There is no way Tesla survives this. They have plummeting sales every quarter, and Elon is pushing out a product that is going to get into an accident within the first week.
Robotaxis are glorified Ubers and will never level up to Waymos. FSD is useless if you have to keep your eye on the wheel and road.
I have a Tesla Model Y. It's my favorite car I've ever driven. I subscribed to FSD as recently as last month for one month to try it out. It doesn't work as well as a Waymo, not even close. It still feels like a really good high school demo but not close to being in the same ballpark as Waymo.
But Elon has killed his brand with his politics, and the robotaxi initiative is a desperate attempt to gain ground. But it's going to kill someone and it will be 100% on him because he's the one pushing this when there's no way it will ever be ready for real world situations the way Waymo is.
Tesla has proven to be very good at lying, moving goalposts, and insisting that this time [x] is right around the corner.
I don't know what would be different this time. They might launch it with, like, 5 cars total, or have it be 100% teleoperated, or limit it to side roads.
They might require the passenger to sit in the drivers seat and take liability, or push it back a few more months, or only run it at 3AM when there is no traffic.
I don't think any of these would tank the stock price, since historically Tesla has gotten away with similar skulduggery. I learned the hard way many years ago to not bet against Tesla, and I don't see anything here that would override that lesson.
Fingers crossed for you though - I definitely think Tesla is irrationally priced, and there would be a certain justice in that overinflated valuation sinking down to reality.
I have had the exact opposite experience. I have been using FSD on HW3 for a few years and the latest version for me is much safer than me driving. I dont care about your opinions of him killing the brand, they are irrelevant. It's clear that the vision only approach will work, it may need more training data but it will get there. Good luck with your short!
Last I checked there were 35K fatal accidents in the US every year - if FSD can bring that down to 3.5K that will be an obvious win.
I also got the free month of FSD about four months ago and tried it out on my Model 3, and was astonished at how bad it was in the greater Seattle area: I had to manually intervene on about 4 out of every 5 drives. I do not understand how I continue to see so many comments online raving about how great FSD is; I have to conclude these are coming from either desperate stock-boosters or people who drive exclusively on wide, flat roads in Salt Lake City early on weekend mornings when the roads are all empty.
Best of luck, I can't tell you how much I have lost having rational takes on Tesla that proved correct in reality, but ultimately immaterial to shareholders.
Tesla stock is a cult stock. People buy it because it goes up. It always goes up. Wall street has long been clued into the brain damage and delusions it's core investors have, and are more than happy to play into the fantasy. Its a company of perpetual massive promises while always carefully dancing around "hammer drop" days.
Robotaxi isn't launching, it's being rolled out over months slowly. Which will turn to years, but like always Elon will be "1 year away" to Tesla paradise. Everything this company is priced on is slowly "rolling out" with the full launch "just around the corner".
Tesla FSD right around the corner
Tesla Semi right around the corner
Tesla <$25k EV right around the corner
Tesla robotaxi right around the corner
Tesla Optimus robot right around the corner
Tesla supercar right around the corner
And people really genuinely believe this is all right around the corner, so load up now while it is still "undervalued"...
But just to be honest here, the people who have full on bought into the hype have made incredible amounts of money.
> But Elon has killed his brand with his politics, and the robotaxi initiative is a desperate attempt to gain ground. But it's going to kill someone and it will be 100% on him because he's the one pushing this when there's no way it will ever be ready for real world situations the way Waymo is.
What if his political allies allow and enable him to push this upon the populace despite it killing people.
After all, other industries have been allowed to kill plenty of people if it makes money and lines the pockets of friendly politicians of all stripes.
Maybe nobody is forcing you to get in a robotaxi, but behavior normalization based on availability is a powerful force.
Tesla can definitely survive this. They've got a low cost base, tens of millions in the bank, a rabid fan base, access to capital. It'll probably change next quarter, but as of right now they're still profitable.
Their trillion dollar capitalization is highly unlikely survive, but Tesla as a company making cars has a very long runway to survive mistakes.
While extremely overvalued by any metric, Musk has also been somehow able to keep the stock flying high for years and years on the force of hyperbolic lies (full self driving only a few month away forever) which most of the market keeps believing. It's a fascinating counterexample to the idea that the stock market can't be fooled.
