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

Announcement: Tesla AI Day Watch Party - Friday at 5:30PM PDT

This site may earn commission on affiliate links.
Last edited:
Happy Watch Party Day. Here are some topics that might be at the party. Be There or Be Square…



 
Last edited:
  • Like
Reactions: danny
Does anyone at Tesla speak English? I couldn't understand a single word anyone said.
Screen Shot 2022-10-01 at 6.23.00 AM.png
Screen Shot 2022-10-01 at 6.22.23 AM.png
 
  • Funny
Reactions: 2101Guy
I enjoyed it. I loved the optimism of the Tesla team. The dojo guy says they hope with the next version of dojo it will be 10x faster. Will be interesting next year's episode when they have HW4 they can talk about.
The robot was very unimpressive. Looked like boston dynamics from more than 5 years ago. What is impressive is Elon's will to succeed. So like Elon said 3+ years down the road it will be impressive.
With fool self driving, I think the theme at the presentation was fail fast. They will continue to fail fast until they succeed.
Also thought it funny their sim environment. They were touting it heavily. Just a couple of years ago, Elon was saying they don't need a sim environment because the real world environment is better.
 
Last edited:
I enjoyed it. I loved the optimism of the Tesla team. The dojo guy says they hope with the next version of dojo it will be 10x faster. Will be interesting next year's episode when they have HW4 they can talk about.
The robot was very unimpressive. Looked like boston dynamics from more than 5 years ago. What is impressive is Elon's will to succeed. So like Elon said 3+ years down the road it will be impressive.
With fool self driving, I think the theme at the presentation was fail fast. They will continue to fail fast until they succeed.
Also thought it funny their sim environment. They were touting it heavily. Just a couple of years ago, Elon was saying they don't need a sim environment because the real world environment is better.
Elon says lots of things he usually gets wrong or backtracks on later.

The next one Im watching is this statement:
  • We will never surrender/settle an unjust case against us, even if we will probably lose.
And will see if that holds up with the twitter case. Since he has this new, hardcore litigation team and all. Lets see them fight that case in court. NO SETTLEMENT's ELON!
 
Nothing new on "single stack" , and when the FSD beta stuff will trickle down to normal AP?
Most of the less technical, FSD-related news came in the Q&A session:
  • Single-stack to release in Oct
  • Improved cross traffic speed estimation in Oct
  • Parking lot capability and, presumably, autopark improvements in Oct
  • Musk doubled down on vision-only in an answer to a question on sensors for Tesla semi.
Of course, as with any Tesla timelines, take with a grain, or two, of salt.

Presentations were, as expected, highly technical with the goal of attracting new engineering talent.
 
Most of the less technical, FSD-related news came in the Q&A session:
  • Single-stack to release in Oct
  • Improved cross traffic speed estimation in Oct
  • Parking lot capability and, presumably, autopark improvements in Oct
  • Musk doubled down on vision-only in an answer to a question on sensors for Tesla semi.
Of course, as with any Tesla timelines, take with a grain, or two, of salt.

Presentations were, as expected, highly technical with the goal of attracting new engineering talent.
OK so basically everything FSD he was asked about is going to be available this month? LOL, cmon.
 
The robot was very unimpressive. Looked like boston dynamics from more than 5 years ago. What is impressive is Elon's will to succeed. So like Elon said 3+ years down the road it will be impressive.

Boston Dynamics has been spending a ton of their time on fluid kinematics. But their robots require lots of heuristic programming to be useful. As stated at AI Day 2, Tesla is aiming for 'as useful as possible as quickly as possible.' So we know Tesla is never going to program their bot to do a back flip unless that is somehow useful. This is the classic form (BD) vs function (Tesla). Most people are going to judge the bots on how pretty they walk because that's all their intellect can handle.

With fool self driving, I think the theme at the presentation was fail fast. They will continue to fail fast until they succeed.

This is just standard Agile practice. The general idea is to work in short iterations so that you can be adaptable to market changes, requirement changes, and of course failures.


Also thought it funny their sim environment. They were touting it heavily. Just a couple of years ago, Elon was saying they don't need a sim environment because the real world environment is better.

Not a correct characterization of their past stance on sims. Musk said there's a place for sims, but warned of an overreliance on sims. Not sure they've changed their stance on this, since for both AI Day 1 & 2, they spoke of sims in context of being particularly useful in 1) rare situations that would take too long to encounter in the real world; 2) situations too dangerous to test in the real world.
 
...
Also thought it funny their sim environment. They were touting it heavily. Just a couple of years ago, Elon was saying they don't need a sim environment because the real world environment is better.
Looks like I didn't remember it quite correctly. Here is what I found related:
The simulation programs engineers use to run their virtual self-driving cars through millions of “edge case” scenarios to rack up billions of driving miles are good but not good enough, Musk said. “We have quite a good simulation, too, but it just does not capture the long tail of weird things that happen in the real world.”