[johnthuss]

All while Waymo is expanding to more and more cities, including Detroit where it will deal with snow and ice. Waymo is years ahead of Tesla in the self-driving race. It's possible Tesla never succeeds in launching a truly self-driving car.

[samwillis]

I think is really interesting how it's often suggested Waymo is at a disadvantage over Tesla due to its reliance on LIDAR and the costs associated with it. But the reality is that it's enabled Waymo to move faster and gain significant more operational experience than Tesla, and that's far more important than front loading with cost reductions in a service business.

Tesla have been operating as a product business, and cost reduction of that product was key to scale and profitability. I completely understand why they have focused on optical sensors for autopilot, lidar was always going to be impossibly expensive for a consumer product.

Waymo on the other hand have always been aiming to build a service business and that changes the model significantly, they need to get to market and gain operational experience. Doing that with more expensive equipment to move faster is exactly what was needed. They can worry about cost of building their cars later, much later.

[b112]

lidar was always going to be impossibly expensive for a consumer product.

Everything is expensive without scale. But lidar will be very cost effective, when scaled to millions upon millions of cars annually.

And with scale, there are reasons to optimise, reduce cost, etc. Large volumes of sales draws more research. Research to reduce cost.

Self driving is a long game. Decades.

[orwin]

It's already cost effective. Lidar prices have been divided by like 10 in a bit more than a decade. I've read a Wall street story about a SV company that wanted to enter the Lidar market for cars, only to bifurcate to a weird scam and a SPAC when they realized as the prices fell that Lidar would never be a very profitable.

Lidar production costs have already scaled. Now it need more miniaturization (which will help with production costs even more) and something against diffraction.

[samwillis]

Oh, yeh, don't get me wrong. I meant impossibly expensive to drop into a mid range car with no scaling up from a higher value lower volume product range first.

[ifwinterco]

It's also quite ugly, not sure you can have it on a convertible etc. - all probably solvable, but not ideal. I can see why Tesla tried the camera only approach but doesn't look like it's working out

[xnx]

> It's also quite ugly

To each their own, but it's possible having sensor bumps on your car become a status symbol that indicates you can afford a private driver.

> not sure you can have it on a convertible etc.

Radically different car shapes are possible when human driving never happens or is very rare. Maybe a small van (like a private lounge on wheels, or a train observation car) with a huge panoramic sunroof becomes en vogue.

[xnx]

> due its reliance on LIDAR

Due to its choice to use LIDAR. Waymo has tested a working system using cameras only, but they choose to use LIDAR because it is safer and does not significantly change cost.

> Waymo on the other hand have always been aiming to build a service business

Waymo's roadmap (from https://www.youtube.com/watch?v=WXLgzP3gv2k):

1) Ride hailing 2) Local delivery 3) Long haul trucking 4) Personally owned cars

Is step 4 still considered a service? It will almost certainly require a subscription.

[philipallstar]

> Waymo on the other hand have always been aiming to build a service business and that changes the model significantly

This isn't the main factor. The main factor is Waymo only does this one thing. Tesla has been: building electric cars and forcing all car makers to do the same; building charging networks; funding and releasing free of charge battery tech research improvements; and doing self driving, and all of it while trying to make a profit to keep running. Waymo is funded by Google, which has infinitely deep ads-spying-on-you pockets. Which is just much, much easier.

[xnx]

> Which is just much, much easier.

Building a self-driving car is more difficult as evidenced by no one having delivered one except for Google. Many companies (including astronomically rich ones like Apple) have tried.

[Workaccount2]

The cost for Waymo is the whole car, so Waymo is in it at ~$100k per operational taxi. They are beholden to hardware manufacturers for their product.

Tesla is trying to get in at ~$20k per operational taxi, with everything made in house.

Assuming Tesla can figure out it's FSD (and convince people it's safe), they could dramatically undercut Waymo on price, while still being profitable. If a Waymo to the airport is $20, and Robotaxi is $5, Tesla will win, regardless of anything else (assuming equal safety).

[xethos]

This is just Google doing what they've done for years now: start with the software, partner with hardware OEMs, then build their own hardware. Android to Nexus to Pixel line is one example, Google Now on Tap to their own smart speakers is another (though they may not have hit the third step there yet), and Waymo / Google Maps to self driving is following the same path.