So I only keep a few puts at a time because the hype surrounding the stock tends to outlive the expiration on the options. Every now and then I cash in big when it has a nice drop though.
There will be incidents and Tesla will just put the blame on drivers, and if you're buying one it most certainly has a clause that you cannot blame Tesla for the problems that the system will cause.
I think if Robotaxis take off, then his politics won't matter at all. People aren't going to care what brand of car comes to pick them up as long as they feel safe. I've been bearish on Tesla for years and they always prove me wrong. Robotaxis/Waymos are going to be national thing within the decade. It's just a bet on who survives, the biggest loser will be gig workers, that part I feel confident about.
I mean, if this were a _normal_ company, that would make sense. But it's Tesla, so it's a very dangerous game, because Tesla's stock price is at best only peripherally related to Tesla's business. You're in great risk of Musk posting a meme or having another child with a silly name or whatever, thus pushing up the stock price 10%.
Best-selling BEVs worldwide January-April 2025, according to new data from EV Volumes:
1) Tesla Model Y
2) Tesla Model 3
3) BYD Seagull/Dolphin Mini
4) Wuling Mini
5) Geely Geome Xingyuan
6) Xiaomi SU7
7) BYD Yuan Plys/Atto 3
8) BYD Yuan Up/Atto 2
9) Wuling Bingo
10) Xpeng M03
I own two Teslas, and drive between Portland and Seattle about once every 10 days, using FSD nearly the entire way.
FSD has gotten amazingly better over the last year, it can and has taken me from driveway to driveway between two cities... when the weather is nice. As soon as there is any weather however, things start to fall apart. It's very clear to me that vision alone wont be the solution to FSD, and is the main reason why I believe Waymo's approach here is simply better.
That's not to say vision wont work when the environment is good - FSD has gotten to the point where when things are optimal, it's a better driver than I am (ie it does better at merging into a lane of traffic better than I could, since it has 360 vision), but it simply isn't stable in the face of dynamic conditions.
I also use FSD HW4 daily, and it really just works well in regular conditions. It's unbelievable how we have a level of autonomy here and now, and not too many people know about it. Out of about 6k miles in the last 7 months, I probably drove 80 or so. I did not have many issues in the rain, but I do have issues in the snow. It's still somewhat unaware of how to drive well in snow. I usually disengage when it's snowy or very icy.
I have a Tesla with HW3 and FSD is an absolute dangerous joke. It nearly pulled me into a curb on its first attempt to enter a freeway.
Also, I think it keeps getting overlooked that freeways are designed from the ground up as a exclusive use of motorized vehicles. FSD performs OK there. But, taxi services are everywhere in cities around the clock in all sorts of weather. And I can't imagine trusting FSD for that use case.
that's literally just I-5 all the way, with probably 5-10 minutes of street driving at either end. So does FSD really offer _that_ much more of a benefit over adaptive cruise control with lane keeping unless you're continuously overtaking?
I've tried enabling FSD on my Tesla twice (got two one-month trials), and really wanted to love it but was disappointed with the results when not on highways. My wife tried it once and won't touch it again.
I live in major metro in the south east. HW4 FSD in a model 3 and it is dangerous. Certainly, it's a lot better than a few years ago but still nowhere near something that could safely carry me home from the bar.
We are potentially about to see a worse incident than Uber in Phoenix or Cruise in SF. 444 miles per Critical Disengagement is terrible. In 2023, Waymo reported 17,000 miles between Critical Disengagements and have made significant (as seen by a functioning robotaxi service) leaps since.
Cruise in SF was a bit of a freak accident. There were systematic issues that exacerbated the issue, but it wasn't an issue of bad disengagement numbers. In fact, Cruise 2023 actually highlights how misleading disengagement numbers can be, as they reported 0 critical disengagements the entire year over 583k miles. Waymo is extremely consistent about their numbers and they typically sandbag themselves relative to their competitors. Tesla of course doesn't officially report CA numbers, so people rely on crowdsourced data that the company and fans maintain are orders of magnitude lower than reality.
