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

[Bobfitz1]

To be fair to Elon, he recanted what he said in the earlier interview in the Faber interview - when Faber asked him about career advice for his kids. Elon went dark instantly

Great catch @SPadival, reminding us of Elon's reaction to Faber's question. I watched it again. Quite frightening to see him go silent for many seconds considering all he knows and feels about where AI is going. Deliberate suspension of disbelief indeed!

 

[Optimeer]

Obviously, this intelligent robotic future is inevitable. Everyone with half a brain must be able to see that as AI's capabilities continue to surprise, right? Optimus looks eerily like terminators IMO. The only real question simply is when. With all the building blocks right in front of us and with a company as capable and visionary as Tesla actually perusing AGI robots, it shouldn't leave a lot to the imagination. I don't know if there'll be a viable product in 5 years or 25 years, but it's certainly exciting! I also don't know if Tesla will be "the company" to ultimately bring this science fiction to the masses. But, I'm comfortable with my Tesla investment in EVs, battery storage, solar roof, FSD, and manufacturing innovations even if Optimus isn't what it might be. My gut tells me that there's a decent chance that Optimus becomes a useful product and I get a huge multiplier on my already great investment. And as for the "when"....well the faster Tesla goes, the more likely it is to be during my lifetime!

I concur.

As a mechanical engineer, until 2008, for a few years, I designed among other things an integrated robotic station for a tooling machine. It is just so much work to do this in the 3d CAD system, since you have to ensure that you can reach all points you need etc. The robot has to be close enough but not too close to reach points. You have 6 axis = 6d which is difficult to handle in a CAD system as with your mouse you only can do 3d. Next, the robot is stupid and could strike someone dead, so there are many safety measures to consider and true collaboration / interaction with other robots or humans is difficult. Despite that, I think robots are often superior to a customized automation with linear axes etc..

Two years ago, I visited a company which is spezialized to automate via robots with my team (in my new job). I couldn‘t really detect any progress in a decade. Proudly they showcased a vision system to assist the robot in grabbing something. It is a while since then, so I am not 100 % sure I remember everything right. But the display was like black and white and was flickering, so painful to watch as a human and the lag of the computation was like 1 s. The robot had no 3d understanding and would crash into something without slowing down. Of course, everything was pre-programmed. I was probably the only one not impressed by the tour. Maybe it‘s just me being arrogant, but it looks to me that this is an industry with hardly any progress.

Even with a conventional control approach, the setting up of a robot system would be so much easier, when the robot has two arms instead of 1, a vision system and can relocate with wheels or legs in the room and can capture the environment to not harm people and stuff. Further you would be flexible. You almost don‘t need any design in the CAD system any more. You just place your stuff somewhere and then tell the robot what he shoud do. Then you can use the robot on Monday for doing this, and Tuesday doing that at another work station etc. When there is an issue, the robot could do a first analysis, like creating a 3d model of the environment, posting the video of what happened and sending it via email to the operator and ask for further orders.

We know for sure what Tesla is able in capturing the environment in 3d. I am currently reading the book atomic habits, and read there that half of the human brain is used for vision. Maybe its included capturing of the environment and interacting with the environment, pattern recognition etc. So what Tesla already has done with FSD is awesome. The HW4 is just a bit, like 50% more powerful than HW3 computation-wise and offers a higher resultion of course. So this says me that Tesla is very confident that what HW3 offers is at least almost enough as soon as the software is optimized. If needed a computation solution like 3x more powerful would have been entirely possible. Of course there will be further improvement, but HW4 imo will be powerful enough for Optimus.

Most people just don‘t anticipate. They ask ChatGPT a few questions, see what is good and what is not good. But they just don‘t anticipate how this will will look like in 6 months, in a year, in two years. I read in a interview a complaint from an „machine learning expert“ that he was surprised that GhatGPT is not better. He has no idea. This is just the beginning. It dosen‘t matter how good something is. The only thing which matters here is the rate of progress.

I think it was after autonomy day in 2019, when I realized which new world is open for robotics. I was thinking that the company which will first use a similar recipe for robotics will eventually be worth 1 trillion. For that no super intelligent robot is needed, this was just unthinkable for me back in 2019. You just need a truly flexible robot who is easy to use and who is no danger for humans (collisions).

For me it is likely that Tesla will be the first company with an useful product and it is likely that it will take less than 5 years.

 

[Cosmacelf]

I concur.

