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

[unk45]

I have talked a bit with Karpathy, very down to earth over the top nice guy. Don’t know him too well but will just paint a story based on public posts by him and what everyone knows.

He is a pretty young at his 35 years, born in Slovakia. Brilliant guy. Got a Phd(with all that means) at Stanford, at the time when deep learning took off. Being a very articulate he got the role to lead the course that many students around the world watched on youtube. He read so many papers he even did his own service to follow new papers more efficiently.

GitHub - karpathy/arxiv-sanity-preserver: Web interface for browsing, search and filtering recent arxiv submissions

Web interface for browsing, search and filtering recent arxiv submissions - karpathy/arxiv-sanity-preserver

github.com

After his Phd, he got a job at OpenAI as a scientist. One year later he joined Tesla, his first major job in the industry doing applied engineering with validation, shipping real products and all that mess. During the Model 3 ramp where everyone had to chip in he was down at the factory line, doing whatever he could like everyone else. He has done this for 5 years. Given his rockstar fame and the demand for top DL talent he was probably paid millions of dollars, plus the same in shares that has 10x since he joined. He is very likely set for life... Also he has a CV that is really impressive and can pretty much get any job he wants if he needs more money.

On his free time, he codes the GPT clone minGPT and other small hobby project to understand the latest devevelopment in AI and make it easier for amateurs to access the latest research. He even coded a bitcoin client from scratch just to understand crypto.

A from-scratch tour of Bitcoin in Python

Musings of a Computer Scientist.

karpathy.github.io

Like I said, he is 35years old and, as far as I know, single. When he takes his first holiday, he goes traveling around the world, meeting fellow AI researchers around the world. He also is a mostly vegan person, “90%+ vegan” and he thinks being vegan is one of the best way to reduce CO2.

So imagine a person like that. Likes to travel the world, has worked pretty damned hard and got to the top of his hiearchy, set for life with both cash and stocks. An academic at heart. Maybe he just wants to be free, travel when he wants, do the research when he wants, keep up with latest development in NLP, spend some time on health and do some dating. Heck even explore spirituality, psychadelics and all that stuff that people who travel around the world, are vegan, into crypto etc like to do.

Should he stay longer? He probably has pushed it ”just another year” a few times already. 5 years is a good number to call it quits. Until FSD is released? It has been next year for a few years now. Can he contribute more? He pretty much defined software 2.0 and made the only known large scale implementation of it. The system is set up, it works well. Now it’s time for the applied industry people to run the development from here.

So no, I don’t think we can infer that FSD is near or far away based on him leaving. He probably just wants to have a break, do what he loves and can afford to do and have some more time for the good stuff in life.

This post is unusually insightful. It helps explain how Musk-led businesses attract people who otherwise would not be attracted to industrial enterprises. Therefore this post belongs here:

Moderators' Choice: Posts of Particular Merit

This thread displays posts from the Investor Sector that the Moderators deem worthy of an especial place where they most easily can be located and re-located. Moderators have been collecting such posts over a period of time; to the extent more time appears on the event horizon then perhaps some...

teslamotorsclub.com

 

[Stretch2727]

If your account has less than $25k in it - you cannot day trade more than (I forget the frequency) a couple of times per 90 days.
If you are trading in a retirement account - you have to wait for the funds to settle from your trades before using that money again - 3 days.

I would imagine it is one of those 2 instances happening.

If you make a retirement account a "Margin" account you do not have to wait for settlement. It does not allow for any borrowing but you can buy again without waiting for settlement. I swing trade one ETF constantly and have no problems with multiple trades per day without settlement.

 

[Matias]

And the bottleneck for speed is just raw compute power. So I think the development of Dojo is far more important than the development of the kinds of ideas that Karpathy brings to the table.

If you're doing wrong/futile things, doing them faster won't help. I'm not saying that this is the case. Just pointing out that the speed alone is not the solution.