It's entirely possible that the opaqueness and small scale of the Tesla rollout could lead to situations where long tail events like the Cruise collision simply don't occur, or aren't allowed to reach public media.
> the automaker posted a new job listing days ago for engineers to help build a low-latency teleoperation system to operate its “self-driving” cars and robots.
It's June 17; they're supposedly launching in five days. Even if it's 95% off-the-shelf software, that timeframe for getting it up and debugged, then hiring enough humans to operate it, makes absolutely no sense. After nearly a decade of broken promises about FSD, has Elon finally trapped himself? Will there be a handful of "robotaxis" circling around a few blocks of Austin with Mexicans hiding in the frunk to drive them?
Thinking of re-upping my put options on $TSLA. This launch is destined to be a fiasco. It's unbelievable to me that they doubled down on vision-only at the expense of lidar. Google/Waymo is going to eat their lunch on the self-driving side, and the Tesla brand is dying with consumers. This company is cooked.
So - the cooked-ness of Tesla may depend on what you think Musk's goal was with Tesla.
If you believe he was out to change the world - maybe he'll let Tesla die. It's unlikely they are going to accomplish much else at this point, better, smarter more experienced competitors are sweeping the board.
If you believe he's out to make as much money as he can, he'll probably course correct at some point. Lidar it is clear will be the future of self-driving, and with all the adopters the economies of scale will kick in at some point.
The danger about shorting TSLA is that Elon may have a staff buying puts and calls to balance on the trading troughs. I was watching the trading volume during the big declines in the spring think the support was dropping out but it always came back. The biggest recovery was when he went to the Middle East with Trump. I think his dog and pony show had a big buy in at a critical time.
That is entirely untrue. Fred Lambert was one of the most full-throated supporters of Tesla just a few years ago. He personally referred around 15 million dollars of sales [1] to Tesla. If even Tesla's most fervent supporters now call them liars and dangerous, you should probably listen.
I personally think it'll be fine. FSD is perfect for all my driving. It's been 3000-4000 miles since my last takeover. I know an uber driver who's driven his 16k miles without a safety intervention. And this is without them superfinetuning and doing custom navigation/mapping on a specific town.
I'm a huge proponent of e2e learning for robotics (worked at two places doing e2e before Tesla adopted it) and personally believe its the right approach long term. I also have FSD on my Model 3 and love it for L2+. That said, my experience with disengagements is very different than yours.. I have a few a week for things like road works, school zones, route map following. Perfectly fine for L2+, L4 it would be unacceptable.
If these robotaxis end up looking more like my experience than yours then another layer of trouble will be root causing and fixing failure modes. Training models e2e makes both of these much more difficult.
Taxis are a useful way to get around, including to and from public transit. That's all that should matter. Whether they fit into a particular urban vision is secondary to the fact that they're desirable to the people living there.
Much of the US was built in ways that make mass transit impractical and inefficient, but those same areas have comprehensive road systems already built out, sometimes excessively. That gives self-driving taxis an opportunity to fill in the transit gap in the huge expanses of this country ill-suited for trains and buses, but well-equipped with roads.
I think the actual concern is around medium-to-large cities and metros, where self-driving cars will compete directly with mass transit, much like Uber does, but potentially much more competitively.
As an austinite I'm nervous about these things. My son and his classmates play along the street and I'm 90% sure I saw one of these driving by our house last week, presumably for testing. The street is legally at a higher speed than most people will drive because there's a lot of activity and no sidewalks which I'm about to argue for changing. Normal people will slow when they see kids around but autonomous cars still drive their normal speed.
I'm curious. Suppose I took a taxi to some specific address, which is a large building the occupies a whole block. There are two entrances to the building on that street, at opposite ends of the block.
With a human driven taxi I'd be able to tell the driver which entrance I'd like to be dropped off near.
Does Waymo provide a way to do this? Has Tesla said anything about if they will allow such a thing?
Could this moment finally be the thing that finally sinks Tesla's ridiculous stock price?
As long as the Robotaxi is just and idea, it can't fail. Once it's real, people can realize what a joke it is.
