6 ms·

Waymos however need thousands of dollars of hardware to achieve this and only work in limited areas. Tesla's bet is a lot more risky but also with a lot more po

by rapsey 2y ago

Waymos however need thousands of dollars of hardware to achieve this and only work in limited areas. Tesla's bet is a lot more risky but also with a lot more potential.

hyfgfh 2y ago
Potential lawsuits
slac 2y ago
I strongly believe that once you have everything working it's much easier to start working on the costs.
seanmcdirmid 2y ago
This is definitely the big bet Tesla is making. However, the hardware Waymo uses will also become cheaper over time and with scale, so either bet has advantages and disadvantages.
ulfw 2y ago
Waymos exist. This intro of nothing doesn't. It's much much easier to make an existing thing cheaper and better over time.
- rapsey 2y ago
  Waymos drive in limited geofenced zones. FSD exists but needs driver supervision. Both are incomplete and it is unknown who will be the winner.
  - imtringued 2y ago
    The people living in the "geofenced" areas don't care about how "incomplete" it is. The point of a self driving taxi is that you don't have to own the vehicle.
    - vidarh 2y ago
      And this is a vast difference. You can expand a self-driving taxi service by running an Uber-like service and dispatching w/safety-drivers for journeys taking you out of the area, and its functionally equivalent - people are buying the mobility, not the self-driving. You get to start rolling out without solving the whole problem.
  - hackernewds 2y ago
    The geofence can extend. Drivers are harder to replace.
    - haliskerbas 2y ago
      Have been seeing both the LA and SF Waymo operating zones increase steadily. Also seen Waymo's being driven manually outside of those ranges presumably for further development.
  - altacc 2y ago
    FSD is also geofenced, just a bigger fence. When/if the robotaxi actually debuts it too will be heavily geofenced and have usage restrictions but it will also be several years behind Waymo in terms of development, testing, technology and regulation. Whilst it is unknown who will be the winner, or even valid competitors, we can predict with high confidence that Tesla has a massive challenge to reach where Waymo is today.
  - Yoric 2y ago
    But isn't it the case that Waymos are actually usable (for this purpose), while FSD/RoboTaxi isn't? Geofencing sounds like a good idea to me. It's a mean to roll things out carefully, while minimizing risk of death. If actual FSD/Robotaxi is ever released, I suspect that they'll need to geofence, too, for a while.
fshbbdssbbgdd 2y ago
The idea that Tesla would win the robotaxi race by not needing LiDAR died sometime between when LiDAR cost $100k and when it cost $1k. Now it’s just Elon being intransigent.
- pr337h4m 2y ago
  Humans don’t use lidar, which clearly shows that a vision-only robotaxi is very much feasible.
  - p_j_w 2y ago
    Humans use our brains to drive. Unless you're planning on popping an actual human brain or something that can perform equivalently into the car, you'd do well to consider more superior sensor suites.
  - trhway 2y ago
    Birds have to flap wings while our planes don't have to. There is absolutely no reason to limit self-driving cars in the same way our bodies are limited. When it comes to AI though, humans are using biological neural net much more capable than any today's AI you can cram into a car. So, even if one accepts your premise of targeting human performance as a design guideline, more sensors is still logical at this point as way to compensate for the weaker AI. Also, if you read how Tesla does vision it is very different from, and i think inferior to, how your eyes and brain build the 3d map of the surroundings. If one is limiting oneself to only vision, the first thing would be to try to get as good as possible that 3d mapping, and the vision seems to be among the simplest and most researched brain functions, ie. easiest to reproduce. As Tesla doesn't seem to be doing it - only may be couple years ago they only started to elicit the 3d model - i think they aren't on the shortest path to success when it comes to FSD.
    - meow_catrix 2y ago
      Planes do ”flap their wings”, just not the ones protruding from the fuselage.
imtringued 2y ago
How does it have more potential? I don't know how much taxi drivers earn in the US but let's pick a random number like 50k USD per year. If Waymo hardware costs $10k but Tesla costs $2k, then the savings are $40k for Waymo in the first year and $48k for Tesla. That is a 20% increase in the worst case for Waymo and the longer the vehicle lasts or the higher the taxi driver salary, the worse it gets for Tesla. Their potential is heavily bounded and gets worse over time.