[beAbU]

The amount of "if"s in your comment is astounding.

[jjice]

The cost discussion on LIDAR always confused a layman like me. How much more expensive is it that it seemed like such a splurge? LIDAR seems to be the only thing that could make sense to me. The fact Tesla does it with only cameras (please correct me understanding if I'm wrong) never made sense to me. The benefits of LIDAR seem huge and I'd assume they'd just become more cost effective over time if the tech became more high in demand.

I'm _way_ out of my depth though.

[madamelic]

> How much more expensive is it that it seemed like such a splurge?

LiDARs at the time Tesla decided against them were $75k per unit. Currently they are $9,300 per car with some promising innovations around solid state LiDAR which could push per-unit down to hundreds of dollars.

Tesla went consumer first so at the time, a car would've likely cost $200k+ so it makes sense why they didn't integrate it. I believe their idea was to kick off a flywheel effect on training data.

[jjice]

Holy - okay never mind, I didn't realize just how expensive LiDAR was...

[robotresearcher]

Lidar will continue to get cheaper, but it has fundamental features that limit how cheap it can get that passive vision does not.

You’re sending your own illumination energy into the environment. This has to be large enough that you can detect the small fraction of it that is reflected back at your sensor, while not being hazardous to anything it hits, notably eyeballs, but also other lidar sensors and cameras around you. To see far down the road, you have to put out quite a lot of energy.

Also, lidar data is not magic: it has its own issues and techniques to master. Since you need vision as well, you have at least two long range sensor technologies to get your head around. Plus the very real issue of how to handle their apparent disagreements.

The evidence from human drivers is that you don’t absolutely need an active illumination sensor to be as good as a human.

The decision to skip LiDAR is based on managing complexity as well as cost, both of which could reduce risk getting to market.

That’s the argument. I don’t know who is right. Waymo has fielded taxis, while Tesla is driving more but easier autonomous miles.

The acid test: I don’t use the partial autonomy in my Tesla today.

[senordevnyc]

Does the "sensor fusion" argument that Tesla made against LiDAR make as much sense now that everyone is basically just plugging all the sensor data into a large NN model?

[robotresearcher]

It's still a problem conceptually, but in practice now that it's end to end ML, plug'n'pray, I guess it's an empirical question. Which gives one the willies a bit.

It'll always be a challenge to get ground truth training data from the real world, since you can't know for sure what was really out there causing the disagreeing sensor readings. Synthetic data addresses this, but requires good error models for both modalities.

On the latter, an interesting approach that has been explored a little is to SOAK your synthetic sensor training data in noise so that the details you get wrong in your sensor model are washed out by the grunge you impose, and only the deep regularities shine through. Avoids overfitting to the sim. This is Jakobi's 'Radical Envelope of Noise Hypothesis' [1], a lovely idea since it means you might be able to write a cheap and cheerful sim that does better than a 'good' one. Always enjoyed that.

[1] https://www.sussex.ac.uk/informatics/cogslib/reports/csrp/cs...

[senordevnyc]

now that it's end to end ML, plug'n'pray, I guess it's an empirical question

Aren't human drivers the same empirical question?

That paper is really interesting, thanks!

[2OEH8eoCRo0]

Tesla's strategy is bad. Build something that works first then optimize.

They optimized for cost first and may never get it to work.

[unglaublich]

Exactly, and the Waymo sensors are practically a superset of those of Tesla, so with all the acquired data, then can build models that slowly phase out the need for Lidars.

[tim333]

It seems to me Tesla's mistake was over optimism about AI. Musk always seemed to believe they'd have it cracked the following year but it seems to be taking its time.

[consumer451]

If I recall correctly, at the last Tesla AI Day, he said ~"FSD basically requires AGI."

There is a lot to unwrap there, but that's what he said. I believe at that moment he was in ML talent recruitment mode, and yet he admitted the true scale of this issue that Tesla faces given the vision-only direction.

[consumer451]

That's an interesting take. So, classic premature optimization for scale?

[xnx]

Yes.

