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

[elasalle]

I have a pinned tweet since Apr 7

S Padival on Twitter

Couple of people close to Musk follow me.. so hopefully, it has been brought to his attention.

One thought I had on this though ... is what is the likely hood of success?

Retail investors will not move the needle much, so will the institutes who are the large holders go along with this .. would they want to increase their stakes?

I think interest in this approach will have to be first sorted out with the big holders before it is offered if it is?

If there is interest, I think SP jump will happen prior to the fact due to leaks etc.

(Just saying, because look where the previous 420 financial Engg got us) ... cheers!!

 

[elasalle]

So we have talked alot over the past week about how NN-based computer vision will enable autonous driving.

Do any of our computer science experts share their opinions how Tesla can use their newly developed computer vision technology to aid manufacturing at GF3/Shanghai and for Model Y in a space/labor constrained environment?

Elon has hinted that manufacturing at GF3 who be 'more advanced', and 'best in the world'. Could this be robotic vision grafted to automated assembly?

So does Tesla have the chops to integrate it's new computer vision technology with existing manufacturing lines to replace labor while minimizing space? How big an 'ask' is this?

Elon dropped another hint during the Q1 conference call that the autonomous charging 'snake' is now possible with NN based computer vision. How big a jump is it from plugging in a charging cable to plugging in a wiring loom?

If possible, I think that would be a huge factor allowing the Model Y to be build at one of Tesla's existing facilities. Perhaps the new Tesla FSD computer will be in more demand than we initially thought?

Same hardware, different training. Train the NN with camera images, then it can drive a robotaxi.

Train the NN with manufacturing process movies, then it can BUILD the robotaxis. Jus' sayin'.

Cheers!

Train flufferbot?

The Kuka robots are all highly software driven. In the future surely AI can be applied for multiple disciplines.
Personally, I think team should focus on just getting FSD working. cheers!!

+ML (Machine Learning) can be applied to Manufacturing process - doesn't have to be NN though. Collect data about processes in the manufacturing pipeline from various components, sensors etc. and conduct data analysis on that data set. It can help detect trends, anomalies etc based on data and provide iterative processes/actions for improvement of the manufacturing process.

 

[Aieukl378]

I'm 100 percent serious. Sorry but acting like a 15 yr. old on Twitter is only entertaining when your company is hugely successful and stock price isn't getting destroyed on a daily basis. Come to think of it it still isn't entertaining...unless you're also 15.

Shiny new M3 goes into a parking lot I’m in yesterday. Owner gets out, people start asking questions. The guy is raving about handling, autopilot, superchargers, etc.

Last time I saw this kind of enthusiasm for a new car was when Lexus first came out. The first Lexus owner in the locker room raves about the car and gets everyone's attention. When the second owner starts raving, it’s harder to discount the first guy as a fanboy. By the time the third guy in the locker room has bought and is raving, the Lexus is a “hot car”.

If there was advertising stoking initial Lexus demand, I don’t remember it. I don’t recall knowing what the car looked like until I started seeing it in the parking lot.

Word of mouth from owners is everything in selling $40k products in large quantities. There’s no way to calculate “steady state demand” until almost every potential buyer knows at least three M3 owners. That will happen pretty fast, since it’s exponential, but that process is just getting started.

I

There's some truth to this simplely because of all the restrictions he has over Twitter now. Him not violating the rules set forth does nothing for the stock. Him violating the rules only drops the stock. So there's really no upside for him being on Twitter since shorts will be applies every letter he now says on Twitter to the rules set forth. So now it's a game of just baiting him to violate in which the shorts are doing after EVERY tweet of his.

I am only talking about the stock price here. I still see value for him using Twitter as free advertisment for all his projects.

OMG yes exactly to your first paragraph. In regards to his using Twitter as "free advertising" if we took the stock price before the "420 tweet" ($354.98)and TSLA last share price($235.14) with a difference of $119.84 and calculated Elon's TSLA exposure and how much less he has now(and don't forget the 40 million in fines, etc. etc.)....that's some expensive "free advertising". Can anybody come up with the actual number?

 

[VValleyEV]

There's some truth to this simplely because of all the restrictions he has over Twitter now. Him not violating the rules set forth does nothing for the stock. Him violating the rules only drops the stock. So there's really no upside for him being on Twitter since shorts will be applies every letter he now says on Twitter to the rules set forth. So now it's a game of just baiting him to violate in which the shorts are doing after EVERY tweet of his.

I am only talking about the stock price here. I still see value for him using Twitter as free advertisment for all his projects.