- hackernewds 2y ago
  Not to mention, Tesla might not be able to do it at all. Plus a driver that makes magnitudes less mistakes, is punctual, doesn't sue, doesn't rest, or is a liability is worth much more than $50k. The human component though we could argue whether humans are meant to be driving cars full time putting themselves and others at human error risk
i2infinity 2y ago
I wouldn’t step into a Tesla robotaxi in bad weather, period. They’d absolutely need a human remotely operating it. Without a steering wheel, passengers can’t take control even if they wanted to. Even in good weather, I’d be genuinely surprised if Teslas, in their current form, could drive around autonomously. I was really hoping Musk would mention new sensors being added for extra safety, maybe spinning it as: “Your Model 3 doesn’t need additional sensors, but just to be safe, we’re adding new ones.”
- bryanrasmussen 2y ago
  can multiple operators from India operate the robotaxis or does it need one on one operation? I mean consider the savings!
  - karel-3d 2y ago
    The latency will kill you. Both figuratively and literally. Maybe something like Mexico would be better.
    - varjag 2y ago
      Latency? That's when Starlink comes into play!
    - bryanrasmussen 2y ago
      >The latency will kill you. I mean - I wasn't thinking I would risk it!!
  - i2infinity 2y ago
    Just ranting here - The psychological disconnect between a remote operator and passengers in a robotaxi needs more research though. Remote operator might have less empathy and responsibility towards passengers possibly causing moral disengagement. The remote driver might never face the real-world aftermath of their actions, which can reduce their sense of remorse or responsibility. There could be complex legal dilemmas too (especially if operator is from a different country) But this has definitely been researched a lot in the field of military drone operators who can make life altering decisions from thousands of miles away.
petters 2y ago
I think “thousands of dollars” in hardware is fine
davedx 2y ago
My Tesla still beeps at me because it thinks I'm about to drive into pedestrians or parked cars because there's a bend in the road. I honestly think at this point Tesla's FSD AI is way, way overfitting on a few US cities.
- jajko 2y ago
  Well my 10 year old BMW F11 does it too sometimes, it really is a stupid primitive technology with tons of badly handled corner cases. Luckily its not obnoxious and I got used to it quickly so ignoring it. But in critical situations it can take away a bit of focus which is pretty bad. Of course can't be turned off. Nobody expected 15 year old design from BMW to perform better I guess. From modern up-to-date teslas who don't even have steering wheel but lidar is a no-no because his ego? I can't imagine it getting approved in Europe, ever. Which is fine, there will be tons of competition for this in few years.
- kalleboo 2y ago
  It's way overfitting on the routes that the CEO and a few YouTubers drive. https://electrek.co/2024/07/09/tesla-insiders-say-elon-optimized-full-self-driving-routes-for-himself-influencers/ https://electrek.co/2024/07/09/tesla-insiders-say-elon-optim...
prmoustache 2y ago
Isn't any Tesla costing thousands of dollars anyway?
specialist 2y ago
Tesla's autobots are free? Waymo & Hyundai announced a partnership. IIRC Waymo has always intended to work with OEMs, vs make their own vehicles. https://seekingalpha.com/news/4156375-hyundai-motor-joins-forces-with-waymo-to-expand-autonomous-vehicle-foundry-business https://seekingalpha.com/news/4156375-hyundai-motor-joins-fo... Having no opinions about the IONIQ 5, I've gleaned that it's well regarded. Maybe not a Model Y, but close enough. Of all the legacy OEMs, Hyundai has a fair chance of surviving the Tesla (& BYD) juggernaut. So I think Waymo chose wisely.
danso 2y ago
But Teslas vehicles can’t operate ANYWHERE autonomously — not even with supervision inside the closed loop tunnels in Las Vegas [0] — 6+ years after Musk said autopilot/FSD was capable of driving itself coast to coast. Whenever Teslas manage to offer autonomous driving, what makes you think LIDAR etc will still cost what it does now? [0] https://www.reviewjournal.com/news/news-columns/road-warrior/new-tunnels-coming-for-vegas-loop-autonomous-driving-on-horizon-3147274/amp/ https://www.reviewjournal.com/news/news-columns/road-warrior...

