As it was a recruiting event with designers and engineers, I just thought there'd be more in-depth design process + technical questions.
All this stuff is beyond me, I like to listen in to be wow'd and participate in being a fanboi.
-
Want to remove ads? Register an account and login to see fewer ads, and become a Supporting Member to remove almost all ads.
Tesla, TSLA & the Investment World: the Perpetual Investors' Roundtable
- Thread starter AudubonB
- Start date
As it was a recruiting event with designers and engineers, I just thought there'd be more in-depth design process + technical questions.
All this stuff is beyond me, I like to listen in to be wow'd and participate in being a fanboi.
Paracelsus
Active Member
Can someone ask him about their plans for mass production? Is it at a gigafactory? What are the plans on producing these hardware quickly? People need to stop asking him about AGI where Elon just goes on a long tangent.
Krugerrand
Meow
No, that’s the biggest innovation for your pocketbook. The biggest innovation is how it all changes the way we live, what the future becomes for mankind.
Krugerrand
Meow
What’s to ask? They told us in depth how they got to this point; failures and all. There really aren’t any questions left to ask unless you (general) weren’t paying attention.
This was predictable.
What is not considered is the cost of those other Robots, how easy they are to manufacture, how much time has been spent on their development, And if they are near some local maximum where their current set of tricks is close to the limit of what they can do. Or each new trick requires many hours of manual training by s skilled operator.
From my point of view the Optimus demo was more or less what I expected, the Optimus hardware was a good design, and overall progress was good.
Dojo and FSD progress was also good.
The FSD part of the presentation sure was detailed and technical, perhaps too long, detailed and technical, but we were warned in advance to expect a detailed presentation.
When experts cut the FSD presentation up into segments and analyse them, we will learn more, and it will make more sense.
For people with the right skills this looks like a great team to join, working with highly intelligent, knowledgeable and hard working people, on interesting challenges, is a great experience. It seemed to me Tesla was close to the leading edge on all aspects, but in particular willing to commit to what it took to get results.
I'm sorry, but I have no idea - again this is all over my head - but I was hoping to be wow'd, as a layperson, by the people asking questions who are supposed to be great designers and engineers (so much so that Tesla is actively bringing them on-site to recruit them). Also, to be wow'd, which I was, by the answers provided to some of the questions by the Tesla team.
It's a different presentation and environment than the shareholder meetings which include the usual PETA question, as an example.
Krugerrand
Meow
What do you mean what are the plans for producing these? Isn’t it obvious?
They’ll do it just like they did for the first cars. When they get to an iteration that’s good enough, they’ll build a production line in one of their 6? factories at that time and start making them, slowly increasing the rate of production, increase efficiencies, reiterate the production line, continue to improve the product with regular engineering changes and OTAs etc….
Your question does not compute.
The questions were unscripted.
From my point of view, 20% of questions were very good, and 20% of questions were totally pointless, many other questions of dubious value.
IMO Elon was slightly optimistic about the quality of the questions they would get, but on another day they might have got better questions.
I would not want to see plants in the audience asking pre-scripted questions, so we need to take our chances with the naturally occurring questions.
In terms of mass production or Optimus, they are designing it for mass production, nothing in the presentation seemed to me like a plan to "go slow".
3-5 years for delivery is because Tesla wants to use most of the initial production in house, and for once Elon is perhaps being cautious about promising delivery dates.
Last edited:
GrandEnigma
Member
What I didn’t hear asked was what are they looking for? More of the same? People to fill specific weaknesses? Seems that would have been an important question to ask.
Oh wow...they are going to build a...PRODUCTION LINE? You don't say.....
I want to know, are the actuators casted? Have they already build the machine that makes the actuators? Is this a raw material in, optimus out type plan? What is the size of the assembly line for this and how automated are we talking about?
It seems that they can pump these robots out with a higher margin than cars if they were to sell it for 20k each.
Elon gave his typical answer to a question about Tesla being like a whole bunch of startups, but really I think this presentation more than any other I have seen from Tesla was all about the massive advantage they have from integration amongst the R&D teams.
Robot simulation was done in Tesla car crash simulation software. Robot vision using the same software as FSD. Ditto neural net training etc. And training to soon be on teslas own hardware, which will eventually becomes its own product segment as a service. Main first customer for these new products is tesla itself.
I found it most curious that Elon went out of his way more than once to point out he doesn’t ultimately control Tesla, and that he could be fired if shareholders really wanted him gone. I wonder what was behind that.
Robot simulation was done in Tesla car crash simulation software. Robot vision using the same software as FSD. Ditto neural net training etc. And training to soon be on teslas own hardware, which will eventually becomes its own product segment as a service. Main first customer for these new products is tesla itself.
I found it most curious that Elon went out of his way more than once to point out he doesn’t ultimately control Tesla, and that he could be fired if shareholders really wanted him gone. I wonder what was behind that.
So Elon made a point to make the Q&A extra long but also made sure people wouldn't have any questions because any question asked will be stupid when it's already answered in the presentation?
Elon wanted to reassure people who might be worried abut the Robot being used in ways which harm humans. But it is hard to give a level of reassurance that will satisfy everyone. And if Tesla doesn't do robots others will.
Krugerrand
Meow
Much of it is over my head as well, but I got enough of a gist to know - Holy donkey balls, Batman!
Nobody else is moving this fast. Where many find the mistakes and failures along the way a negative, I see them for what they were - an opportunity to learn at an accelerated rate, changing direction repeatedly until you’re facing the right way and then onto the next problem. Unless you are a creative problem solver at heart, you wouldn’t get the excitement and satisfaction of learning these people are living.
I thought there were a couple of insightful questions, but mostly people playing at being smarter than the folks on stage. That’s why they’re in the audience and not working at Tesla already.
Don’t worry that this may not be received well by WS or even many here. Everything is on track at Tesla and moving at full speed. Serious poo is happening. Buy with impunity. Hold like your life depends on it.
Bet TSLA
Active Member
Well, sure it's obvious, but you left out the critical difference from all their production lines done in the past. This time the ramp won't be complete until they are producing Optimus Prime with no humans on the line, all of them replaced by Optimus Prime. That's how they know it's good enough.
Krugerrand
Meow
Don’t be ridiculous. Q&A is an opportunity for people to get clarification (which they did) or to ask about the human aspect of Optimus (which they did) and other things that weren’t touch upon - you know like the question about semi.
There are so many questions no one clarified. Like how is training done on optimus that scales? Right now it seems like they are doing motion capture and that does not scale. How will optimus understand intent? Planning and logic is perhaps FSD's highest on the needs improvement list. Also how does driving video data translate to usable data for a robot geared to do completely different tasks?
Krugerrand
Meow
Those are entirely different questions than what you first proposed - specific vs general - and logically these ones are likely a bit ahead of where they are at this point. They likely have some ideas. They’d be making the parts by hand at this point, MAYBE seeking out some suppliers but knowing Elon - planning to do as much of it as possible themselves to be able to control costs and quality, mitigate risks etc…
Nvidia's H100 is suppose to be 6-30x better than A100(I don't know, it's probably lots of marketing and very application specific). So it should provide ...X amount of performance increase for the same cost like Dojo. A good question would be if the H100 will be purchased by Tesla or how it compares to that when it comes to cost savings.
Similar threads
- Locked
- Locked
- Locked