 

[AudubonB]

To give you an idea of how bad the natural gas crunch is going to be for Germany this winter, here’s a forecast study that assumes people will start resorting to burning wood as a replacement. This is de facto de-industrialization. Hopefully this will be fixed within a year. But it highlights very real risks for Tesla’s Berlin plant this winter since I assume it uses natural gas.

https://twitter.com/x/status/1547180581141766144

I can anecdotally support just that. Our close friend/former associate just flew back from his home in northern Germany; among other dire news he reported he’d spent the prior days preparing a winter’s worth of firewood for his mid-80s parents - who can barely take care of themselves under the best of circumstances - and told them to prepare for camping out in their wood stove-equipped living room + kitchen all winter.
He shares the frustration others have voiced about Germany intentionally having hamstrung itself by decommissioning its nukes without putting replacements in service.

 

[UltradoomY]

If you make a retirement account a "Margin" account you do not have to wait for settlement. It does not allow for any borrowing but you can buy again without waiting for settlement. I swing trade one ETF constantly and have no problems with multiple trades per day without settlement.

Correct - but you do have to apply for and receive the margin.
It is not automatic, I was just stating the reasons you would not be able to day trade. There are multiple ways around that, but you have to do the work to set it up.
Cheers

 

[bkp_duke]

The Limiting Factor has their part 2 up on the 4680 cell:

Fortunately much more digestable at 11 minutes long.

TL;DW - energy density looks pretty good at 272 wh/kg (cell level), but there is room to improve energy density, which is no surprise to anyone here.

Also, NO SILICON yet in the 4680, so some low-hanging fruit for improved energy density. (His cell is from 6 months ago, and Munro's cells are the latest version - so there may already be changes).

I encourage you all to watch the full video.

 

[elasalle]

To give you an idea of how bad the natural gas crunch is going to be for Germany this winter, here’s a forecast study that assumes people will start resorting to burning wood as a replacement. This is de facto de-industrialization. Hopefully this will be fixed within a year. But it highlights very real risks for Tesla’s Berlin plant this winter since I assume it uses natural gas.

https://twitter.com/x/status/1547180581141766144

GF Berlin - should go solar/ mega packs in a big way.

If Germany/EU ordered more mega packs - maybe they should be given higher priority.

Hornsdale Megapack bet was for 100 days ...

(+ after Starlink satellites, what better way to piss off .... )

 

[GOVA]

The good thing is that the FUD targeted against Tesla is actually obscures the importance of non-Tesla EV cars. The mob that lives by the FUD always thinks in terms ICE vs Tesla and keeps buying ICE.

I strongly think that the last nail into ICE coffin will be the non-Tesla EVs that seem to pop out left and right. Tesla will be fine, ICE will die.

I've been following the Out of Spec Reviews and his latest EU reportings look really interesting. Do check his latest vides.

 

[Usain]

The Limiting Factor has their part 2 up on the 4680 cell:

Fortunately much more digestable at 11 minutes long.

TL;DW - energy density looks pretty good at 272 wh/kg (cell level), but there is room to improve energy density, which is no surprise to anyone here.

Also, NO SILICON yet in the 4680, so some low-hanging fruit for improved energy density. (His cell is from 6 months ago, and Munro's cells are the latest version - so there may already be changes).

I encourage you all to watch the full video.

Just gave it a watch and I really liked what I heard. We have a cell that is already very good with LOTS of room for improvement.

Now it's down to manufacturing speed. Then raw materials supply chain. Then a $10,000 stock price.

 

[jerry33]

If you make a retirement account a "Margin" account you do not have to wait for settlement. It does not allow for any borrowing but you can buy again without waiting for settlement. I swing trade one ETF constantly and have no problems with multiple trades per day without settlement.

After all, it wouldn't do to let have retail have the same prerogatives that the MMs have

 

[bkp_duke]

Just gave it a watch and I really liked what I heard. We have a cell that is already very good with LOTS of room for improvement.

Now it's down to manufacturing speed. Then raw materials supply chain. Then a $10,000 stock price.

I like that given this is the BASE cell, as you said with room for improvements, we are already at parity on energy density with the 2170s.

