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

[Linkmeup]

They're trying to get to 1 Starship per day!

SpaceX wants to build 1 Starship megarocket a day with new Starfactory

"This will enable us to increase our production rate significantly as we build toward our long-term goal of producing one Ship per day."

www.space.com

Re-posting this here too as it relates to how much effort was put into this factory over Tesla in recent years, IMO. Elon is back in Tesla and shoring up efficiencies again.

2 years from now is going to be pretty dang epic if they reach the anywhere close to 1 Starship a day.

 

[cheqsgravity]

Ok, I would like to ask the many experts here a question related to part of the current discussion. How do you define AI? To a doofus who read way too much Asimov in the 40s and 50s, I think of AI as being sentient. When Mr Musk said that he had realized that when Tesla had solved FSD, they would have solved AI, I thought that was what he meant. When I try FSD on my car and it works, I sometimes get the feeling it is human like. When Tessie drives like a human in the future, will she pass the Turing test? And other tests?sorry to interrupt the flow of the thread, but I would like a definition.

IMO, AI is what it stands for Artificial intelligence. The human brain for the most part is a black box. There are have been several leaps made in understanding the human brain. But not entirety. AI is similar as in its a black box. In Tesla/EMs words its video in/control out. Increasingly decisions made by AI will become harder and harder to reason/decipher.
Having stated the above, the best way to look at AI is through the lens of evolution. So the human brain is one of the largest brains in the animal world. Its complexity and energy needs are supreme. This is what chatgpt says about human vs chimp

So b/c of structural and energy differences, chimps are limited in what they can do. AI will be several levels (as of now no limit) above human brain. Currently it can do certain things better w/ inputs. Like my cars on FSD will see pedestrians when i don't see them (esp at night). Or will go really quickly to next lane to avoid car coming in b/c it has already checked blind spots to see target lane is clear. But capabilities will start increasing over time. Once LLMs get introduced in cars, they will start conversing w/ us. And the list will just grow from there. Bots are a natural progression from there.
So yes AI will eventually become sentient like Robert e Olivaaw. And yes IMO i can see them being heads of state (which i very much prefer to the current bunch we have). But all this is contingent on the evolution of AI. They can evolve into a caring and nurturing sentience. That is why what google did w/ gemini is shocking. EM stated AI should be made to be truth seeking and curious and I tend to agree with him. No one really knows how AI will evolve especially when geo political considerations are taken. A possibility that it will become nurturing (don't know how much weight the possiblity is). But it can also go awry. But IMO, the age of the dominance of human intelligence is close to an end. As ilya of open ai put it, it will be like we treat pets. We don't go around killing our pets. We nurture them. But we don't ask for their opinion when we need to build a highway

Also read Iain M. Banks culture series. I read those after the foundation series and robot short stories. The player of games is a good one. If things go right, good chance our world/universe can be like that: Amazon.com

 

[EVMeister]

If he wants his package, he should probably quit posting stuff like this. He looks like an ass and a crazy MAGA person.

https://twitter.com/x/status/1799417761539805431

Irrespective of his pay packet, he should be more selective about what he shares with the world on X... It's not a trait I admire him for, but for better or worse it's part of his DNA.

 

[mrmage]

I have no qualms about your opinion of Musk as CEO. People have different opinions and I respect that.
I have not been following this entire discussion but I want to share something that may provide an alternate perspective to Musk's words.

When Musk made the very specific proclamation 7 and 1/2 years ago of US coast to coast driving - you called that a mistake.
Now he is being vague (when all known bugs . . ); you seem upset about this vagueness .
You go on to say, "a man who does not learn from his mistakes" . . . but perhaps he has learned.

When choosing between saying something likely to be wrong or something misleading, one can choose to say nothing at all.

 

[cliff harris]

A few thousand dollars of sensor costs annually (or much less in future years) is not an "insane" burden.

You think?
How many trips does each waymo car do a day? 10? lets say 300 trips a month, lets say 3,000 trips (very generously) in a year, given maintenance, quiet periods, etc.
So yup, a few thousand dollars a year is going to still be a dollar per trip. And thats without the fact that cameras require TRIVIAL maintenance (just wipe em!) and have almost 100% uptime. LIDAR has rapidly moving parts and a lot to go wrong, and draw more power.
waymo's vehicles are like the SLS jobs-program rocket designed by Boeing/lockheed. Lots if impressive looking and very impressive technology, but in terms of competing against starship, a complete joke.