Wait. The new settlement actually reduces twitter restrictions. A lot. (Assuming it is approved by the judge).

I agree with you that Twitter is a valuable tool for free advertisement, therefore a value to shareholders, and this new agreement will mean that Twitter may now be used for that purpose with way less concern for threats from the SEC.

If you are reading that bullet list in the new settlement and thinking “gosh here are a long list of new restrictions” you are missing the point. In the settlement agreement before this one, there was vague and ambiguous language that the SEC was trying to use to challenge virtually any statement from Elon as a violation, whether the statement had any new or material information or not. That language was removed in the new agreement, and that weapon wielded by the SEC no longer exists.

Now, again assuming this settlement is approved, there is agreement that only new and material information needs vetting by Tesla attorneys, and there is a list of what could reasonably be interpreted as material information. This is in fact what Elon had been doing since the previous agreement, limiting twitter to avoid release of new material information, and consulting an attorney when there was a grey area. The SEC attacked anyway, using that vague language provision, now they will no longer be able to do so.

Am I anywhere near close, @Fact Checking ?

 

[Intl Professor]

etc...

Speaking of the nittiest nits to pick, it always annoys to see ellipses without the period denoting an end to the sentence. Is that a coy ploy to suggest there is more to come? (Admittedly, I never had a college English class. After 4 years at MIT I wrote less well than as a senior in high school, so always insecure about writing. Audie, the master grammarian once caught me in a capitalization error. Fortunately, it was not a decapitalization error, as frequently happens in an unnamed Middle Eastern country.)

 

[Aieukl378]

He might have to vow to quit saying FSD is right around the corner, and that they will be profitable every quarter from here on out, also.

Wait. The new settlement actually reduces twitter restrictions. A lot. (Assuming it is approved by the judge).

I agree with you that Twitter is a valuable tool for free advertisement, therefore a value to shareholders, and this new agreement will mean that Twitter may now be used for that purpose with way less concern for threats from the SEC.

If you are reading that bullet list in the new settlement and thinking “gosh here are a long list of new restrictions” you are missing the point. In the settlement agreement before this one, there was vague and ambiguous language that the SEC was trying to use to challenge virtually any statement from Elon as a violation, whether the statement had any new or material information or not. That language was removed in the new agreement, and that weapon wielded by the SEC no longer exists.

Now, again assuming this settlement is approved, there is agreement that only new and material information needs vetting by Tesla attorneys, and there is a list of what could reasonably be interpreted as material information. This is in fact what Elon hard been doing since the previous agreement, limiting twitter to avoid release of new material information, and consulting an attorney when there was a grey area. The SEC attacked anyway, using that vague language provision, now they will no longer be able to do so.

Am I anywhere near close, @Fact Checking ?

But we ALL know Elon will not comply and will continue to push SECs buttons regarding his stupid Twitter usage. This is another reason institutional investors are bailing(they know there will be more SEC/Twitter drama ahead.) God it's just so f...ing ridiculous.

 

[SO16]

I

OMG yes exactly to your first paragraph. In regards to his using Twitter as "free advertising" if we took the stock price before the "420 tweet" ($354.98)and TSLA last share price($235.14) with a difference of $119.84 and calculated Elon's TSLA exposure and how much less he has now(and don't forget the 40 million in fines, etc. etc.)....that's some expensive "free advertising". Can anybody come up with the actual number?

Probably a bit disingenuous to blame him on Twitter for the majority of this massive price drop. If Tesla had been much closer to guidance Q1 and profitable Q2 (to help dispel the “demand is falling” narrative), I don’t think this huge drop would have occurred.

 

[Artful Dodger]

Train flufferbot?

The Kuka robots are all highly software driven. In the future surely AI can be applied for multiple disciplines.
Personally, I think team should focus on just getting FSD working. cheers!!

None of those robots use NN based computer vision. Their software has to be laboriously and intricately programs, tested, and debugged.

Vision based robots could be shown how to do a take, literally, and train themselves to do the task by watching repetative examples.

FSD is a separate task. Each task gets its own NN build suited to that task. So although each solution is task-specific, the technique for building the solution is general, and in fact is what Tesla has constructed.
It's also eminently suitable for parallel development of solutions.

I expect NN vision to take over as Tesla's core manufacturing technology. I expect a revolution at Tesla Grohmann when the robots that build the robots can see with their own eyes.

It's the singularity.

 

[Dan Detweiler]

He might have to vow to quit saying FSD is right around the corner, and that they will be profitable every quarter from here on out, also.

Unless of course it actually is and they will. Just sayin'.