Things I quickly gleaned that are not yet in the design, which can improve over they next 1-2 years:
1) Only the Anode is DBE currently, leaving improvement in the cathode when it moves to DBE
2) Decrease shell thickness moderately (currently 500-600 um, vs. 125-250 um thickness for 2170).
3) Add silicon for anode (5-10% immediate boost)

 

[Knightshade]

Hey remember the argument about how Tesla was making thousands of pre-production berlin cars and that they'd have to scrap em (or find some other non-sales use for em)?

Tesla looks to sell "good" quality early Model Y units from Giga Berlin: report

Tesla is reportedly looking to sell some units from its early Model Y production units from Giga Berlin to customers.

www.teslarati.com

 

[JRP3]

Decrease shell thickness moderately

Since it's part of the structural integrity of the pack I wouldn't expect that to change.

 

[insaneoctane]

One surprising thought about Karpathy leaving is the opportunities he's actually leaving. With humanoid robot AI just starting -Optimus in its infancy, FSD still not done (opportunity to be around when it dazzles the world) and the $$ due to options/stock to what, 5x-10x, within the next decade? He would have carte blanche at Tesla to play and pursue almost unlimited interests with more resources than ANYWHERE. I wish him good luck, but am honestly surprised...

 

[Cosmacelf]

GF Berlin - should go solar/ mega packs in a big way.

If Germany/EU ordered more mega packs - maybe they should be given higher priority.

Hornsdale Megapack bet was for 100 days ...

(+ after Starlink satellites, what better way to piss off .... )

Berlin doesn’t have high solar radiance, so, while possible, it may not be the cheapest. The problem isn’t electricity supply, that’s being looked after in Germany through the use of cheap coal. The problem is natural gas availability which is used in industrial processes.

 

[LYTRIDR]

I like that given this is the BASE cell, as you said with room for improvements, we are already at parity on energy density with the 2170s.

Things I quickly gleaned that are not yet in the design, which can improve over they next 1-2 years:
1) Only the Anode is DBE currently, leaving improvement in the cathode when it moves to DBE
2) Decrease shell thickness moderately (currently 500-600 um, vs. 125-250 um thickness for 2170).
3) Add silicon for anode (5-10% immediate boost)

Since only the anode is DBE, that aligns with our prior knowledge that there was some issues with denting of the calendering rollers by harder metals that make up the cathode. Since this cell is at least 6 months old, Jordan rightly points out that once he gets the new cell from Munro's 4680 Y teardown, they will repeat the testing and do additional testing.

Jordan also mentions that he was hoping to do cathode/anode permeability testing but decided against it due the incoming newer production version of the cell he hopes to get soon. This can be important in understanding how Tesla was able to increase the thickness of the active materials which seems to be responsible for the better than expected energy density figures.

I'm looking forward to seeing the new specs on the newest cells, especially to find out if the cathode has moved to DBE. Jordan estimates it will take a few months to get the data back from the lab.

It is great content as usual, but he has mentioned on twitter that he is thinking about making his videos available only for patrons due to a lot of relatively stupid criticism on social media. I'd not be surprised to see him bifurcate his content and release less for free to the public. It's a big +1 from me to spend the 11 minutes to watch this video.

 

[Cosmacelf]

One surprising thought about Karpathy leaving is the opportunities he's actually leaving. With humanoid robot AI just starting -Optimus in its infancy, FSD still not done (opportunity to be around when it dazzles the world) and the $$ due to options/stock to what, 10x, within the next decade? He would have carte blanche at Tesla to play and pursue almost unlimited interests with more resources than ANYWHERE. I wish him good luck, but am honestly surprised...

I’ve said it before, but no one in the deep learning industry believes me. The industry’s and Tesla’s current AI technology is not up to the task for TeslaBot. My opinion is you are going to need continuous learning for that which is fundamentally different from the AI architecture currently in use. Maybe that’s why Karpathy is leaving, he realizes this and wants/needs to do exploratory work on even more cutting edge stuff.

