Welcome to Tesla Motors Club
Discuss Tesla's Model S, Model 3, Model X, Model Y, Cybertruck, Roadster and More.
Register

Tesla, TSLA & the Investment World: the Perpetual Investors' Roundtable

This site may earn commission on affiliate links.
This makes it even more clear that nobody can possibly compete. Its bad enough that you need terrabytes of specially collected video footage to even train reasonable FSD, but if the last 1% of perfection can only use a trivial percentage of that data, then its even more essential that you have it.
Unless another company sells >1 million EVs with over-the-air video data collection and builds up $10billion in computing and networking infrastructure to support it, then they absolutely will not be able to offer FSD.
Tesla could absolutely screw every dollar out of FSD licensing deals with other companies, and there is nothing they can do about it.
Wouldn't the fact that Waymo is able to run cars in L4 with no major issues be a counterpoint to that?
 
FSD issue is now finding the 99 %



*In specific carefully mapped areas with a remote operator ready to take over when the car messes up or gets stuck (and they do).
So why isn't Tesla running robotaxis with the same conditions?
I find Elon's comment that 'interventions are so rare' to be disingenuous. He obviously isn't seeing my car's data, and it sounds like he gets his data from watching the choir preach on X
 
*In specific carefully mapped areas with a remote operator ready to take over when the car messes up or gets stuck (and they do).
We've been over this a number of times. They're obligated to report how many times that happens and it's roughly once in 17000 miles, as per official data submitted to the DMV. I think that's pretty good.

PS: As per the information available and which was posted on this forum a number of times, the remote operators can't control the car directly. Also each of their intervention counts as a disengagement.
 
We've been over this a number of times. They're obligated to report how many times that happens and it's roughly once in 17000 miles, as per official data submitted to the DMV. I think that's pretty good.

PS: As per the information available and which was posted on this forum a number of times, the remote operators can't control the car directly. Also each of their intervention counts as a disengagement.
We've also gone over the fact comparing Waymo's "on rails" solution to Tesla's "go anywhere" solution is comparing apples to oranges, yet here we are.
 
So why isn't Tesla running robotaxis with the same conditions?
I find Elon's comment that 'interventions are so rare' to be disingenuous. He obviously isn't seeing my car's data, and it sounds like he gets his data from watching the choir preach on X
He's getting his data from the entire pool of data that Tesla processes. A lot more insight than any individual tester has. It is a work on progress and it currently works better in some areas than others, but it has far more potential than Waymo's system. They're fundamentally different approaches.
 

^China sales


April EU sales do not look good. Down QoQ and YoY.

Good plot with long term comparison my piloly (also see the following tweet for more):

 
We've also gone over the fact comparing Waymo's "on rails" solution to Tesla's "go anywhere" solution is comparing apples to oranges, yet here we are.

I don't think it's valid to compare a solution that's currently level 4 with one that's level 2, but which promises to jump straight to level 5. Promises are cheap. Also, without knowing Waymo's algorithms in detail, nobody can say how specialized / how much high definition mapping their solution needs. It's virtually impossible to say how a Waymo would behave if taken outside of its geographical limits and how many miles it would do before critical disengagements.
 

^China sales


April EU sales do not look good. Down QoQ and YoY.
Last year same April week 5900 this year 10900 for me it's not bad sorry !
 
  • Informative
Reactions: UkNorthampton
So why isn't Tesla running robotaxis with the same conditions?
I find Elon's comment that 'interventions are so rare' to be disingenuous. He obviously isn't seeing my car's data, and it sounds like he gets his data from watching the choir preach on X
Because half baked robotaxie is a money losing business? Look at exhibit A. Waymo and exhibit B. CRUISE.

I say if the area is well mapped, I am seeing zero disengagements myself. I don't think I have ever disengaged since v12.3.4 due to traffic manipulation. It's been a few weeks.
 
Last year same April week 5900 this year 10900 for me it's not bad sorry !
It's getting to the point where China data no longer build any trend. If you are using it to judge demand, it introduces more noise than signal. One week it's 2000 and next week is 10k. Then 4k then 16k. It's like a bunch of random nonsense now.
 
I don't think it's valid to compare a solution that's currently level 4 with one that's level 2, but which promises to jump straight to level 5. Promises are cheap. Also, without knowing Waymo's algorithms in detail, nobody can say how specialized / how much high definition mapping their solution needs. It's virtually impossible to say how a Waymo would behave if taken outside of its geographical limits and how many miles it would do before critical disengagements.
This just proves how restrictive it is to think in terms of "levels" of autonomy. "Promises are cheap"... Yeah, so is a hypothetical situation where a Waymo is taken out of its comfort zone and performs well. It's abundantly clear those very restrictions exist because the system is not designed to operate outside of its set parameters or specific areas. Their system is based on well-defined rules in carefully mapped areas, let's not try to pretend otherwise.
 
If you understand the limitations of the approach they're taking this is obvious.

It's pretty clear that nobody on this forum (and only a handful of people that post online) have the required background and inside knowledge required to comment with such certainty. Just think for a second that if they were that tied to pre-processing, they wouldn't be able to handle any situation that hasn't already been pre-mapped - traffic cops, construction etc.
 
Allow me to discard your opinion, unless you're an expert in AI and self-driving, your opinion is merely based on hopes and dreams. :)

Tesla will be soon, starting with safety drivers. Within a year of driverless rollout, Tesla's system will outperform Waymo by a large margin.

And when do you think the driverless rollout will happen? Just so I can set a reminder to come back to this topic.