Investors should've known they were being taken for the proverbial 'ride' when Tesla didn't even trust their driverless cars to operate without a driver in an access-restricted Las Vegas tunnel, where Tesla controls the environment top to bottom.
netsharc · 1d ago
I was wondering if their taxi turned from an inconvenient 2-seater to a very incovenient 1-seater plus a dummy driver, but I guess they're testing with model Xes, not the 2-seaters they demoed...
toomuchtodo · 1d ago
They are cosplaying robotaxis to hold the stock price up.
SR2Z · 1d ago
I wonder if Musk will just give up on this whole "camera only" thing and use LiDAR and RADAR the way a Waymo would, but with less redundancy and huge economies of scale.
This could actually be a decent opportunity for Tesla to branch out into manufacturing these sensors.
Of course, it's much more likely he clings to "camera only" until he dies.
Zigurd · 18h ago
I think the discussion about sensors leaves out some important aspects. If you don't have the geospatial data, a lidar point cloud isn't going to be as useful. Tesla has a lot of 2-D image data and could easily get a lot more. But if 3-D sensing and geospatial data turn out to be a large advantage, what's the value of all that 2-D imagery?
Of course there are some inherent advantages to lidar sensing without Google's data muscle behind it, like sensing at night and in glaring sun. And then there's the radar and maybe sonar sensing for up-close obstacles. But, overall, I think the value of Google's data is underappreciated.
SR2Z · 15h ago
Musk did give up on the "no maps" thing a while ago. Tesla does map places for FSD.
toomuchtodo · 18h ago
Musk could deploy a bespoke robotaxi fleet today, with only those units upgraded with lidar and sensor fusion after they roll off the manufacturing line, similar to Waymo’s fleet. Waymo only has ~2k-3k vehicles in their fleet, for example. Why does he not? Is it because he really believes camera over lidar? Or it because he doesn’t care enough to because investors keep believing the marketing which continues to keep him wealthy and in power with no effort?
kelnos · 1d ago
It is absolutely hilarious to me that they call these "robotaxis". But this is just another in a long line of examples of Tesla naming things in ways just so they can sound like their capabilities are much more than they are.
No comments yet
1123581321 · 1d ago
Is there any publicly accessible data on the engagement rate of the safety monitors?
anonym29 · 1d ago
Missing from the headline:
1. This is not true for the existing commercial robotaxi fleet that the public can get rides in today.
2. This is true only for the engineering sample vehicles, which do not have the robotaxi decal applied, and which the public cannot get a ride in.
3. This is being done in the engineering sample vehicles because they are testing performance on highways now. s̶o̶m̶e̶t̶h̶i̶n̶g̶ ̶W̶a̶y̶m̶o̶ ̶a̶n̶d̶ ̶o̶t̶h̶e̶r̶ ̶c̶o̶m̶p̶e̶t̶i̶t̶o̶r̶s̶ ̶a̶r̶e̶n̶'̶t̶ ̶e̶v̶e̶n̶ ̶a̶t̶t̶e̶m̶p̶t̶i̶n̶g̶ ̶t̶o̶ ̶t̶e̶s̶t̶
FireBeyond · 1d ago
> This is not true for the existing commercial robotaxi fleet that the public can get rides in today.
The public? Must have opened things up, then, because it was only available to a select few recently (generally Tesla-positive influencers and vloggers).
The existing commercial robotaxi fleet? You mean the "10-20" vehicles in Austin? Sorry, Tesla announced that they'd increased the fleet size by 50%, so "15-30".
> This is being done in the engineering sample vehicles because they are testing performance on highways now - something Waymo and other competitors aren't even attempting to test.
Thanks for the heads up about Waymo. I've updated my original post.
TheAlchemist · 1d ago
You can also update the rest of it. There is no "existing commercial robotaxi fleet that the public can get rides in today" - it's open to pretty much influencers only, not to the public. But I suppose you know that already.
anonym29 · 14h ago
Your response is factually inaccurate. Tesla's Robotaxi is open the public:
Trying to throw people under the bus as "You are wrong and factually inaccurate" for comments made over the last few days by...
... by linking to a news article TODAY about how Tesla is STARTING to WAITLIST users of its app?
I'm sure you could be more disingenuous. But you'd have to make an effort to do so.
Also, I am completely unimpressed with a "Robotaxi" service that has someone literally sitting in the driver's seat, let alone trying to claim some form of superiority over the competition.
TheAlchemist · 12h ago
Having spend quite a bit of time of following Tesla propaganda machine on Twitter, this guy's answers are eerily similar to their usual strategies of disinformation. Most of these guys are influencers that are either paid or affiliated with Tesla, and now they have the advantage of being boosted by the platform owner.
anonym29 · 13h ago
18 hours ago is "last few days"? It's open to the public right now. Members of the public are riding in Tesla Robotaxis right now. You cannot simply assume that no members of the public are able to do this just because your account is still on the waitlist.
There's a difference between disingenuity and information advantage. You're welcome for the early tip-off.
