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

[AndrewZ]

People should read into the reporting about the legality of what Tesla is attempting to do with the new shareholder vote.

I’d guess this saga is far from over, even if the vote goes in Elon’s favour.

 

[AudubonB]

Housekeeping:

Too many of you (even one is too many but it’s quite a bit more than that) have gotten sloppy with your overabundant use of abbreviations.
YOUR microseconds of saved time absolutely do not balance out against the time-wasting collective hours of others’ musings as to what it is to which you are referring.

Besides, use of them suggests that you don’t care about forum rules. And that never ends well.

 

[Musskiah]

Doesn’t the nanny cam cut FSD if it sees your eyes are not on the road. How did you get away with texting without the system disengaging? FSD 2.3.x won’t work if the lens of the camera is covered either.

Ever use the onboard texting?

I have my secrets for phone texting as well...

 

[AudubonB]

I am amused by the following; I have no idea if others are:

My reaction to the news today that Tesla appears ready to re-submit to shareholders Mr Musk’s 2018 incentive plan was:
THIS time I very well might vote for it. I opposed it the first time because (1) my bullishness regarding the company at that time was such that I envisioned the goals as very likely to be achieved; and
(2) which if so, very likely would cause exactly the kind of legal backlash - a lawsuit of some kind - that would adversely affect…Tesla, probably, and TSLA utterly without question.

So…..
This time, for me the single best path I can see to register my disgust at the Delaware Court’s decision is to reverse my own decision, and Vote With Extreme Fervor FOR the reinstatement of the 2018 Compensation/Incentive package.

Conflicting? Contradictory? Hypocritical? I do not believe so. My targets are different; my goal is the same: to enhance the value and the price of TSLA.

 

[Buckminster]

https://twitter.com/x/status/1780705171129196683

Top so far:

ANONYMOUS RETAIL INVESTOR ASKS
Retail
Following Tesla's Robotaxi unveil on 8/8, what is the realistic timeline for launching a revenue generating robotaxi network?


1.7K Votes
1.2M TSLA Shares Represented

ANONYMOUS RETAIL INVESTOR ASKS
Retail
Can we make FSD transfer permanent?(until FSD is fully delivered / Level 5 autonomy)


1.3K Votes
925.1K TSLA Shares Represented

ANONYMOUS RETAIL INVESTOR ASKS
Retail
What is tbe progress on the cheaper next generation vehicle?

 

[mrmage]

Limited inputs? Go was explicitly chosen because it is a complicated game with FAR more possible plays than chess.

From Scientific American: “Go's complexity is bigger, much bigger. With its breadth of 250 possible moves each turn (Go is played on a 19 by 19 board compared to the much smaller eight by eight chess field) and a typical game depth of 150 moves, there are about 250^150, or 10^360 possible moves.”

That’s a lot for “limited inputs”. And that’s using technology from over a decade ago.

And if you asked most people if you could just show a neural network videos of cars driving and it would be able to do it pretty darn well with almost no control code, they wouldn’t have believed you either.

And if I told you I could show a random picture of a cat to a computer and it could identify that there’s a cat in the picture, it would have been thought of as impossible just two decades ago.

The car can already identify people of all different kinds, shapes, sizes, and with different clothing with almost 100% reliability. Why do you think the capability stops there?

How is it that a human brain can drive in an infinite world of infinite possibilities? Yes, we have more neurons, but we also have to do a lot more than just drive.

Neural nets don’t have to be perfect drivers. No human is a perfect driver either.

Go is complicated for a game, but not compared to predicting appropriate behaviors in the real world. Alphago was also a blend of algorithms and ML, whereas Tesla gave up on rules / algorithms because they didn't work in the real world.

As far as complexity, Alphazero has about 20M parameters. A really small crummy LLM has 100M, but that's nearly useless because it's constantly wrong. LLMs need 5-10B to get started, and closer to 100B to be useful in the real world. FSD probably has 10x to 100x more parameters than Alphago, so yes, AlphaGo is very simple in comparison to most AI models.

As you noted, image classification took decades because it required multiple breakthroughs. NN, CNN, transformers, more compute/memory, etc. Tesla will get there too, but might also require multiple breakthroughs or more compute/memory to reach all those nines.

As far as FSD progress, it seems unlikely that a ML task that gets exponentially more difficult (each nine is an exponent!) would progress linearly, much less exponentially, when the amount of compute for inference is constrained.

Mod edit: You appear to have Mod posts on block. All of NINETEEN MINUTES after a Mod post!
Next time you pull this, you’ll be pulling a vacation.

 

[dl003]

https://twitter.com/x/status/1780705171129196683

Top so far:

ANONYMOUS RETAIL INVESTOR ASKS
Retail
Following Tesla's Robotaxi unveil on 8/8, what is the realistic timeline for launching a revenue generating robotaxi network?


1.7K Votes
1.2M TSLA Shares Represented

ANONYMOUS RETAIL INVESTOR ASKS
Retail
Can we make FSD transfer permanent?(until FSD is fully delivered / Level 5 autonomy)


1.3K Votes
925.1K TSLA Shares Represented

ANONYMOUS RETAIL INVESTOR ASKS
Retail
What is tbe progress on the cheaper next generation vehicle?

I see 3 huge red candles in our future.

 

[AndrewZ]

I really hope the answer to half those questions isn’t “March of 9s”

Especially the question that asks about the March of 9s

 

[jamaatta]

Potus wants to triple steel and Aluminum import tariffs. Can he do it alone or does it need Congress to vote?