by rapsey 2y ago

Waymos however need thousands of dollars of hardware to achieve this and only work in limited areas. Tesla's bet is a lot more risky but also with a lot more potential.

hyfgfh 2y ago

Potential lawsuits

slac 2y ago

I strongly believe that once you have everything working it's much easier to start working on the costs.

seanmcdirmid 2y ago

This is definitely the big bet Tesla is making. However, the hardware Waymo uses will also become cheaper over time and with scale, so either bet has advantages and disadvantages.

ulfw 2y ago

Waymos exist. This intro of nothing doesn't. It's much much easier to make an existing thing cheaper and better over time.

rapsey 2y ago
Waymos drive in limited geofenced zones. FSD exists but needs driver supervision. Both are incomplete and it is unknown who will be the winner.
- imtringued 2y ago
  The people living in the "geofenced" areas don't care about how "incomplete" it is. The point of a self driving taxi is that you don't have to own the vehicle.
  - vidarh 2y ago
    And this is a vast difference. You can expand a self-driving taxi service by running an Uber-like service and dispatching w/safety-drivers for journeys taking you out of the area, and its functionally equivalent - people are buying the mobility, not the self-driving. You get to start rolling out without solving the whole problem.
- hackernewds 2y ago
  The geofence can extend. Drivers are harder to replace.
  - haliskerbas 2y ago
    Have been seeing both the LA and SF Waymo operating zones increase steadily. Also seen Waymo's being driven manually outside of those ranges presumably for further development.
- altacc 2y ago
  FSD is also geofenced, just a bigger fence. When/if the robotaxi actually debuts it too will be heavily geofenced and have usage restrictions but it will also be several years behind Waymo in terms of development, testing, technology and regulation. Whilst it is unknown who will be the winner, or even valid competitors, we can predict with high confidence that Tesla has a massive challenge to reach where Waymo is today.
- Yoric 2y ago
  But isn't it the case that Waymos are actually usable (for this purpose), while FSD/RoboTaxi isn't? Geofencing sounds like a good idea to me. It's a mean to roll things out carefully, while minimizing risk of death. If actual FSD/Robotaxi is ever released, I suspect that they'll need to geofence, too, for a while.

fshbbdssbbgdd 2y ago

The idea that Tesla would win the robotaxi race by not needing LiDAR died sometime between when LiDAR cost $100k and when it cost $1k. Now it’s just Elon being intransigent.

pr337h4m 2y ago
Humans don’t use lidar, which clearly shows that a vision-only robotaxi is very much feasible.
- p_j_w 2y ago
  Humans use our brains to drive. Unless you're planning on popping an actual human brain or something that can perform equivalently into the car, you'd do well to consider more superior sensor suites.
- trhway 2y ago
  Birds have to flap wings while our planes don't have to. There is absolutely no reason to limit self-driving cars in the same way our bodies are limited. When it comes to AI though, humans are using biological neural net much more capable than any today's AI you can cram into a car. So, even if one accepts your premise of targeting human performance as a design guideline, more sensors is still logical at this point as way to compensate for the weaker AI. Also, if you read how Tesla does vision it is very different from, and i think inferior to, how your eyes and brain build the 3d map of the surroundings. If one is limiting oneself to only vision, the first thing would be to try to get as good as possible that 3d mapping, and the vision seems to be among the simplest and most researched brain functions, ie. easiest to reproduce. As Tesla doesn't seem to be doing it - only may be couple years ago they only started to elicit the 3d model - i think they aren't on the shortest path to success when it comes to FSD.
  - meow_catrix 2y ago
    Planes do ”flap their wings”, just not the ones protruding from the fuselage.

imtringued 2y ago

How does it have more potential? I don't know how much taxi drivers earn in the US but let's pick a random number like 50k USD per year. If Waymo hardware costs $10k but Tesla costs $2k, then the savings are $40k for Waymo in the first year and $48k for Tesla. That is a 20% increase in the worst case for Waymo and the longer the vehicle lasts or the higher the taxi driver salary, the worse it gets for Tesla. Their potential is heavily bounded and gets worse over time.