Relevant Silicon Valley:
"If you show revenue, people will ask 'HOW MUCH?' and it will never be enough. The company that was the 100xer, the 1000xer is suddenly the 2x dog. But if you have NO revenue, you can say you're pre-revenue! You're a potential pure play... It's not about how much you earn, it's about how much you're worth. And who is worth the most? Companies that lose money!"
A public display of seething hate aimed at someone adding to the world from someone only attempting to critique it. Fred, checks notes, tracks human beings that kill each other for sport for money, to make money. It must not be lucrative enough so he gets the rest of his green from writing half-truth hit pieces about companies that have more success. I'm surprised he's not writing for business insider.
Tesla had more chance to succeed than Waymo. It's impossible for Waymo to match Tesla in real world training and testing data. Somehow Waymo is winning.
Waymo chose an approach that seems geared towards actually doing the task.
Tesla tried for the moonshot - They wanted a consumer car with cheap sensor hardware to perform the job. Trusting that computing "smarts" could solve the rest of the problem.
I'm in Atlanta where Waymos have started popping up left and right - the sensor bank on these things is HUGE. You can spot 'em from way far off... Giant sensors on top. Big sensors on the front wheel wells back wheel wells, big sensors on both front and back. Big sensors basically all over them.
I'm of the opinion now that Tesla was just way, WAY off base about what sort of requirements exist for sensing, and that they don't, in fact, have much more real world training data because their data is just garbage from the cameras.
Waymo is winning because Waymo accepted the actual requirements early. Tesla is off in lala land with a dead end solution. Lots of great marketing from tesla... but very little progress now in years. They really seem stuck in a local optimum with the camera-only approach, and it's not close to delivering the promised experience.
Waymo has ~40 million miles of test data. Tesla has ~4B miles. But for training both quantity and quality are important. Tesla only uploads excerpts and metadata for its 4B miles, and it's video only. Waymo can fully analyze all 40M miles, and it's much richer, with LIDAR and other sensors.
Waymo has already lapped Tesla. They sell 250,000 rides per week and rising, for a total well over 10 million. That compares to precisely zero rides sold commercially by Tesla. It's possible Tesla could catch up, but first they need to stop losing ground and develop technology that can actually compete with their leading competitor. Plus, right now the Tesla brand is synonymous with the most obnoxious blowhard in America. Meanwhile the Waymo brand is on the way to becoming the "Kleenex" of self-driving technology.
Tesla could have more camera data in sum (that's not even clear - transmitting and storing data from all the cars on the road is no easy task - L4 companies typically pysically remove drives and use appliances to suck data off the hard drives), but Waymo has more camera data per car (29 cameras) and higher fidelity data overall (including lidar, radar, and microphone data). Tesla can't magically enhance data it didn't collect.
This is a crippling disadvantage. Consider what it takes to evaluate a single software release for a robotaxi.
If you have a simulator, you can take long tail distribution events and just resimulate your software to see if there are regressions against those events. (Waymo, Zoox)
If you don't, or your simulator has too much error, you have to deploy your software in cars in "ghost mode" and hope that sufficient miles see rare and scary situations recur. You then need to find those specific situations and check if your software did a good job (vs just getting lucky). But what if you need to A/B test a change? What if you need to A/B test 100 changes made by different engineers? How do you ensure you're testing the right thing? (Tesla)
And if you have a simulator that _sucks_ because it doesn't have physics-grounded understanding of distances (i.e. it's based on distance estimates from camera), then you can easily trick yourself into thinking your software is doing the right thing, right up until you start killing people.
Another way to look at it is: most driving data is actually very low in signal. You want all the hard driving miles, and in high resolution, so that you can basically generate the world's best unit testing suite for the software driver. You can just throw the rest of the driving data away -- and you must, because nobody has that much storage and unit economics still matter.
This is to say nothing of the fact that differences between hardware matter too. Tesla has a bunch of car models out there, and software working well one one model may not actually work well on another.
The Tesla approach was kind of magical thinking; put enough data into the magic box, and eventually a perfect driver will pop out. In some ways it's similar to the "just make a big enough LLM and it will be GAI" thing (a model which has pretty much fallen out of favour even with the most relentlessly optimistic LLM vendors in favour of so-called reasoning models).
