should not be on the road.
There have been several other less notable incidents where I was worried for my safety.
Does your family ride roller coasters, is that not giving up your autonomy? Waymo has done enough rides to have statistically significant risk data, and they are very safe and SF strikes me as hard a city to drive in as almost anywhere with all the rain fog pedestrians traffic, and weird routes and elevation changes.
Automotive has a similar standard called ISO 26262. ISO-26262-6 is closest to DO-178B, though ISO-21448 is also highly relevant. Both are widely used throughout the industry, including Waymo. The norm in the US is self-certification. European homologation generally requires third party certification, and Waymo is almost certainly in that process right now, as part of their discussions with European regulators.
Avionics software has a much more stringent and regulated software process (and pilots still there just in case). Are you advocating the same for cars?
It does actually, my father was a pilot for Delta and he saw the automation come thru and there are horror stories there too.
On top of this, the software quality is certified, built over decades, and continuously tested on each plane regularly - and of course the pilots are still there ; )