 

[Chenkers]

But it highlights very real risks for Tesla’s Berlin plant this winter since I assume it uses natural gas.

This question has come up a bit of how Giga Berlin would be affected by a rationing of natural gas to industrial users in Germany. So I thought I would do a bit of research to assess the likely impacts. Btw my background is in building services mechanical engineering so I have some knowledge of natural gas systems.

I watch all of the Giga Berlin drone videos that come out each week and have also gone through the detailed architectural and services plans for Giga Berlin that were released for planning. The main uses for Natural Gas that I can so far determine for Giga Berlin are as follows:

1. Giga Press furnaces - Used to melt the aluminium ingots to the required temperature to be fed into the gigapress to form the rear underbody castings (2 operational, 2 in progress, 4 more planned).
2. Natural Gas Boilers in the main central thermal plant building (the shed next to where new cars are marshalled for loading). Used for general heating requirements through the heating hot water pipework reticulated throughout the site. (there are also chilled water pipes for cooling but these are from electric chillers, not gas)
3. Domestic Hot Water - Bathrooms, showers etc. This could be electric or gas depending on the system but there are some hot water plantrooms on the plans in places.
4. Various equipment that may directly use natural gas - there's no way to tell if and how much of this there is as details aren't provided (eg the battery cell building has a large gas meter room but no large boilers).

No.1 is potentially the biggest problem as these furnaces are likely to rely on gas. The most logical solution I can see is to import excess rear giga castings from either Fremont, Shanghai or Texas. Based on installed capacity, stored parts and current utilisation, there is likely to be some spare capacity across the other giga factories that could keep production going at Berlin during it's ramp. It will add logistical cost and complexity but nothing Tesla can't handle and no harder than getting battery packs and drives from Shanghai.

No.2 doesn't appear much of an issue as the main boilers in the central utility building are still being installed (flues still being connected this week) and not on-line yet. Tesla have made do so far with various temporary heat pumps scattered around the outside of the building so I expect they could keep these running or use more if needed. The main ventilation plant has air to air heat exchangers and there are also a lot of localised heat pumps used so the demand for gas boilers should not be that great.

No.3 Not much of an issue (cold showers anyone). Temporary electric units are easy to install and would be used to date along with the temporary heat pumps so not a drama for now.

No.4 is harder to predict as it's a bit of an unknown. It's not likely that much modern plant would directly use gas, particularly if trying to be green. Drying etc in the paint shop would be the biggest risk but then the Geico Taikisha paint shop is supposed to be their Zero Environmental Impact version. I can't see this being heavily reliant on a fossil fuel like natural gas. The battery cell production could use gas for process heat but that won't be online for a while yet.

So overall I don't see the imposition of natural gas restrictions in Germany being a major impediment to Giga Berlin continuing to ramp. It could create some headaches but not anything beyond the Tesla teams ability to overcome.

 

[dmvevguy]

One surprising thought about Karpathy leaving is the opportunities he's actually leaving. With humanoid robot AI just starting -Optimus in its infancy, FSD still not done (opportunity to be around when it dazzles the world) and the $$ due to options/stock to what, 5x-10x, within the next decade? He would have carte blanche at Tesla to play and pursue almost unlimited interests with more resources than ANYWHERE. I wish him good luck, but am honestly surprised...

I'm disappointed he left as he's the real tech AI leader there, and probably the #1 in the world. It must be burnout, exhaustion, etc. b/c he did take sabbatical. I'm sure there are super smart people like him already at TSLA, but the optics are just not so favorable. Hopefully it's a small blip in the road, b/c Elon just tweeted they are still rolling out V11 soon, but when I first heard this I was very disappointed (and still am).

 

[Drumheller]

Since it's part of the structural integrity of the pack I wouldn't expect that to change.

My guess is that Tesla has engineers, maybe SpaceX engineers, exploring using other materials for the shell. Maybe something like carbon fiber. Cost / weight / ease of manufacture trade offs, of course.