Dan

 

[Fact Checking]

How did you come up with 10 year $8k per car? Any rough logic?

I see it as for $50k car insurance will be $150 a month ($1800/year), and the combined ratio will be 95% (best in industry). Say investment gains in 6%, they can profit 10%, which is only $180 a year for 10 years is $1800.

Also if the insurance take rate is only 20-40%, the per car equation is even lower. Am I missing something?

That's the beauty of Tesla Insurance IMHO, the possible gains aren't just the investment income from the premiums collected and held - like it is for regular insurance companies.

Instead the gains are from the spread between the much lower underwriting costs that Tesla can achieve via high-tech solutions such as Sentry Mode, much safer cars, FSD advantages, 7 cameras recording every accident, etc.

I.e. say there's an average insurance rates of $150 a month per car on average from a generic car insurance provider, which covers underwriting costs (payouts for insurance claims or re-insurance costs), corporate overhead and expected profits.

If Tesla can reduce this cost to say $100 per month, then they can keep $50 of pure profits and the $18 generic profit built into the generic insurance rate - 3x-4x higher profits that no other insurance company has access to.

This is a big assumption though, and it might not materialize to that extent - but it's an interesting experiment nevertheless.

 

[SPadival]

One thought I had on this though ... is what is the likely hood of success?

Retail investors will not move the needle much, so will the institutes who are the large holders go along with this .. would they want to increase their stakes?

I think interest in this approach will have to be first sorted out with the big holders before it is offered if it is?

If there is interest, I think SP jump will happen prior to the fact due to leaks etc.

(Just saying, because look where the previous 420 financial Engg got us) ... cheers!!

Bailee Gifford had a standing offer to fund additional capital if Musk needs it. He will have to re-confirm if the offer still stands. Them and Ron Baron are my preferred institutional investors. Note that SP is pretty close to the avg price that Baron funds bought at ~ $230. They were the Johnny come lately and may have an appetite to load up further

Tesla backer ready to put more money behind Musk : teslamotors

 

[mongo]

So we have talked alot over the past week about how NN-based computer vision will enable autonous driving.

Would any of our computer science experts share their opinion how Tesla can use their newly developed computer vision technology to aid manufacturing at GF3/Shanghai and for Model Y in a space/labor constrained environment?

Elon has hinted that manufacturing at GF3 would be 'more advanced', and 'best in the world'. Could this be robotic vision grafted onto automated assembly?

Does Tesla have the chops to integrate it's computer vision technology into existing manufacturing lines to replace labor while minimizing space? How big an 'ask' is this?

Elon dropped another hint during the Q1 conference call that the autonomous charging 'snake' is now possible with NN based computer vision. How big a jump is it from plugging in a charging cable to plugging in a wiring loom?

If possible, I think that would be a huge factor allowing the Model Y to be build at one of Tesla's existing facilities. Perhaps the new Tesla FSD computer will be in more demand than we initially thought?

Same hardware, different training. Train the NN with camera images, then it can drive a robotaxi. Train the NN with manufacturing process movies, then it can BUILD the robotaxis.

I bet computer vision based robots could relieve some production bottlenecks due to 'scare' labor on Panasonic battery cell lines at GF1 too. Jus' sayin'.

Cheers!

Vision assisted assembly would be great, but if they haven't already implemented it on the existing robots at Fremont, I don't see them running it as an experiment in GF3.

For the 3, I expect Tesla to be risk adverse (plus Manual assembly is much lower capex than robots) and implement the following:
Direct copy of working/ in use robots and fixtures from Fremont.
Improved material flow system (line running along a lot of loading bays)
Hybrid GA line, working robots plus designed for human stations.
Same parts as current 3 (possibly cost down interior).

Given the changes needed to support robotic wiring, I think that will not happen till Y. They may hedge and go manual install to start on the first line with provisions for robots. Or Grohmann is already building/testing the install bots so there will not be issues. That would be a low risk development path.

 

[JohnDinger]

So about Tesla insurance.

Allstate provides a discount plus sustained discount using their drive wise system that monitors when you drive, how fast, and how many harsh braking events.

Tesla can monitor you in a whole new level. I wonder if they will change your rates on the fly for doing risky things like max acceleration events and how often you run red lights. Essentially the car pretty much knows everything about you.. even know if you are cheating on your spouse or not. So I don't really know how hardcore they are willing to go with this. Once underwriting is assessing risks of an individual, the sky is kind of the limit.

Tesla will have a lot of cool things they can bring to auto insurance. But I don't think it will be a significant business for them for a long time.