LIDAR and other sensors means more to install, more to go wrong, more to power, more to maintain, more to deal with in sensor fusion software. Its about as sensible an approach as Toyotas dogged insistence that hydrogen vehicles will be a thing.

I will eat an entire humble pie the day waymo can announce that they are profitable, in multiple cities, in multiple countries, in all times of day, in all weather.

 

[Christine69420]

If he wants his package, he should probably quit posting stuff like this. He looks like an ass and a crazy MAGA person.

https://twitter.com/x/status/1799417761539805431

So people who are perceived as asses should not expect to get their pay? And should expect deals to be broken?

And people who are seen as being on the far right politically should expect the same?

Who decide when people fall into these categories?

Disclaimer: I do not live i the US and may miss some context

 

[mrmage]

https://twitter.com/x/status/1799272436132614343

FSD 12.4.1

Very eager to see how well this works. From what I've seen from 12.3, at least 4 more nines are needed, probably 5.

For FSD to be ready in 2 years, that's 2 nines a year, implying this release needs to increase city miles between disengagements by a factor of 5, since it took 4 months to come out.

 

[Snerruc]

IMO, AI is what it stands for Artificial intelligence. The human brain for the most part is a black box. There are have been several leaps made in understanding the human brain. But not entirety. AI is similar as in its a black box. In Tesla/EMs words its video in/control out. Increasingly decisions made by AI will become harder and harder to reason/decipher.
Having stated the above, the best way to look at AI is through the lens of evolution. So the human brain is one of the largest brains in the animal world. Its complexity and energy needs are supreme. This is what chatgpt says about human vs chimp

View attachment 1054665
So b/c of structural and energy differences, chimps are limited in what they can do. AI will be several levels (as of now no limit) above human brain. Currently it can do certain things better w/ inputs. Like my cars on FSD will see pedestrians when i don't see them (esp at night). Or will go really quickly to next lane to avoid car coming in b/c it has already checked blind spots to see target lane is clear. But capabilities will start increasing over time. Once LLMs get introduced in cars, they will start conversing w/ us. And the list will just grow from there. Bots are a natural progression from there.
So yes AI will eventually become sentient like Robert e Olivaaw. And yes IMO i can see them being heads of state (which i very much prefer to the current bunch we have). But all this is contingent on the evolution of AI. They can evolve into a caring and nurturing sentience. That is why what google did w/ gemini is shocking. EM stated AI should be made to be truth seeking and curious and I tend to agree with him. No one really knows how AI will evolve especially when geo political considerations are taken. A possibility that it will become nurturing (don't know how much weight the possiblity is). But it can also go awry. But IMO, the age of the dominance of human intelligence is close to an end. As ilya of open ai put it, it will be like we treat pets. We don't go around killing our pets. We nurture them. But we don't ask for their opinion when we need to build a highway

Also read Iain M. Banks culture series. I read those after the foundation series and robot short stories. The player of games is a good one. If things go right, good chance our world/universe can be like that: Amazon.com

Thanks. So, if this is the pathway, we are in early innings, correct? I think this makes Asimov all the more important. We are on a potentially very dangerous path. I for one don’t want HAL.

 

[CHGolferJim]

I voted "yes" for the compensation plan, both in 2018 and again last month, but there is a part of me which agrees with you as well. For the past year or two Elon has not impressed me much with some of his decisions:

- I don't like that he bought Twitter nor how he bought it.
- I don't agree with his slowing Tesla auto production down to focus resources on Robotaxis before FSD is ready.
- I certainly don't agree with some of the more controversial things he says, and I do believe his antics have hurt the Tesla brand.
- I don't like how he fired the SC team and is now slowing down SC deployments.

That said, I also believe he's better at running Tesla than I am or could be, and I think (for now) Tesla is better with him than without him. I also believe he earned the 2018 compensation plan and should be allowed to keep it. Even though I don't feel he's deserved it for the past few years, a deal is a deal and he earned it before then so I voted yes.

But I can certainly see why so many are voting "no" as well. It's hard to argue against many of their reasons, Elon has made the arguments for them.

Next week is going to be a roller coaster. No matter which way this vote plays out we are in for a rough ride in its aftermath.

Agree with your well said post. I voted with the board on all points except Kimbal. I believe shareholders should stand up for Elon vs. the DE court, and support him to receive the incentive plan we previously voted for. However, it would not bother me at all if it got rejected by shareholders. A new plan doesn’t have to yield the same net value to Elon (past earnings), and he should be required to spend a quantifiable majority of his future time at Tesla and be restricted in how he can sell shares given his massive position and the potential harm to shareholders from the careless process used a few years ago.