A normal approach is:

1) Make it work

2) Make it right(/safe)

3) Make it fast/cheap

[consumer451]

The surprising thing to me is when you look at Starlink, there was a very expensive blocker there: consumer phased array. Prior to Starlink, I think the cheapest consumer unit was around $50k. That did not stop Musk from charging ahead.

Is there some technological thing about LiDAR that would prevent similar cost reductions? Or, is it just the philosophical difference over pre-mapping, and not doing so?

[xnx]

LIDAR has seen many cost reductions as capabilities continue to increase. I don't know the area well enough to speculate how much optimization might be left.

> Or, is it just the philosophical difference over pre-mapping, and not doing so?

It seems to be a "burn the ships" style bet that the Tesla engineers will get to camera-only self driving first without having ever relied on LIDAR. It's equally as likely (or moreso) that Waymo could get there first with better ground truth data from the LIDAR.

[BobaFloutist]

I think it's that with Tesla he had hardware to sell (and maybe already sold?) to existing customers with the contractual promise that they'd get self-driving as soon as TESLA cracked it. Retrofitting LIDAR into all those already sold cars would have been pretty expensive at the time, and the more he doubles down the more monstrously expensive it'll get.

With Starlink, there was no baseline consumer product to sell before getting it working.

[vardump]

Have they already solved the issue of LIDAR destroying CMOS cameras with its laser?

[seanhunter]

> lidar was always going to be impossibly expensive for a consumer product.

I just don't buy this at all

>"The new iPad Pro adds ... a breakthrough LiDAR Scanner that delivers cutting-edge depth-sensing capabilities, opening up more pro workflows and supporting pro photo and video apps." [1]

Yes of course the specs of LiDAR on a car are higher but if apple are putting it on iPads I just don't buy the theory that an affordable car-spec LiDAR is totally out of the realm of the possible. One of the things istr Elon Musk saying is that one of the reasons they got rid of the LiDAR is the problem of sensor fusion - what do you do when the LiDAR says one thing and the vision says something different.

[1] https://www.apple.com/uk/newsroom/2020/03/apple-unveils-new-...

[Veedrac]

Tesla got rid of radar because of sensor fusion, and particularly for reasons that wouldn't apply to high resolution radar. Sensor fusion with a high resolution source like LiDAR isn't particularly tricky.

[robotresearcher]

The iPad lidar has a range of a handful of meters indoors and is not safety critical.

Higher specs can make all the difference. A model rocket engine vs Space Shuttle main engine, for an extreme example. Or a pistol round vs an anti-armor tank round. The cost of the former says nothing at all about the latter.

[seanhunter]

OK, how about this? Volvo EX90, a consumer SUV on sale now in the UK. Fitted with Lidar.

https://www.volvocars.com/uk/support/car/ex90/article/47d2c9...

[robotresearcher]

They are getting there. But that link has big caveats. Not sure how cool it is to endanger other people’s cameras.

From your linked page:

> Important Use responsibly

The lidar and features that can rely on it are supplements to safe driving practices. They do not reduce or replace the need for the driver to stay attentive and focused on driving safely.

Safe for the eyes

The lidar is not harmful to the eyes.

Lidar light waves can damage external cameras

Do not point a camera directly at the lidar. The lidar, being a laser based system, uses infrared light waves that may cause damage to certain camera devices. This can include smartphones or phones equipped with a camera.

[CamperBob2]

How many cameras don't use IR filters? At one time, at least, they were quite common.

I suppose one example might include fixed security cameras with IR-based night vision capability.

[hnburnsy]

And that Lidar is not assisting drivers and prevously sold US EX90s now need a computer upgrade to get it to work

[thiago_fm]

That's what I've been saying for years.

Waymo can just add the cameras exactly the way Tesla has, and train based only on that information.

Now it has tons and tons of data, they could gradually remove the Lidar on cities that they've driven over and over again. IF driving without Lidar is worth it... maybe it isn't even worth it and we should pursue using Lidars in order to further reduce accidents.

Meanwhile people use Tesla sporadically in a few spots they consider safe, they will always have data that isn't useful at all, as it can already drive on those spots.