As a mechanical engineer, until 2008, for a few years, I designed among other things an integrated robotic station for a tooling machine. It is just so much work to do this in the 3d CAD system, since you have to ensure that you can reach all points you need etc. The robot has to be close enough but not too close to reach points. You have 6 axis = 6d which is difficult to handle in a CAD system as with your mouse you only can do 3d. Next, the robot is stupid and could strike someone dead, so there are many safety measures to consider and true collaboration / interaction with other robots or humans is difficult. Despite that, I think robots are often superior to a customized automation with linear axes etc..

Two years ago, I visited a company which is spezialized to automate via robots with my team (in my new job). I couldn‘t really detect any progress in a decade. Proudly they showcased a vision system to assist the robot in grabbing something. It is a while again, so I am not 100 % sure I remember everything right. But the display was like black and white and was flickering, so painful to watch as a human and the lag of the computation was like 1 s. The robot had no 3d understanding and would crash into something without slowing down. Of course, everything was pre-programmed. I was probably the only one not impressed by the tour. Maybe it‘s just me being arrogant, but it looks to me that this is an industry with hardly any progress.

Even with a conventional control approach, the setting up of a robot systems would be so much easier, when the robot has two arms instead of 1, a vision system and can relocate with wheels or legs in the room and can capture the environment to not harm people and stuff. Further you would be flexible. You almost don‘t need any design in the CAD system any more. You just place your stuff somewhere and then tell the robot what he shoud do. Then you can use the robot on Monday for doint this, and Tuesday doing that at another work station etc. When there is an issue, the robot could do a first analysis, like creating a 3d model of the environment, posting the video of what happened and sending it via email to the operator and ask for further orders.

We know for sure what Tesla is able in capturing the environment in 3d. I am currently reading the book atomic habits, and read there that half of the human brain is used for vision. Maybe its included capturing of the environment and interacting with the environment, pattern recognition etc. So what Tesla already has done with FSD is awesome. The HW4 is just a bit, like 50% more powerful than HW3 computation-wise and offers a higher resultion of course. So this says me that Tesla is very confident that what HW3 offers is at least almost enough as soon as the software is optimized. If needed a computation solution like 3x more powerful would have been entirely possible. Of course there will be further improvement, but HW4 imo will be powerful enough for Optimus.

Most people just don‘t anticipate. They ask ChatGPT a few questions, see what is good and what is not good. But they just don‘t anticipate how this will will look like in 6 months, in a year, in two years. I read in a interview a complaint from an „machine learning expert“ that he was surprised that GhatGPT is not better. He has no idea. This is just the beginning. It dosen‘t matter how good something is. The only thing which matters here is the rate of progress.

I think it was after autonomy day in 2019, when I realized which new world is open for robotics. I was thinking that the company which will first use a similar recipe for robotics will eventually be worth 1 trillion. For that no super intelligent robot is needed, this was just unthinkable for me back in 2019. You just need a truly flexible robot who is easy to use and who is no danger for humans (collisions).

For me it is likely that Tesla will be the first company with a useful product and it is likely that it will take less than 5 years.

Absolutely. One basic thing people get wrong is that they equate Optimus to existing factory robots or Boston robotics. Optimus is end to end AI, unlike factory robots or the Boston robots. This is a huge difference, as anyone, like you, who has worked with factory automation can attest.

The other big difference is price. Boston robots are hugely expensive as they are one ofs made under military contracts for demo purposes. Teslabots will be mass produced general purpose bots, making them economical to use in a much, much bigger variety of use cases.

 

[kbM3]

Great thread on the advantages of the 48V system over and above reduced copper, and how it helps enable the unboxed next gen manufacturing technique.

https://twitter.com/x/status/1659330720610004994

 

[bkp_duke]

I hope the 4680 problems are resolved ever so soon. For tesla investors or investors in anything hoping a new technology comes along smoothly is not a recipe for profits.

I posted about it over a year ago and go lambasted. A year later it turns out I was right, sadly. I don't see any evidence that they've solved the DBE situation yet so if someone is counting on DBE CT sales to goose profits this year, or next, I'd suggest caution. I feel much more strongly that the Semi should ramp smoothly and add immediate profits that are synergistic with profitable energy megapack solutions. Further, the semi is going to make a meaningful contribution to fighting climate change whereas I am not sure if the CT will turn out to be just a suburban homeowner 3rd vehicle toy. In which case the net impact may actually be negative for the environment.

I'll get lambasted again, I'm sure. I did a year ago.