Inflation… huh? Could be just political posturing, but there is funding in the Infastructure bill for Steel Mfg in Pennsylvania and Ohio. Quoting the Whitehouse here.

Congress, “through the Bipartisan Infrastructure Law and the Inflation Reduction Act, recently announced up to $1.5 billion for six clean iron and steel projects as part of a historic broader investment to lower emissions from energy-intensive industries. These projects are set to demonstrate innovative technologies that can eliminate the vast majority of emissions from steelmaking, and enable the industry to phase out more traditional carbon-intensive production methods.”

(Really?)

“They will support the economic comeback of steel communities in Pennsylvania, Ohio, and the South and Midwest, so the U.S. steel industry can remain competitive as the global leader in low-carbon iron and steel products.“

Maybe related?

"SSAB has been selected for awards negotiations for up to 500 MUSD in funding from the U.S. Department of Energy for a potential construction of a HYBRIT[®] manufacturing facility, capable of producing fossil free iron by using green hydrogen instead of fossil fuels. The project will also explore the possibility to expand capacity and capability at SSAB’s steel manufacturing operations in Montpelier, Iowa, including increased use of renewable energy."

SSAB selected by U.S Department of Energy to explore possibilities for production of fossil-free steel in the U.S.

SSAB has been selected for awards negotiations for up to 500 MUSD in funding from the U.S. Department of Energy for a potential construction of a HYBRIT[®] manufacturing facility, capable of producing fossil free iron by using green hydrogen instead of fossil fuels. The project will also explore...

www.ssab.com

 

[mongo]

Investors can save about $25 billion dollars in payroll taxes by ratifying the 2018 CEO compensation plan.

You are confusing the GAAP accounting impact (paper costs) of new stock options with payroll taxes.

For illustrative purposes, the Company’s accounting team informed the Special Committee that a new grant of 300 million fully vested options — functionally equivalent to what Mr. Musk had before the Tornetta Opinion — would potentially result in an accounting charge in excess of $25 billion, depending on certain timing and valuation factors.

Payroll taxes would be basically 1/2 of the 2.9% Medicare amount plus %0.9 for amounts >$200k, so 2.35%.
$25B would require a payout of 25/.0235 or just over a trillion dollars.

 

[Artful Dodger]

Aaron Greenspan's appeal has been rejected by the courts:

https://twitter.com/x/status/1780841602040336694

I will not expand the acronym GFY for Fredtrik; this is a family site.

Cheers!

 

[CitizenKane]

Not much new but a pretty good summary of sentiment: https://www.barrons.com/articles/tesla-stock-price-problem-demand-fef78a84?mod=Searchresults

 

[Artful Dodger]

I'M GOING TO DISNEYLAND! ... now imagine how cool it would be to get pulled over by a Cybertruck cruiser?

https://twitter.com/x/status/1780840031470686695

 

[FEERSUMENDJIN]

Logically speaking a 2-seater Robotaxi is probably ideal for 95% of trips, but isn't suitable for the other 5%.

Retrofitting any required hardware to existing models seems to me like the most efficient way of satisfying that last 5%.

I am referring to comments like :- "I will not use a Robotaxi" / " I will not pay 12K for FSD" - clearly Tesla is hoping there are customers for the services they intend to provide, I no longer play golf, but it does seem that the game is somehow surviving without my contribution.

Once robotic cars are on the road at scale, insurance companies will drive human drivers off the road immediately.

 

[Ellec]

Once robotic cars are on the road at scale, insurance companies will drive human drivers off the road immediately.

Exactly this, it will be like asking them to insure your car with no seat belts, air bags or locks, they might do it but it will be very expensive.

 

[Todd Burch]

Ok, so back to value of FSD:

Suppose Tesla insurance charges some per-mile premium when driving manually, and a significantly less per-mile premium while driving on FSD.

Safety statistics could, in the not-too-distant future, support a situation where the insurance savings alone could pay for most of the $99/month FSD subscription cost.

If the net result due to insurance savings is that the true cost of FSD to a consumer is only say $10 or 20/month, what would the take rate be? I would wager it would be extremely high.

And if statistics bear out that insurance payouts are extremely low for drivers on FSD, that becomes nearly pure profit for Tesla, on a recurring basis, for the forseeable future. I would switch to Tesla insurance immediately.
That may be the piece of the FSD affordability puzzle that is currently missing.

We are close to the point where this can realistically occur, and Tesla has the means to see when that point gets hit and already has the mechanism to make it happen via FSD. And it doesn’t even require full autonomy.

Tesla should highlight this on the earnings call. With data to back it up. I don’t think analysts have considered it.

 

[dedicatedtek]

I'M GOING TO DISNEYLAND! ... now imagine how cool it would be to get pulled over by a Cybertruck cruiser?

https://twitter.com/x/status/1780840031470686695

" Throw 'em in the frunk".

 

[replicant]

https://europe.autonews.com/sales-market/europe-sales-fell-3-ev-demand-declined-11

Europe sales drop in March as EV weakness persists
Sales of plug-in hybrid vehicles rose nearly 1%, outpacing BEVs and gasoline and diesel models.

 

[jerry33]

That’s a good point about insurance. Why wouldn’t other insurance providers offer a discount as well.

Perhaps because insurance companies are heavily invested in fossil fuels? (my understanding)

 

[Usain]

Joe Tegtmeyer posted a new drone video today showing exactly that, with cybertrucks exiting from the new SouthWest door of Giga Texas.

Yeah, I got my wish like an hour later! So glad Cybertruck is back on track.

Now if we could get a good handle on the run rate. Something tells me I'll have to wait a bit longer for that one.