 

[Leftlane22]

Hypothetical alternate scenario:

Let’s say back in 2018, instead of this particular comp plan, the demand was that Tesla must buy Twitter in its entirety and give it to Elon, or else he will not focus on Tesla any more and will develop technologies at his other private companies instead. Seems totally ridiculous, but is essentially what happened here. He had the percentage ownership he now wants, but he felt the need to sell shares and spend $44 billion on that piece of garbage, taking his ownership down. Now, he wants the 25% stock ownership again, which would cost more than what he spent on Twitter. The board should have never gone along with something so egregious, even if the goals were completely pie in the sky. But they alsways do whatever he wants, and that’s the problem. Okay, he earned it based on the metrics - but it never should have been the reward in the first place. If he wanted that share ownership level he could have passed on Twitter and focus on the one company he runs that is, you know, actually public and has shareholders.

Would you have voted yes for that plan?

 

[2Pearls]

Mid-summer* break from your favorite Moderator’s regularly-scheduled hiatus from 24h/d riding herd over y’allz -

If a CEO of any company for which I sat on the BoD ever said to the Board words analogous to:

We are starting to get to the point where, once known bugs are fixed, it will take over a year of driving to get even one intervention

He or she would be 37th floor defenestrated before he could stop for a water break.

I don’t care about “starting to get”; I care about getting.

I don’t even care about “getting to the point”; I care about being at the point.

I don’t care about “once known bugs are fixed”; I care about all bugs having been fixed.

For the umpteenth time on this thread, words matter. And those words demonstrate that we have just been told absolutely nothing at all.

Disgraceful.

*Last frost today! We hope. And I think I plowed snow for the final time this season last week.

There are always people who are fanbois and keep blinders on...they are usually followers and react too late when the sh*t really hits the fan. We can be owners of TSLA stock, love the companyTesla, and not agree with Musk 100% of the time. Remember, free speech!

 

[mrmage]

You think?
How many trips does each waymo car do a day? 10? lets say 300 trips a month, lets say 3,000 trips (very generously) in a year, given maintenance, quiet periods, etc.
So yup, a few thousand dollars a year is going to still be a dollar per trip. And thats without the fact that cameras require TRIVIAL maintenance (just wipe em!) and have almost 100% uptime. LIDAR has rapidly moving parts and a lot to go wrong, and draw more power.
waymo's vehicles are like the SLS jobs-program rocket designed by Boeing/lockheed. Lots if impressive looking and very impressive technology, but in terms of competing against starship, a complete joke.

LIDAR and other sensors means more to install, more to go wrong, more to power, more to maintain, more to deal with in sensor fusion software. Its about as sensible an approach as Toyotas dogged insistence that hydrogen vehicles will be a thing.

I will eat an entire humble pie the day waymo can announce that they are profitable, in multiple cities, in multiple countries, in all times of day, in all weather.

A full time driver makes 10 to 20x more than that. Eliminate that cost and there's lots of extra margin left. Plus, sensor costs keep going down with volume. Just read that LIDAR is expected to drop to $100 in volume. So probably just hundreds of $ amortized cost in the future.

As for maintenance, a car has thousands of moving parts. They're not dropping like flies, so I'm sure their sensor array and lidar will be fine. Google is not clueless on how to mass produce products.

Btw, consider what you're implying if a cost of 3 figures or even low $1k annually is a deal breaker. Assuming $50k capital upfront, annual ROI of couple %?

Either robo TaaS is wildly profitable, in which case an extra annual $1k cost is fine, or it's so unprofitable that Tesla's strategy is doomed. Can't be both. Which is it?

 

[Doggydogworld]

Waymo's problem is that there just is no clear path to profitability. Waymo's hardware costs are too high and more importantly, Waymo's operational costs are too high.

You don't know their costs.

That's why Waymo won't have 50k robotaxis any time soon.

Simple extrapolation says 3 years to 50k, so summer 2027. When do you predict Tesla will have 50k driverless robotaxis in service?

It's not at all true that each Waymo car replaces a normal car plus its driver. Each Waymo car requires remote monitoring, remote intervention, and sometimes a relief driver. Those all require a lot of people with relatively higher salaries than an Uber driver.

How will Tesla will avoid these costs?

I think Tesla will start its robotaxi business quite soon, like next year.

They can start a service with safety drivers today. I'm sure they will on 8/8, along with multiple grandiose claims.
Make a real prediction. How many paid driverless rides will they provide in December 2025? Waymo should do ~3 million that month.

Tesla's cost structure will always be far below Waymo's.

