Welcome to Tesla Motors Club
Discuss Tesla's Model S, Model 3, Model X, Model Y, Cybertruck, Roadster and More.
Register
This site may earn commission on affiliate links.
Waymo and Cruise already have functional robotaxis, and you're saying it's still going to take Tesla years to get there? Huge lead Cruise / Waymo can build in major metros.
"Fit bit already has health monitoring watches, and you're saying it's still going to take Apple years to release an apple watch?"
"Diamond already has mp3 players, are you saying it's still going to take apple years to release a mp3 players?"
"Nissan already have affordable EVs, are you saying it's still going to take Tesla years to release the model 3?
"Motorola already have smart phones, are you saying it's going to take apple years to release the iphone?"

I can go on and on.....

The product waymo and cruise provide sucks. Limited, expensive, and slow. You know the perfect ingredients needed for a competent mode of public transportation...said by no one ever.
 
So I tentatively believe a pretty advanced L2 system will come from V12 sometime early next year. Robotaxi level is still hard to estimate simply because of accuracy thresholds needed - there still may need to be a lot of iterations to get it to a useable point.
Most of those iterations will simply be retaining using existing data.

The key IMO is of Tesla has all of the data needed to train for Robotaxis, or can quickly source additional data when interventions occur.

Assuming all of the data is present, it is a matter of cranking the training flywheel.

It then comes down to how long between versions? They may need to drive a version of software around for at least 1 week to find most of the interventions for that version.

The next unknown question is how many V12 versions are needed, and what is the rate of improvement?

They could have up to 20 rounds of V12 training on HW3 between now and the end of the year.

Will V12 be better than Cruise and Waymo by the end of the year?

Clearly better than Cruise and Waymo could be the trigger for starting the regulatory approvals process, Approval might take another 12 months.

If they need fairly major architecture revision, then I agree that blows the timeline out. That many be like to dropping back to training round 1 after say 15 rounds of training. IMO because there is no need to change written software, this is likely to be faster than previous pivots.
 
I think I already saw that in ‘The Terminator’. Works well.

Is it fate that the next installment of "The Terminator" francise will be v11?
  1. The Terminator
  2. Terminator 2: Judgment Day
  3. Terminator 3: Rise of the Machines
  4. T2-3D: Battle Across Time
  5. Terminator Salvation
  6. Terminator Salvation: The Machinima Series
  7. Terminator: The Sarah Connor Chronicles
  8. Terminator: Dark Fate
  9. Terminator Genisys
  10. Terminator Genisys: Future War
I'll be back! -- James Cameron

P.S. Elon did not 'miss-a-meme' by not choosing August 29th for the 1st AI to become self-aware: Instead it was 2600.72% of a common year... ;)
 
So looks like I'm called out... :)

In Japan, there are two address systems. Make it A and B.
System A (Chiban system) identifies the owner of the land. You can't use it to send cars or deliver mail in some locations because one address could be one mountain or a vast rice field with many houses (owned by one person).
The system B (Banchi system) identifies the buildings. You can use it to deliver mail. However, there is no system-B address on the land without structures.

So, in most urban areas, system-B addresses are used, whereas in other rural regions, system-A address is used. If you own land with a system-A address of xxx-1 and sell half of the land, the sold land will have the address of xxx-2. At the same time, if there is a new building on the xxx-2, you will need to talk to the local government to reorganize and assign a new system-B address to the new building.

Google Maps seems to recognize system-A addresses in rural areas and system-B addresses in urban areas.
 
Thanks, @hiroshiy.

As the one who brought this up, am I being too harsh were I to suggest “address” does not mean the same thing - or rather perhaps I should write “does not bring to mind the same concept “ - in Japan as it does in…many…other parts of the world?
 
So I tentatively believe a pretty advanced L2 system will come from V12 sometime early next year. Robotaxi level is still hard to estimate simply because of accuracy thresholds needed - there still may need to be a lot of iterations to get it to a useable point.

Can you share more what you believe the features of this pretty advanced L2 system are going to look like?