Most people choose insurance based on what's cheapest, not brand or customer service.

That's why margins for the auto- insurance industry are around 5% -- it's a race to the bottom.

Even if Tesla could get 30% gross margin, that's still only 400,000 cars * $100 monthly premium * 30% = $12 million per month.

Probably not a significant part of their business until they can get past selling 500k cars per year and start eating into the ICE industry.

I suspect it's more of a marketing tactic. Insurance rates can prompt sticker-shock for people who are on the line between buying/not buying. I'm sure they'd get cheaper rates if they shopped around, and this way, Tesla guarantees a low price.

 

[Aieukl378]

Probably a bit disingenuous to blame him on Twitter for the majority of this massive price drop. If Tesla had been much closer to guidance Q1 and profitable Q2, I don’t think this huge drop would have occurred.

Yes, of course it's not "only" twitter and yes terrible forecasting/communication regarding financials BUT add those issues to Twitter usage.

You're an institutional investor and you're evaluating whether to add, stay or go(TSLA):

While stock plummets and company seems in peril CEO on Twitter is
1. spending time creating rap videos
2. poking SEC with Twitter jabs(leaf blower latest)
3. retweeting porn sites
4. acting like a 15 year old
5. and on and on and on

Removing himself for Twitter is technically simple(2 clicks?) and would at least remove all the uncertainty and risk of his Twitter usage.

 

[Silent Ludicrosy]

Speaking of the nittiest nits to pick, it always annoys to see ellipses without the period denoting an end to the sentence. Is that a coy ploy to suggest there is more to come?

 

[⚡️ELECTROMAN⚡️]

If Tesla Insurance proves successful, they will have bought another group of Tesla-haters....insurance company workers. Tesla already has plenty of haters bound to lose from the disruption:

- Fossil fuel industry workers
- Gas station owners
- Car dealers
- Auto mechanics
- Auto parts manufacturers, distributors and retailers
- Car tuners and aftermarket parts suppliers
- Dash cam manufactures
- Taxi services (future)
- Truck Drivers (future)

And next, insurance providers.

The disruption is great, but it invites a lot of haters who are bound to lose their livelihoods.

Reminds me of Apple:

-Flashlight makers and sellers
-GPS makers and sellers
-Camera makers and sellers
-Pedometer makers and sellers
-EKG makers and sellers
-Music stores
-CDs and CD player makers and sellers
-Calculator makers and sellers
-Thomas Guide
-Phone books
-etc.

 

[VValleyEV]

But we ALL know Elon will not comply and will continue to push SECs buttons regarding his stupid Twitter usage. This is another reason institutional investors are bailing(they know there will be more SEC/Twitter drama ahead.) God it's just so f...ing ridiculous.

What evidence to you have that Elon has not been complying with the previous settlement all along? Have you read the statement from Elon to the court to that effect? What evidence is there that there will be additional SEC/Twitter drama, especially when the new settlement gives the SEC LESS opportunity to attack Elon and/or Tesla?

Perhaps you are just saying that evidence and facts do not matter, everyone will continue to believe what they read in headlines and ignore the substance? This always happens with every news cycle on pretty much any subject, but if the reality is really different, the issue will fade away.

 

[gavine]

Reminds me of Apple:

-Flashlight makers and sellers
-GPS makers and sellers
-Camera makers and sellers
-Pedometer makers and sellers
-EKG makers and sellers
-Music stores
-CDs and CD player makers and sellers
-Calculator makers and sellers
-Thomas Guide
-Phone books
-etc.

Yes! Also, clock radio/alarm clock makers and sellers.

 

[ZachF]

Indeed, especially when we need seven league boots right now. But I think the new S is already sporting 6 league boots.

Cheers!

How many cubits tall is an S?

 

[elasalle]

None of those robots use NN based computer vision. Their software has to be laboriously and intricately programs, tested, and debugged.

Vision based robots could be shown how to do a take, literally, and train themselves to do the task by watching repetative examples.

FSD is a separate task. Each task gets its own NN build suited to that task. So although each solution is task-specific, the technique for building the solution is general, and in fact is what Tesla has constructed.
It's also eminently suitable for parallel development of solutions.

I expect NN vision to take over as Tesla's core manufacturing technology. I expect a revolution at Tesla Grohmann when the robots that build the robots can see with their own eyes.

It's the singularity.

Didn't EM just say some tasks are better suited for humans?

I hope NN etc not applied to manufacturing at least for now. (Another distraction, waste of resources etc etc)

More simplicity in the Tent approach was what got TSLA over the hurdle last year.