Not lambasted, I think you (like a lot of analysts) assumed 4680 would just be eventually "turned on". Battery day made it VERY clear it would be an iterative process that would take 5-6 years to hit targets. That's what those of us disagreeing with you are focused on - we are not expecting it to "get there" yet.

 

[MP3Mike]

I doubt it. The govt doesn't tax typewriters, computers or industrial robots as workers now. No tractor tax to pay for unemployed farm workers.

Depends on where the farm is. In the county I am in there is an annual tax on the tools of your business. So they absolutely tax tractors, typewriters, computers, mechanic's tools, etc.

 

[dhanson865]

Does anyone have Germany April 2023 sales by model? Like a list with these suspects and actual totals?

• Volkswagen Golf
• Tesla Model Y
• Volkswagen T-Toc
• Volkswagen Tiguan

I've been able to find random articles for Feb and March but not April. I'm still waiting for Model Y to consistently beat the VW Golf on it's home turf (It did so once in Sept 2022 but hasn't repeated the lead since)

Spoiler: prior articles

• Feb 2023 Golf #1, Model Y #2 Germany February 2023: VW Golf and Tesla Model Y dominate in another disappointing month (+2.8%)
• Mar 2023 more of the same Golf #1, Model Y #2 https://vrumcar.com/news/germany-car-sales-march-2023-one-car-ruins-the-big-volkswagen-party/5825/

Bueller, Bueller, anybody?

Do we have any German members here or am I just asking too soon to get April numbers?

 

[Tronguy]

Oh wow, you made me go down the rabbit hole of a quick re read of I Robot. In each short story, Asimov postulates some problem with a particular type of robot and what effect it would have. It is indeed an interesting exploration into the issues we might have with robots once we introduce AI powered ones into general society.

I was born 40 years too early. All my life I’ve fantasized about being a robot psychologist. That may very well end up being a real profession soonish.

Thing is.. In the Asimov stories, there's but one company that makes positronic brains, which, in all the Asimov stories, are the one and only, "thinking" machine that can be placed in a robot: U.S. Robotics. The stories allude to the idea that the design and build of these brains is too high-tech for a third party to come up with on its own, period. And that one company and its successors keep the three (or maybe four, counting the zeroth law) in all the robots.

Under those circumstances, the robots tend to be, well, friendly. Which is cool.

Except we live in a world where any fool can put together neural networks or ChatGPT things that are.. getting.. rather close to sentience. And while there are some companies making attempts to build something that looks vaguely like the three laws into their products, there's plenty who don't give a hoot: There's Money To Be Made, there's a land-rush mentality, and, so, we're all going to find out the hard way whether these new technologies will work with us without us all getting killed. Or the entities running the planet won't be humans any more, the ones doing it will be 'way more intelligent than us. It's nice to think that superintelligent would be nice to humans, Ian Banks style, but there's no guarantees. And no working guardrails.

And, even beyond the question of niceness or not, there's the fundamental bit about whether hyperintelligent entities would even be sane. There are plenty of species on the face of the planet which are social. Us, for example. Ants, bats, bees, deer, and so on are as well. In a scarcity of food environment with significant resources devoted to reproduction (hello, Darwin!) those species that didn't socialize in one form or another died out over the eons. Hyperintelligent entities don't have that built-in genetic-built evolutionary behavior. So what these entities might do might not make any sense from any point of view, because they don't have guardrails. Blow up the world because one doesn't see why not?

People have thought about this. One thought revolves around the Fermi paradox. Humans have been around for around 100,000 years. Say that there's an intelligent species out there with a million idle years on its hands. Assuming that they would travel, they'd be all over the galaxy. So, Fermi asked: "Where is everybody?"

One thought people have had is that, once a biological species reaches a certain level, they invent computers, get them smarter.. and then the whole thing falls down, for scary reasons given above.

Hang onto your hats, it's going to be a bumpy decade.

 

[mongo]

I hope the 4680 problems are resolved ever so soon. For tesla investors or investors in anything hoping a new technology comes along smoothly is not a recipe for profits.

I posted about it over a year ago and go lambasted. A year later it turns out I was right, sadly. I don't see any evidence that they've solved the DBE situation yet so if someone is counting on DBE CT sales to goose profits this year, or next, I'd suggest caution. I feel much more strongly that the Semi should ramp smoothly and add immediate profits that are synergistic with profitable energy megapack solutions. Further, the semi is going to make a meaningful contribution to fighting climate change whereas I am not sure if the CT will turn out to be just a suburban homeowner 3rd vehicle toy. In which case the net impact may actually be negative for the environment.