Why? H/W cost? What's the incremental h/w cost per mile? A nickel? A dime? What will it be in two years? Decent solid-state lidar is already sub-$500 with a specification war driving rapid improvement.

 

[spacerfc]

He no longer has my vote of confidence and I would be happy to see his replacement at the helm, sooner rather than later.

It's beyond ridiculous to claim Tesla investors should replace a CEO who just accomplished this:

https://twitter.com/x/status/1798792222743122164

Get this: Tesla is nothing without Elon Musk, without him Tesla is just a bigger Rivian, and one would be better off to just buy RIVN since it's a lot cheaper.

And who do you plan to replace him with? Trevor Milton? Peter Rawlinson?

 

[mrmage]

So people who are perceived as asses should not expect to get their pay? And should expect deals to be broken?

And people who are seen as being on the far right politically should expect the same?

Who decide when people fall into these categories?

Disclaimer: I do not live i the US and may miss some context

Happens all the time. Lots of people get fired or reprimanded by their boss, when they say something that might violate their company social media policy. Many companies would have fired employees for some statements that Elon has made.

For CEO though, that duty and responsibility would also fall to the board. Of course, this board would never dream of that.

Which is related to this issue, because Delaware essentially said that the board was not doing its job, and therefore the pay package was invalid. Because the package would not have been excessive in the first place if the board had done its job.

 

[AudubonB]

It's beyond ridiculous to claim Tesla investors should replace a CEO who just accomplished this:

https://twitter.com/x/status/1798792222743122164

Get this: Tesla is nothing without Elon Musk, without him Tesla is just a bigger Rivian, and one would be better off to just buy RIVN since it's a lot cheaper.

And who do you plan to replace him with? Trevor Milton? Peter Rawlinson?

I have zero intention of desiring him to leave SpaceX.

 

[jerry33]

Happens all the time. Lots of people get fired or reprimanded by their boss, when they say something that might violate their company social media policy. Many companies would have fired employees for some statements that Elon has made.

For CEO though, that duty and responsibility would also fall to the board. Of course, this board would never dream of that.

Which is related to this issue, because Delaware essentially said that the board was not doing its job, and therefore the pay package was invalid. Because the package would not have been excessive in the first place if the board had done its job.

The pay package wasn't excessive in view of the perceived difficulties of success. I don't believe any BoD at the time would believe that Elon could actually accomplish everything stated. Hindsight is of no value unless you have a time machine.

 

[The Accountant]

CYBERCAB FINANCIALS
Until now, I have never performed an analysis on the potential financial impact of the Robotaxi/Cybercab.
The cybercab was excluded from my long range estimates.

I completed a high level analysis recently. I took Uber's 2023 financials and adjusted them as if they ran a robotaxi business.
I stripped out driver payments, added costs such as charging, maintenance & repairs, Insurance, back office logistics, etc.

In this analysis, Uber's EBITDA went from $5B (25% of sales) to $30B (50% of sales).
Note: Sales are higher in the robotaxi model as Uber's revenues exclude the portion of billings given to drivers.
EBITDA= Earnings before Interest, Taxes, Depreciation and Amortization

I was surprised by these results and as I dug into it i realized the following:
Since everything is practically outsourced at Uber, profit is shared with:
- the driver
- the gas station
- the auto insurance company
- the mechanic
- the car wash

When you have no driver and you have your own infrastructure (service centers, charging stations, insurance, etc), you keep a higher percentage of the end to end profit. Of course the robotaxi model requires a huge ongoing capital investment with the vehicles, but in Tesla's case, they mfg the vehicles and thus acquire them "at cost" - no markup to share there either.

I'm not ready to share the analysis but hope to prior to the Aug 8 event.

 

[DarkandStormy]

https://twitter.com/x/status/1799345802202210601

I thought the inventory discounts ended so as to simplify the sales process?

 

[sunwarriors]

Saw this recent article about Australia inventory buildup:

Tesla 'graveyard' photo shows major shift in Aussie EV market

The dramatic images show a major change trickling through the EV market – and it's a win for consumers.

au.news.yahoo.com

I've always felt SpaceX success is more due to Shotwell who is 100% there. She has 21 direct reports including the Starlink team. Maybe Tom Zhu could be groomed to eventually be the COO or something similar/higher profile as how SpaceX is structured and Elon can step back with more vision CEO stuff.

That said, SpaceX has limited benefit to individual Tesla investors since no matter how SpaceX does, it's private and you aren't an owner. Similar to if xAI does insanely well, it does no one here any good in terms of direct financials.