hackernewds 2y ago
Not to mention, Tesla might not be able to do it at all. Plus a driver that makes magnitudes less mistakes, is punctual, doesn't sue, doesn't rest, or is a liability is worth much more than $50k. The human component though we could argue whether humans are meant to be driving cars full time putting themselves and others at human error risk

i2infinity 2y ago

I wouldn’t step into a Tesla robotaxi in bad weather, period. They’d absolutely need a human remotely operating it. Without a steering wheel, passengers can’t take control even if they wanted to. Even in good weather, I’d be genuinely surprised if Teslas, in their current form, could drive around autonomously. I was really hoping Musk would mention new sensors being added for extra safety, maybe spinning it as: “Your Model 3 doesn’t need additional sensors, but just to be safe, we’re adding new ones.”

bryanrasmussen 2y ago
can multiple operators from India operate the robotaxis or does it need one on one operation? I mean consider the savings!
- karel-3d 2y ago
  The latency will kill you. Both figuratively and literally. Maybe something like Mexico would be better.
  - varjag 2y ago
    Latency? That's when Starlink comes into play!
  - bryanrasmussen 2y ago
    >The latency will kill you. I mean - I wasn't thinking I would risk it!!
- i2infinity 2y ago
  Just ranting here - The psychological disconnect between a remote operator and passengers in a robotaxi needs more research though. Remote operator might have less empathy and responsibility towards passengers possibly causing moral disengagement. The remote driver might never face the real-world aftermath of their actions, which can reduce their sense of remorse or responsibility. There could be complex legal dilemmas too (especially if operator is from a different country) But this has definitely been researched a lot in the field of military drone operators who can make life altering decisions from thousands of miles away.

petters 2y ago

I think “thousands of dollars” in hardware is fine

davedx 2y ago

My Tesla still beeps at me because it thinks I'm about to drive into pedestrians or parked cars because there's a bend in the road. I honestly think at this point Tesla's FSD AI is way, way overfitting on a few US cities.

jajko 2y ago
Well my 10 year old BMW F11 does it too sometimes, it really is a stupid primitive technology with tons of badly handled corner cases. Luckily its not obnoxious and I got used to it quickly so ignoring it. But in critical situations it can take away a bit of focus which is pretty bad. Of course can't be turned off. Nobody expected 15 year old design from BMW to perform better I guess. From modern up-to-date teslas who don't even have steering wheel but lidar is a no-no because his ego? I can't imagine it getting approved in Europe, ever. Which is fine, there will be tons of competition for this in few years.
kalleboo 2y ago
It's way overfitting on the routes that the CEO and a few YouTubers drive. https://electrek.co/2024/07/09/tesla-insiders-say-elon-optimized-full-self-driving-routes-for-himself-influencers/ https://electrek.co/2024/07/09/tesla-insiders-say-elon-optim...

prmoustache 2y ago

Isn't any Tesla costing thousands of dollars anyway?

specialist 2y ago

Tesla's autobots are free? Waymo & Hyundai announced a partnership. IIRC Waymo has always intended to work with OEMs, vs make their own vehicles. https://seekingalpha.com/news/4156375-hyundai-motor-joins-forces-with-waymo-to-expand-autonomous-vehicle-foundry-business https://seekingalpha.com/news/4156375-hyundai-motor-joins-fo... Having no opinions about the IONIQ 5, I've gleaned that it's well regarded. Maybe not a Model Y, but close enough. Of all the legacy OEMs, Hyundai has a fair chance of surviving the Tesla (& BYD) juggernaut. So I think Waymo chose wisely.

danso 2y ago

But Teslas vehicles can’t operate ANYWHERE autonomously — not even with supervision inside the closed loop tunnels in Las Vegas [0] — 6+ years after Musk said autopilot/FSD was capable of driving itself coast to coast. Whenever Teslas manage to offer autonomous driving, what makes you think LIDAR etc will still cost what it does now? [0] https://www.reviewjournal.com/news/news-columns/road-warrior/new-tunnels-coming-for-vegas-loop-autonomous-driving-on-horizon-3147274/amp/ https://www.reviewjournal.com/news/news-columns/road-warrior...