I'm completely unimpressed with Waymo's economics.
Waymo is locking itself into economically non-viable LiDAR setups. The 5th-gen Jag iPace cars were $70k + another ~$80k in hardware, making them $150k each just to get into the same ballpark as Tesla's sub-$40k cybercabs. While the zeekr-based / gen 6 Waymos are considerably cheaper, with a base vehicle costing only $33k and equipment costs way down, they're making progress, but they're still reportedly $80k+ per vehicle, and the LiDAR is simply unnecessary.
The safest human driver on the planet only has 2 eyes and 2 ears. The existing camera + mic setup on a Tesla vastly exceeds that, and thus the upper threshold on Tesla's camera-only safety approach vastly exceeds the safest human driver on the planet even without LiDAR.
As time has shown, the enormous trove of real-world data that Tesla has continues to improve the FSD stack in the real world (not just the theoretical upper limit) over time. You don't have to like Elon Musk, or Tesla, but the fact of the matter is that FSD (despite being only level 3 autonomy at the moment) is already far safer than human drivers are, just like Waymo.
Waymo (who has only a couple thousands of cars total, and only plans to produce tens of thousands per year) will never have as much real-world data as Tesla (who has millions of cars on the road collecting training data, and continues to produce over a million per year), and thus will never breach Tesla's data moat.
In order for environmentally-friendly EV mass transit via a robotaxi network to be a viable future for the nation's transportation needs, the unit economics and resulting per-mile costs have to be attractive not just to the ivy tower millionaire tech and finance workers along the west coast and New England, but to the ordinary, working class people that coastal HN elitists love to denigrate.
The big picture goal here is not to have the best robotaxi service at any price point in 2025, it's to make convenient, safer-than-human-driver individualized mass transit (the US transportation network, for better or worse, is ultimately built for individual automobiles first and foremost, rather than trains, busses, etc) economically and environmentally sustainable for all people, not just for those on six-figure incomes.
Waymo's a cool idea on paper, and their approach to safety may be laudable, but they do not have an economic trajectory to replace personally owned automobiles. Tesla does, and that matters more to the big picture than Waymo's 2025 snapshot-in-time safety advantage over Tesla, which has and will continue to shrink as Tesla's data moat expands at an exponential rate over Waymo's.
I'm happy about this. Helping all people (including low income people) access affordable, convenient, and environmentally sustainable transportation is a laudable goal. It's disappointing that the politically left-of-center has lost track of cheer-leading this important goal to play a game of partisan tribalism that seeks to attack the entity that is doing more to advance towards this goal than any other entity.
essdas · 13h ago
Plenty in there to pick apart. Pointing out just one: Things that humans have that Teslas don’t:
1. Binocular vision
2. A brain that understands the world
FireBeyond · 12h ago
> 18 hours ago is "last few days"? It's open to the public right now. Members of the public are riding in Tesla Robotaxis right now. You cannot simply assume that no members of the public are able to do this just because your account is still on the waitlist.
They only opened the ability to even join the waitlist today.
Those comments WERE factually accurate when they were made.
> but the fact of the matter is that FSD (despite being only level 3 autonomy at the moment) is already far safer than human drivers are, just like Waymo.
Oh god, it's the FSD talking points. Based on what? Oh yes, Tesla's stats where they compare a subset of driving in a subset of roads in a subset of conditions in a subset of vehicles to "all drivers, all roads, all conditions, all weather" and come to conclusions that their statisticians know to be false because they wouldn't be statisticians if they didn't know that.
And yeah, Tesla's stats, the same ones that 1) don't count any incident without airbag deployment as an accident, even if it involves the total write off of the vehicle, or a fatality, and 2) oh yeah, any fatality accidents.
> Tesla's sub-$40k cybercabs
That don't actually exist, yet (if at all). Are they coming out before or after the $39K CyberTruck?
> Waymo (who has only a couple thousands of cars total, and only plans to produce tens of thousands per year) will never have as much real-world data as Tesla
Still hundreds of times more Robotaxis than Tesla has on the road right now.
This could actually be a decent opportunity for Tesla to branch out into manufacturing these sensors.
Of course, it's much more likely he clings to "camera only" until he dies.
Of course there are some inherent advantages to lidar sensing without Google's data muscle behind it, like sensing at night and in glaring sun. And then there's the radar and maybe sonar sensing for up-close obstacles. But, overall, I think the value of Google's data is underappreciated.
No comments yet
1. This is not true for the existing commercial robotaxi fleet that the public can get rides in today.
2. This is true only for the engineering sample vehicles, which do not have the robotaxi decal applied, and which the public cannot get a ride in.
3. This is being done in the engineering sample vehicles because they are testing performance on highways now. s̶o̶m̶e̶t̶h̶i̶n̶g̶ ̶W̶a̶y̶m̶o̶ ̶a̶n̶d̶ ̶o̶t̶h̶e̶r̶ ̶c̶o̶m̶p̶e̶t̶i̶t̶o̶r̶s̶ ̶a̶r̶e̶n̶'̶t̶ ̶e̶v̶e̶n̶ ̶a̶t̶t̶e̶m̶p̶t̶i̶n̶g̶ ̶t̶o̶ ̶t̶e̶s̶t̶
The public? Must have opened things up, then, because it was only available to a select few recently (generally Tesla-positive influencers and vloggers).