I'll get lambasted again, I'm sure. I did a year ago.

https://twitter.com/x/status/1649028439138086913

 

[Captkerosene]

Depends on where the farm is. In the county I am in there is an annual tax on the tools of your business. So they absolutely tax tractors, typewriters, computers, mechanic's tools, etc.

My response was very specific about not taxing a robot as a "worker." Taxing authorities can be creative though ... eventually, I can see the cops hauling my refrigerator of for not paying its income taxes. (It's always had a rebellious nature.)

 

[Artful Dodger]

My gut tells me that there's a decent chance that Optimus becomes a useful product and I get a huge multiplier on my already great investment.

Indeed, my gut* tells me that the purpose of new 'Tesla Restaurants' to be built along major Supercharging corridors are NOT just for flipping burgers, it's for flipping the skeptics as they watch Optimus Prime at work in the 'Glass Kitchen of the Future'.



Cheers!

*Best way to a man's heart is through his stomach.

 

[Fred42]

Indeed, my gut* tells me that the purpose of new 'Tesla Restaurants' to be built along major Supercharging corridors are NOT just for flipping burgers, it's for flipping the skeptics as they watch Optimus Prime at work in the 'Glass Kitchen of the Future'.

View attachment 939951

Cheers!

*Best way to a man's heart is through his stomach.

'Toasters' using toasters to make my breakfast.

(Even if I drive up in a blue Lincoln?)

(Optimus Prime look to me like the 'toasters' in Battlestar Gallactica.)

 

[wtlloyd]

That’s the whole point of the pricing system and capitalism, it determines what something is worth, completely fairly and equitably.

I feel much better about the stock market now. Thanks!
(Your argument is ridiculous)

 

[T@oo]

Thing is.. In the Asimov stories, there's but one company that makes positronic brains, which, in all the Asimov stories, are the one and only, "thinking" machine that can be placed in a robot: U.S. Robotics. The stories allude to the idea that the design and build of these brains is too high-tech for a third party to come up with on its own, period. And that one company and its successors keep the three (or maybe four, counting the zeroth law) in all the robots.

Under those circumstances, the robots tend to be, well, friendly. Which is cool.

Except we live in a world where any fool can put together neural networks or ChatGPT things that are.. getting.. rather close to sentience. And while there are some companies making attempts to build something that looks vaguely like the three laws into their products, there's plenty who don't give a hoot: There's Money To Be Made, there's a land-rush mentality, and, so, we're all going to find out the hard way whether these new technologies will work with us without us all getting killed. Or the entities running the planet won't be humans any more, the ones doing it will be 'way more intelligent than us. It's nice to think that superintelligent would be nice to humans, Ian Banks style, but there's no guarantees. And no working guardrails.

And, even beyond the question of niceness or not, there's the fundamental bit about whether hyperintelligent entities would even be sane. There are plenty of species on the face of the planet which are social. Us, for example. Ants, bats, bees, deer, and so on are as well. In a scarcity of food environment with significant resources devoted to reproduction (hello, Darwin!) those species that didn't socialize in one form or another died out over the eons. Hyperintelligent entities don't have that built-in genetic-built evolutionary behavior. So what these entities might do might not make any sense from any point of view, because they don't have guardrails. Blow up the world because one doesn't see why not?

People have thought about this. One thought revolves around the Fermi paradox. Humans have been around for around 100,000 years. Say that there's an intelligent species out there with a million idle years on its hands. Assuming that they would travel, they'd be all over the galaxy. So, Fermi asked: "Where is everybody?"

One thought people have had is that, once a biological species reaches a certain level, they invent computers, get them smarter.. and then the whole thing falls down, for scary reasons given above.

Hang onto your hats, it's going to be a bumpy decade.

Except that it seems a common concept in certain circles, that the first one who gets AGI up and running will be the last one and as a consequence the only one to do so. What that exactly means .... I leave it up for your imagination. I tend to agree with this thought and I have my own opinion on how this may look like.

 

[Nrkl]

The system is always unstable....while seeking stability.

"We" don't know what is coming and any discussion on reality is...while entertaining....incomplete.
It is for those inclined an exciting ride.

I only wish I was younger, so I had more time to enjoy....but who knows....maybe I have forever!