--

Another thing, we can definitely afford to have Lidars on every car, if that would make our cars safer.

Imagine if China does a huge supply chain of Lidars, I bet the cost would be very tiny. And this is supposing there aren't any more automation and productivity gains in the future, which is very unlikely.

Lidar production just doesn't have that big scale, because it's a very tiny market as of now. With scale, those prices would fall like batteries and other hardware have fallen with the years.

TLDR: Tesla lost badly

[testing22321]

I have no doubt Waymo is ahead of Tesla right now.

Why does it then follow that Tesla will never release a truly self driving car? That makes no sense.

Is it somehow impossible for an achievement to be reached by more than one person/company?

Of course not. The hyperbole is not needed, and does nothing but remove credibility from your statement

[esalman]

"Tesla will never release a truly self driving car" unless they significantly change direction.

It is possible to fool a camera with some specs of dust at close range. They have interior safety camera, and everyone I know put a cover on it all the time.

[xnx]

> Why does it then follow that Tesla will never release a truly self driving car?

Never is a long time, but it's possible that Tesla adopts LIDAR before it gets camera-only to work.

There are all kind of historical technological dead-end examples (e.g. planes that flap their wings).

[testing22321]

Of course it’s not only possible but virtually a certainty that an approach to doing anything difficult will be refined, modified and improved as lessons are learned.

Does anyone now care how much SpaceX changed their reuse strategy for the falcon 9? Of course not. It only matters that they found a way that works.

[testing22321]

You mean to say a company or person will change their approach, adapt and grow as they learn how to best achieve their goal?

Ummmm yes. That is the definition of doing something difficult.

[xoa]

>You mean to say a company or person will change their approach, adapt and grow as they learn how to best achieve their goal?

There's nothing wrong with that in a vacuum, but it's objectively a different story when the company chooses to turn it into an enormous financial liability. That is a massive practical difference, with extremely pertinent example being Tesla vs SpaceX. SpaceX has had enormous success running a hardware rich testing regime, and they're very willing to move fast and break things... but NEVER with customer payloads. Which is how that aphorism was supposed to be applied. You can and often should be extremely aggressive exploring the entire problem space in testing before you move into final build and then production, but once you're dealing with customers then it has to be done very differently.

Tesla absolutely could have experimented to their heart's desire without making any financial promises let alone actually entering into end customer contracts and taking hundreds of millions of dollars in preorder money for a future feature on ass-pulled timelines. But that's not the path they took, and that leaves them in a much uglier position when it comes to reacting to data and changing their approach because they've already locked in hardware to paying customers. That's the exact opposite of what SpaceX or any normal responsible company does.

[chneu]

There are serious concerns about Tesla even being around in a few years.

They have very large legal issues right now. Legal issues that are going to cost them tens of billions, depending on the legal proceedings.

[testing22321]

What legal issues?

They also have $41 billion in the bank…

[reaperducer]

It's possible Tesla never succeeds in launching a truly self-driving car.

It doesn't have to. The money's been made. The stock's been pumped. Mission accomplished.

As for calling it fraud, well the government's closed. The moose out front should have told 'ya.

[hnburnsy]

It would be sad if Tesla or some other entity doesn't compete in the space? I dont want a monolpoly or even a duopoly. Give me 4 or 5 players for true choice and competition.

[krupan]

Someone should have kept investing in Cruise for that. They were basically on par with Waymo, but then GM stopped the funding and shut Cruise down

[LurkandComment]

For a fraction of Musk's bonus he could buy Waymo

[drivingmenuts]

What value does he bring to Tesla? or SpaceX? Or any of his other companies, for that matter?

[pstuart]

He pumps up the meme stock price -- the reason it's so high anyway is from his showboating (autonomous driving! robots! AI! robotaxi! etc...)

[fluidcruft]

Musk is gonna buy Alphabet? Fascinating.

[Octoth0rpe]

Alphabet might be willing sell off just waymo.

[dylan604]

Has Waymo become successful reaching a critical mass of users yet? If so, they would most likely shut it down based on historical examples. Imagine if they had sold those deprecated products instead. Maybe not a financial significance, but there'd be some interesting products still kicking

[madamelic]

Considering Boston Dynamics sat around for like 15 years being a research lab and only started commercializing when they were sold... I'd agree.