In reality, just shovelling data into a box is not sufficient.
blindriver|8 months ago
There is no way Tesla survives this. They have plummeting sales every quarter, and Elon is pushing out a product that is going to get into an accident within the first week.
Robotaxis are glorified Ubers and will never level up to Waymos. FSD is useless if you have to keep your eye on the wheel and road.
I have a Tesla Model Y. It's my favorite car I've ever driven. I subscribed to FSD as recently as last month for one month to try it out. It doesn't work as well as a Waymo, not even close. It still feels like a really good high school demo but not close to being in the same ballpark as Waymo.
But Elon has killed his brand with his politics, and the robotaxi initiative is a desperate attempt to gain ground. But it's going to kill someone and it will be 100% on him because he's the one pushing this when there's no way it will ever be ready for real world situations the way Waymo is.
cowthulhu|8 months ago
They might require the passenger to sit in the drivers seat and take liability, or push it back a few more months, or only run it at 3AM when there is no traffic.
I don't think any of these would tank the stock price, since historically Tesla has gotten away with similar skulduggery. I learned the hard way many years ago to not bet against Tesla, and I don't see anything here that would override that lesson.
Fingers crossed for you though - I definitely think Tesla is irrationally priced, and there would be a certain justice in that overinflated valuation sinking down to reality.
WorkerBee28474|8 months ago
misiti3780|8 months ago
Last I checked there were 35K fatal accidents in the US every year - if FSD can bring that down to 3.5K that will be an obvious win.
Powdering7082|8 months ago
It's been a great way to lose money so far.
guywithahat|8 months ago
Analemma_|8 months ago
Workaccount2|8 months ago
Tesla stock is a cult stock. People buy it because it goes up. It always goes up. Wall street has long been clued into the brain damage and delusions it's core investors have, and are more than happy to play into the fantasy. Its a company of perpetual massive promises while always carefully dancing around "hammer drop" days.
Robotaxi isn't launching, it's being rolled out over months slowly. Which will turn to years, but like always Elon will be "1 year away" to Tesla paradise. Everything this company is priced on is slowly "rolling out" with the full launch "just around the corner".
Tesla FSD right around the corner
Tesla Semi right around the corner
Tesla <$25k EV right around the corner
Tesla robotaxi right around the corner
Tesla Optimus robot right around the corner
Tesla supercar right around the corner
And people really genuinely believe this is all right around the corner, so load up now while it is still "undervalued"...
But just to be honest here, the people who have full on bought into the hype have made incredible amounts of money.
stronglikedan|8 months ago
I think a lot of people would like to think this is true, ironically because of his politics.
danans|8 months ago
What if his political allies allow and enable him to push this upon the populace despite it killing people.
After all, other industries have been allowed to kill plenty of people if it makes money and lines the pockets of friendly politicians of all stripes.
Maybe nobody is forcing you to get in a robotaxi, but behavior normalization based on availability is a powerful force.
bryanlarsen|8 months ago
Tesla can definitely survive this. They've got a low cost base, tens of millions in the bank, a rabid fan base, access to capital. It'll probably change next quarter, but as of right now they're still profitable.
Their trillion dollar capitalization is highly unlikely survive, but Tesla as a company making cars has a very long runway to survive mistakes.
turnsout|8 months ago
Gud|8 months ago
laidoffamazon|8 months ago
jjav|8 months ago
I also have Tesla puts but that is brave!
While extremely overvalued by any metric, Musk has also been somehow able to keep the stock flying high for years and years on the force of hyperbolic lies (full self driving only a few month away forever) which most of the market keeps believing. It's a fascinating counterexample to the idea that the stock market can't be fooled.
So I only keep a few puts at a time because the hype surrounding the stock tends to outlive the expiration on the options. Every now and then I cash in big when it has a nice drop though.
coliveira|8 months ago
ortusdux|8 months ago
partiallypro|8 months ago
enslavedrobot|8 months ago
rsynnott|8 months ago
I mean, if this were a _normal_ company, that would make sense. But it's Tesla, so it's a very dangerous game, because Tesla's stock price is at best only peripherally related to Tesla's business. You're in great risk of Musk posting a meme or having another child with a silly name or whatever, thus pushing up the stock price 10%.
freshmatrix|8 months ago
- https://x.com/SawyerMerritt/status/1934623551694508456/photo... - "(Obi) The Road Ahead:Pricing Insights On Waymo, Uber and Lyft" https://www.documentcloud.org/documents/25973106-obi-waymo-6... - https://www.bloomberg.com/news/videos/2025-06-16/why-tesla-c...