One can think of many thought experiments to make this easier to understand. Even with simple systems, like a thermostat, if you adjust the temperature repeatedly before the previous adjustment have taken effect you will be in “trouble”. With organic systems that are networked, such as what the biosphere is all about, many effects might take a long time to manifest. So if you make constant changes in these systems the outcome will be wild oscillations and a crash of the system.

none of this would be as dramatic if we at least wouldn’t expand the spatial scale at the same time we compress the temporal scale. So that there would be rapid localized experiments with innovation and technology, some of those expeirments end up crashing due to piling negative unintended consequences, and some would survive. This is how evolution works and evolution itself is a well proven process.

Fantasies of centrally controlled robot utopias are nothing but a central engineering god doing intelligent design instead of “trusting” evolution.

 

[hobbes]

Interesting new statistics from the German registration authority - this is not monthly registrations but percentage of current total fleet as of Apr 1 this year.
Tesla (12.5%, up 1.3% yoy) is second behind Volkswagen (17.7%, down 1.5% yoy, this is VW brand not group).

Title: "Share of the 15 top brands with BEV total of passenger vehicle fleet as of Apr 1 2023 in percent"

Source:

https://www.kba.de/SharedDocs/Downloads/DE/Statistik/Fahrzeuge/FZ27/fz27_202304.xlsx?__blob=publicationFile&v=11

 

[hobbes]

Tesla will set quotas for the German sales team from June, salary dependent on success​

Exklusiv: Tesla gibt deutschem Verkaufsteam ab Juni Quoten vor, Gehalt erfolgsabhängig

Tesla stellt den Vertrieb im deutschsprachigen Raum um. Verkäufer werden zu Spezialisten – und zum Teil abhängig vom Erfolg bezahlt.

teslamag-de.translate.goog

 

[ItsNotAboutTheMoney]

Bueller, Bueller, anybody?

Do we have any German members here or am I just asking too soon to get April numbers?

Has a few mentioned specifically as the top models:

Germany: All-Electric Car Sales Increased In April 2023 By 34%

In April, all-electric car sales in Germany increased by 34% year-over-year to 29,740 new registrations, which represents about 14.7% of the market.

insideevs.com

Last month, the most popular model happened to be the Volkswagen ID.4/ID.5 duo with some 2,723 units, followed by the Volkswagen ID.3 (2,050). We also saw not bad 1,405 Skoda Enyaq iV and 1,370 Audi Q4 e-tron, which are direct cousins of the ID.4/ID.5.

The Tesla Model Y had some 1,636 registrations in April. However, its massive advantage over others so far this year (17,487 units YTD) is unthreatened.

So, ID.4 family, ID.3 and Tesla Model Y.

 

[Bobfitz1]

Indeed, my gut* tells me that the purpose of new 'Tesla Restaurants' to be built along major Supercharging corridors are NOT just for flipping burgers, it's for flipping the skeptics as they watch Optimus Prime at work in the 'Glass Kitchen of the Future'.

View attachment 939951

Cheers!

*Best way to a man's heart is through his stomach.

My gut tells me Tesla needs Franz to redesign the Optimus for better curb appeal. Stainless steel so customers can choose the wrap they like.

 

[Zaddy Daddy]

Pulverize powder well? I'm not sure you understand 4680 technology, which is a huge red flag on your opinions.

Oh really? Please explain what the current bottleneck is for limited DBE yield, because Tesla hasn't told us lately.

What they had explicitly said was that originally the calendaring machinery was getting dented rolling the material into a film. Conjecture / discussion / youtube videos then they were changing roller material, # of rollers, and trying to cool / freeze the material before going into the rollers.

Are those still the issues? If not, what is? Does anyone know? Considering this is cost a good amount of profit the last few quarters and has limited growth, don't you kinda think you would like to know?

BTW the limited growth most definitely showed up in limited Powerwall production in 2022. If you remember a while ago (I think Summer 2021), Musk was quite confident that Powerwall production would grow immensely in 2022. Then it didn't happen, at all. Likely because they had to take that 2170 supply and throw it at Model Y.

Not lambasted, I think you (like a lot of analysts) assumed 4680 would just be eventually "turned on". Battery day made it VERY clear it would be an iterative process that would take 5-6 years to hit targets. That's what those of us disagreeing with you are focused on - we are not expecting it to "get there" yet.

Please explain where he said 5-6 years. He said 3 years to achieve all the cost savings. We are near 3 years and have achieved negative cost savings.