Argue with that as you like but Google _loves_ to sit around on good ideas and, in my opinion, hamstring them away from pushing their products to commercialization.

[senordevnyc]

What's the highest revenue product that Google shut down? Maybe Google+? But I wouldn't be surprised if Waymo is already making way more revenue than Google+ ever did, while barely scratching the surface of immediate demand.

[dylan604]

was G+ around long enough to generate revenue at all? there was no activity on it that would have been worthy to buy ad space within it. obviously, theGoog had the advantage of owning the ad space just like they had an advantage of converting all G users to G+ users without consent from those users. Even with that, they still had no users

[4kchiefofstaff]

Alphabet is not stupid. They are doing this to be part of the ecosystem.

The same with robots.

Before they sell it to some competitor, they will shelf it and will bet on nvidias autonomoes car platform.

[fluidcruft]

So the idea is a memestock that can't deliver can just memebuy its competitor with fake money?

Pure anti-Capitalist nonsense. Companies fail. Let them die.

[octocop]

They are in completely different leagues, the Tesla robotaxi vehicle is probably cost 10x less, at least.

[dachworker]

Isn't the lesson from the success of TSLA, that you don't compete on price? That's what made Tesla the first successful EV. Because unlike the rest, they didn't try to compete on price and offer a mass market consumer vehicle. Instead they started with a roadster and then a luxury saloon both targeting the upper end of the market. I don't see the point of a budget taxi car. After all even the human driven counterparts tend to be higher end luxury saloons or SUVs.

[madamelic]

Robot Taxis will be competing on price. Whoever can release the lowest cost per mile and most reliable taxi will take lion's share simply because consumers are generally price conscious about transport. Very few will be analyzing the data if two are judged to be 'safe enough', it will come down to price.

Companies like BYD and Tesla are positioned well for that if they can get their AV functionality proven out as both are fully integrated car manufacturers.

Waymo doesn't have in-house manufacturing and is, to my knowledge, purely software so they have lots of vendors along with a relatively low output of vehicles. Their 2025 and 2026 plan is to build 2,500 new cars per year. Each Waymo car currently costs over $100k. Even if Tesla was pushing out Model Ys as their robotaxi platform, they could flood the market very easily in both scale and price per mile _if_ UFSD (unsupervised FSD) was proven.

[4kchiefofstaff]

I did a basic napkin calculation in the other comment. The price of the car is not that relevant per km than you make it to be.

I think self driving will be a commodity in the long term and every car will be able to do it. If Tesla will solve it purly by cameras, every other car manufacturer will be able to add this too. Perhaps a few years later but they will be able to do it too.

So Tesla has to leverage the first mover advantage, and they are loosing this already.

And while Musk says robot taxis are fundamental to tesla, the taxi market is actually not that big. All the broad nice areas like small cities etc. will buy a small fleet of cars and i don't think the price point of a Tesla will that crazy much cheaper than whatever everyone else will have that it will be obvous for everyone to just buy the Tesla model.

I alone will not use Tesla alone for Musk. Despite that, people might want to pay a euro more to have a SVU to have space or higher entry point than choosing the cheapest Tesla model to drive with.

Tesla can't flood the market very easily. If they could, they would have done it. And its expected that Tesla will not suddenly find the solution to their problems. They are optimizing away the next 9 at the 9x% reliability. Every additional 9 will take the same amount as the previous 9. And the nines are quite relevant if you look how many km these cars will have to drive.

[dachworker]

Fair point, but here's my counter: consumers won't analyze the data but insurance companies will.

[vel0city]

If Americans were price conscious about transport they wouldn't be driving $60,000, 15mpg, oversized pickup trucks to go drop off their kids at daycare and commute to their office job, they'd be riding the bus.

Most Americans don't seem to consider the cost of their transportation in the slightest.

[jjav]

> If Americans were price conscious about transport they wouldn't be driving $60,000, 15mpg, oversized pickup trucks to go drop off their kids at daycare and commute to their office job, they'd be riding the bus.