Best-selling BEVs worldwide January-April 2025, according to new data from EV Volumes:
1) Tesla Model Y 2) Tesla Model 3 3) BYD Seagull/Dolphin Mini 4) Wuling Mini 5) Geely Geome Xingyuan 6) Xiaomi SU7 7) BYD Yuan Plys/Atto 3 8) BYD Yuan Up/Atto 2 9) Wuling Bingo 10) Xpeng M03
jjcm|8 months ago
FSD has gotten amazingly better over the last year, it can and has taken me from driveway to driveway between two cities... when the weather is nice. As soon as there is any weather however, things start to fall apart. It's very clear to me that vision alone wont be the solution to FSD, and is the main reason why I believe Waymo's approach here is simply better.
That's not to say vision wont work when the environment is good - FSD has gotten to the point where when things are optimal, it's a better driver than I am (ie it does better at merging into a lane of traffic better than I could, since it has 360 vision), but it simply isn't stable in the face of dynamic conditions.
ilikeatari|8 months ago
philips|8 months ago
Also, I think it keeps getting overlooked that freeways are designed from the ground up as a exclusive use of motorized vehicles. FSD performs OK there. But, taxi services are everywhere in cities around the clock in all sorts of weather. And I can't imagine trusting FSD for that use case.
insane_dreamer|8 months ago
that's literally just I-5 all the way, with probably 5-10 minutes of street driving at either end. So does FSD really offer _that_ much more of a benefit over adaptive cruise control with lane keeping unless you're continuously overtaking?
I've tried enabling FSD on my Tesla twice (got two one-month trials), and really wanted to love it but was disappointed with the results when not on highways. My wife tried it once and won't touch it again.
FireBeyond|8 months ago
Let's try Pittsburgh in snow in December, and then see.
Mind you as recently as 12 months ago there were still videos of Teslas happily taking straight lines through roundabouts.
coliveira|8 months ago
root_axis|8 months ago
AlexandrB|8 months ago
misiti3780|8 months ago
stingrae|8 months ago
AlotOfReading|8 months ago
It's entirely possible that the opaqueness and small scale of the Tesla rollout could lead to situations where long tail events like the Cruise collision simply don't occur, or aren't allowed to reach public media.
username223|8 months ago
> the automaker posted a new job listing days ago for engineers to help build a low-latency teleoperation system to operate its “self-driving” cars and robots.
It's June 17; they're supposedly launching in five days. Even if it's 95% off-the-shelf software, that timeframe for getting it up and debugged, then hiring enough humans to operate it, makes absolutely no sense. After nearly a decade of broken promises about FSD, has Elon finally trapped himself? Will there be a handful of "robotaxis" circling around a few blocks of Austin with Mexicans hiding in the frunk to drive them?
I'm glad I don't live there.
turnsout|8 months ago
RajT88|8 months ago
If you believe he was out to change the world - maybe he'll let Tesla die. It's unlikely they are going to accomplish much else at this point, better, smarter more experienced competitors are sweeping the board.
If you believe he's out to make as much money as he can, he'll probably course correct at some point. Lidar it is clear will be the future of self-driving, and with all the adopters the economies of scale will kick in at some point.
detourdog|8 months ago
kreetx|8 months ago
Veserv|8 months ago
[1] https://electrek.co/2019/01/17/tesla-roadster-free-killed-re...
kajecounterhack|8 months ago
matt3210|8 months ago
leesec|8 months ago
tuckerman|8 months ago
If these robotaxis end up looking more like my experience than yours then another layer of trouble will be root causing and fixing failure modes. Training models e2e makes both of these much more difficult.
JKCalhoun|8 months ago
But we don't want drivers because....
So strange to me.