The existing commercial robotaxi fleet? You mean the "10-20" vehicles in Austin? Sorry, Tesla announced that they'd increased the fleet size by 50%, so "15-30".
> This is being done in the engineering sample vehicles because they are testing performance on highways now - something Waymo and other competitors aren't even attempting to test.
You're out of date. Waymo has been testing and is able to operate on California highways for the last year and a half (https://www.theverge.com/2024/3/2/24088454/waymo-california-...).
- https://www.cnet.com/roadshow/news/want-to-try-teslas-robota...
Trying to throw people under the bus as "You are wrong and factually inaccurate" for comments made over the last few days by...
... by linking to a news article TODAY about how Tesla is STARTING to WAITLIST users of its app?
I'm sure you could be more disingenuous. But you'd have to make an effort to do so.
Also, I am completely unimpressed with a "Robotaxi" service that has someone literally sitting in the driver's seat, let alone trying to claim some form of superiority over the competition.
There's a difference between disingenuity and information advantage. You're welcome for the early tip-off.
I'm completely unimpressed with Waymo's economics.
Waymo is locking itself into economically non-viable LiDAR setups. The 5th-gen Jag iPace cars were $70k + another ~$80k in hardware, making them $150k each just to get into the same ballpark as Tesla's sub-$40k cybercabs. While the zeekr-based / gen 6 Waymos are considerably cheaper, with a base vehicle costing only $33k and equipment costs way down, they're making progress, but they're still reportedly $80k+ per vehicle, and the LiDAR is simply unnecessary.
The safest human driver on the planet only has 2 eyes and 2 ears. The existing camera + mic setup on a Tesla vastly exceeds that, and thus the upper threshold on Tesla's camera-only safety approach vastly exceeds the safest human driver on the planet even without LiDAR.
As time has shown, the enormous trove of real-world data that Tesla has continues to improve the FSD stack in the real world (not just the theoretical upper limit) over time. You don't have to like Elon Musk, or Tesla, but the fact of the matter is that FSD (despite being only level 3 autonomy at the moment) is already far safer than human drivers are, just like Waymo.
Waymo (who has only a couple thousands of cars total, and only plans to produce tens of thousands per year) will never have as much real-world data as Tesla (who has millions of cars on the road collecting training data, and continues to produce over a million per year), and thus will never breach Tesla's data moat.
In order for environmentally-friendly EV mass transit via a robotaxi network to be a viable future for the nation's transportation needs, the unit economics and resulting per-mile costs have to be attractive not just to the ivy tower millionaire tech and finance workers along the west coast and New England, but to the ordinary, working class people that coastal HN elitists love to denigrate.
The big picture goal here is not to have the best robotaxi service at any price point in 2025, it's to make convenient, safer-than-human-driver individualized mass transit (the US transportation network, for better or worse, is ultimately built for individual automobiles first and foremost, rather than trains, busses, etc) economically and environmentally sustainable for all people, not just for those on six-figure incomes.
Waymo's a cool idea on paper, and their approach to safety may be laudable, but they do not have an economic trajectory to replace personally owned automobiles. Tesla does, and that matters more to the big picture than Waymo's 2025 snapshot-in-time safety advantage over Tesla, which has and will continue to shrink as Tesla's data moat expands at an exponential rate over Waymo's.
I'm happy about this. Helping all people (including low income people) access affordable, convenient, and environmentally sustainable transportation is a laudable goal. It's disappointing that the politically left-of-center has lost track of cheer-leading this important goal to play a game of partisan tribalism that seeks to attack the entity that is doing more to advance towards this goal than any other entity.
1. Binocular vision 2. A brain that understands the world
They only opened the ability to even join the waitlist today.
Those comments WERE factually accurate when they were made.
> but the fact of the matter is that FSD (despite being only level 3 autonomy at the moment) is already far safer than human drivers are, just like Waymo.
Oh god, it's the FSD talking points. Based on what? Oh yes, Tesla's stats where they compare a subset of driving in a subset of roads in a subset of conditions in a subset of vehicles to "all drivers, all roads, all conditions, all weather" and come to conclusions that their statisticians know to be false because they wouldn't be statisticians if they didn't know that.
And yeah, Tesla's stats, the same ones that 1) don't count any incident without airbag deployment as an accident, even if it involves the total write off of the vehicle, or a fatality, and 2) oh yeah, any fatality accidents.
> Tesla's sub-$40k cybercabs
That don't actually exist, yet (if at all). Are they coming out before or after the $39K CyberTruck?
> Waymo (who has only a couple thousands of cars total, and only plans to produce tens of thousands per year) will never have as much real-world data as Tesla
Still hundreds of times more Robotaxis than Tesla has on the road right now.