> Most Americans don't seem to consider the cost of their transportation in the slightest.

Time is also an important cost. It would take us about 90 minutes from home to school to drop off my kid by bus (plus walking, since no bus stops near the school).

By car, it is 15 minutes worst case if I hit all red lights.

By car we leave home 8:15, kid is in school on time and I'm in my work meetings easily by 9:00.

By bus, we'd have to leave home at 6am and I might just barely make it in time for 9am meetings, or often be late.

So yes, people do consider the cost of transportation but it is not just dollars, also time.

[vel0city]

The level of bus service is a societal choice. One could get bus service to be a lot more competitive in time. But I do agree, individuals do not have much control over bus schedules and people do what they can with what's currently available.

The context of the previous comments was clearly about monetary cost though, not other kinds of cost. There's also obviously environmental, health outcomes, etc, costs in question.

[madamelic]

People who use ride share use more than one app because they can pick the one that is the cheapest. The people who use these will be price conscious.

Of course there will be other factors like amenities.

Personally, I think 'style' is going to be a non-insignificant factor to it as well. Few normies will want to get out of a 'nerd car' that has bulbous sensors all over it if they can pay a bit more to have a cooler looking ride, it's the Prius effect.

The style thing is just my opinion though but price will be the major one. People will tolerate an ugly robotaxi if it is significantly cheaper or more convenient.

[bryanlarsen]

With ride share you can take the cheap option 99 times out of 100, and then rent a limo for your hot date. And then rent a truck or movers when you need to move something. People buying vehicles usually buy something that covers all their needs, however rare.

[xnx]

> If Americans were price conscious about transport ... they'd be riding the bus.

Have you considered that Americans might value their time differently than you? That might change your equation.

[vel0city]

So in other words, they value other things much higher than cost? Gee, sure sounds like exactly what I stated. People care more about other things than the cost, like how cool they look or how much departure angle they can achieve while they drive around in a parking garage. Overall TCO rarely figures into it.

[Workaccount2]

They do, but they price it in cost per month.

Finacialization is what made $65,000 cars "cheap".

[LorenDB]

I assume you mean "sedan" rather than "saloon"?

[dachworker]

It's the British English equivalent.

[tverbeure]

Does that include the cost of a full-time chaperone?

That aside, the cost of a Waymo is estimated to be between $150 and $200k. A model 3 based Tesla robotaxi doesn’t cost less than $20k…

[Workaccount2]

The reason for the cybercab is to produce an absolute bare bones self driving vehicle for $20kish.

[tverbeure]

Every Google result says below $30k, not $20k-ish, and we all know the reliability of Tesla’s future looking statements. We’ll see.

[gibolt]

Cost to manufacture is likely around $25k

[4kchiefofstaff]

Calculate how much a car can drive (200-400k km), then the avg cost of a car (50k vs. 100k) and the avg taxi route (5-30km).

The car itself is a price point of 10 to 40 cent pro km which has impact on the journey for sure but a lot less that it might be the reason.

if you tell me, that i can take the saver car and pay 1 euro more with a 20 euro fair, I wouldn't care.

Nonetheless, economy of scale has happened already at lidar and continues to happen.

If tesla can't get it running properly in bad weather but waymo can, they can also compensate it just by driving at situations were tesla doesn't want to drive.

But hey its just brainstorming at this point as tesla is not close enough to waymo to compare it properly. And while waymo exists, plenty of other companies exist too doing this. Nvidia itself will keep building their car platform which will level the playfield even more.

Whatever market selfdriving cars are, it will be split between everyone and no tesla will not just 'win' this. It will be a race to the bottom for everyone reducing the revenue to a commodity.

[don_neufeld]

LIDAR, like other technologies, declines in cost over time, has done so substantially, and will continue to do so.

Tesla is just on the wrong side of that bet.

[xnx]

> They are in completely different leagues

Agreed. Waymo has a working self-driving vehicle that currently operates in many cities. Tesla has a buggy tech demo in a portion of Austin.

> the Tesla robotaxi vehicle is probably cost 10x less, at least.

Very unlikely. Waymo vehicles also carry twice as many people.

[4b11b4]

Yeah but, it doesn't work good