JumpCrisscross|8 months ago
Taxis are a useful way to get around, including to and from public transit. That's all that should matter. Whether they fit into a particular urban vision is secondary to the fact that they're desirable to the people living there.
standardUser|8 months ago
I think the actual concern is around medium-to-large cities and metros, where self-driving cars will compete directly with mass transit, much like Uber does, but potentially much more competitively.
srj|8 months ago
tzs|8 months ago
With a human driven taxi I'd be able to tell the driver which entrance I'd like to be dropped off near.
Does Waymo provide a way to do this? Has Tesla said anything about if they will allow such a thing?
electrondood|8 months ago
darkwater|8 months ago
gizzlon|8 months ago
seydor|8 months ago
xnx|8 months ago
As long as the Robotaxi is just and idea, it can't fail. Once it's real, people can realize what a joke it is.
Relevant Silicon Valley:
"If you show revenue, people will ask 'HOW MUCH?' and it will never be enough. The company that was the 100xer, the 1000xer is suddenly the 2x dog. But if you have NO revenue, you can say you're pre-revenue! You're a potential pure play... It's not about how much you earn, it's about how much you're worth. And who is worth the most? Companies that lose money!"
https://www.youtube.com/watch?v=BzAdXyPYKQo
more_corn|8 months ago
honeybadger1|8 months ago
iw7tdb2kqo9|8 months ago
horsawlarway|8 months ago
Tesla tried for the moonshot - They wanted a consumer car with cheap sensor hardware to perform the job. Trusting that computing "smarts" could solve the rest of the problem.
I'm in Atlanta where Waymos have started popping up left and right - the sensor bank on these things is HUGE. You can spot 'em from way far off... Giant sensors on top. Big sensors on the front wheel wells back wheel wells, big sensors on both front and back. Big sensors basically all over them.
I'm of the opinion now that Tesla was just way, WAY off base about what sort of requirements exist for sensing, and that they don't, in fact, have much more real world training data because their data is just garbage from the cameras.
Waymo is winning because Waymo accepted the actual requirements early. Tesla is off in lala land with a dead end solution. Lots of great marketing from tesla... but very little progress now in years. They really seem stuck in a local optimum with the camera-only approach, and it's not close to delivering the promised experience.
bryanlarsen|8 months ago
My feeling is that Waymo has the data advantage.
standardUser|8 months ago
kajecounterhack|8 months ago
This is a crippling disadvantage. Consider what it takes to evaluate a single software release for a robotaxi.
If you have a simulator, you can take long tail distribution events and just resimulate your software to see if there are regressions against those events. (Waymo, Zoox)
If you don't, or your simulator has too much error, you have to deploy your software in cars in "ghost mode" and hope that sufficient miles see rare and scary situations recur. You then need to find those specific situations and check if your software did a good job (vs just getting lucky). But what if you need to A/B test a change? What if you need to A/B test 100 changes made by different engineers? How do you ensure you're testing the right thing? (Tesla)
And if you have a simulator that _sucks_ because it doesn't have physics-grounded understanding of distances (i.e. it's based on distance estimates from camera), then you can easily trick yourself into thinking your software is doing the right thing, right up until you start killing people.
Another way to look at it is: most driving data is actually very low in signal. You want all the hard driving miles, and in high resolution, so that you can basically generate the world's best unit testing suite for the software driver. You can just throw the rest of the driving data away -- and you must, because nobody has that much storage and unit economics still matter.
This is to say nothing of the fact that differences between hardware matter too. Tesla has a bunch of car models out there, and software working well one one model may not actually work well on another.
rsynnott|8 months ago
In reality, just shovelling data into a box is not sufficient.
misiti3780|8 months ago
https://x.com/niccruzpatane/status/1928477936845226469
https://x.com/TeslaCamera/status/1929167075731239226
https://x.com/friscolive415/status/1881181885445063041
https://x.com/LAMultiBroker/status/1885370114054512921
https://x.com/ananayarora/status/1808679192432927153
https://x.com/niccruzpatane/status/1929590206488883246
https://x.com/TheTrailerDan/status/1934618081881370766
https://x.com/JeffTutorials/status/1778188663253574035
https://x.com/ScannerPacific/status/1935016023536844960
https://x.com/greggertruck/status/1864836542402797812
https://x.com/neil_csagi/status/1803229926033858746
I could post these all day ...