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

[SOULPEDL]

Yes, I have. Again you either refuse to, or are unable to, understand them.

For example a while back when Tesla had already run entirely out of single-node compute and was forced to spill over to Node 2 on HW3-- making it impossible it would be usable for RT (because they'd have to have at least redundant nodes to be safe enough for an RT.... Waymo, since you love to compare to them, has entire fully independent computers in their cars).




But again you have no basis for that assumption beyond hope and not understanding why those aren't the same thing.





Again, they are not.

Unless you define the problem at a 50,000 foot level and ignore all the details.

Which admittedly seems to be your approach to everything.


Almost NOTHING about how Waymo is doing things will be relevant to how Tesla would create or deploy a self driving system.





Sure I have.

But the fact they don't have a system capable of >L2 operation, and certainly not of L4, seems to be a pretty good one to start with.

We went over, in great detail, why that is a fact and as I recall you mostly just denied or misunderstood what the levels mean and what elements Teslas system is missing over in the FSD related threads (lack of a complete OEDR, lack of a fallback mechanism, etc)-- and it wasn't like I was the only one trying to explain why you were wrong either.

There's many OTHER reasons potentially in the way (insufficient redundancies, insufficient ways to clear the few sensors Tesla uses when obscured, insufficient compute, insufficient side visibility at occluded intersections, insufficient distance measurement capabilities esp. on the <HW4 cameras, etc--- some of which may or may not be soluble in SW but it's far from a given on any of em).






What other ones can you cite, specifically?

Yo, I'm misquoted there, and not going to debate this. I'm already beyond the agree to disagree. Let's get there shall we?

 

[Usain]

Yes, I have. Again you either refuse to, or are unable to, understand them.

For example a while back when Tesla had already run entirely out of single-node compute and was forced to spill over to Node 2 on HW3-- making it impossible it would be usable for RT (because they'd have to have at least redundant nodes to be safe enough for an RT.... Waymo, since you love to compare to them, has entire fully independent computers in their cars).




But again you have no basis for that assumption beyond hope and not understanding why those aren't the same thing.





Again, they are not.

Unless you define the problem at a 50,000 foot level and ignore all the details.

Which admittedly seems to be your approach to everything.


Almost NOTHING about how Waymo is doing things will be relevant to how Tesla would create or deploy a self driving system.





Sure I have.

But the fact they don't have a system capable of >L2 operation, and certainly not of L4, seems to be a pretty good one to start with.

We went over, in great detail, why that is a fact and as I recall you mostly just denied or misunderstood what the levels mean and what elements Teslas system is missing over in the FSD related threads (lack of a complete OEDR, lack of a fallback mechanism, etc)-- and it wasn't like I was the only one trying to explain why you were wrong either.

There's many OTHER reasons potentially in the way (insufficient redundancies, insufficient ways to clear the few sensors Tesla uses when obscured, insufficient compute, insufficient side visibility at occluded intersections, insufficient distance measurement capabilities esp. on the <HW4 cameras, etc--- some of which may or may not be soluble in SW but it's far from a given on any of em).






What other ones can you cite, specifically?

I refuse to get into a discussion about the SAE levels because it's just a classification system. It is actually irrelevant to determining whether a vehicle can really operate as a robotaxi. For that, you need a proof case.

Again, Waymo is the proof case that robotaxi can be done. So point out something essential that Waymo can do but Tesla end-to-end with vision only can not do relatively soon.

The "spillover" argument is a red herring because the need for node 2 was under a different software architecture. We don't know the onboard performance characteristics of the end-to-end system. There is zero evidence that the new system will not be able to utilize redundancy, even under HW3.

The ability to clear sensors won't come into play any time soon. A robotaxi trial can easily get started without that feature. Perhaps it will be solved on the Gen 3 vehicle or by "robotaxi" add-on hardware. But it's not needed to start if it's really needed at all.

There is no hard evidence that "insufficient compute" will be a problem. Again, hardware may be a future limitation, but at this point it's just a guessing game.

There is no evidence that occluded intersections would keep Tesla from starting a robotaxi service.

As for distance measurement, I'm not sure exactly what you are complaining about. But you seem to think it will not be a likely problem on HW4. Requiring HW4 for a robotaxi trial would be just fine.

 

[SOULPEDL]

Here's Cathie's Camp...

Humor me...

 

[Cal1]

She has not been right on Telsa for years. Amazing she is still followed.

 

[Zaddy Daddy]

Thats not good investing though. Thats waiting until something has happened, then buying at the same price as everyone else. Good investing is seeing whats imminent, and buying in advance at a quarter the price.
When its absolutely clear that FSD is a reality and Tesla's can be relied upon to drive you without paying attention, the stock is $1,000. I'd much rather buy now and wait a few years to quadruple my money.

I was partially joking. You can try to predict a quarter before you think it's going to happen, but you'll need quantitative trends indicating a projection for 100x reduction in critical disengagements.

The thing is most of the market isn't simply paying as much attention as people here. I guarantee that if videos aren't showing any disengagements for a month, the market definitely will not have priced it all in. Share price would probably be $250-$300 but nowhere near $1000. But $1000 would probably come within 6 months - 1 year. I'd prefer avoiding the flat period (which could last 2-5 years), buy with 2x leverage and I wouldn't lose anything. That's my personal preference

 

[Knightshade]

Yo, I'm misquoted there, and not going to debate this. I'm already beyond the agree to disagree. Let's get there shall we?

Corrected via edit- I have now ascribed the earlier quote to the right poster, and only the one that's actually you is cited to you now-- thanks for catching this, mutli-quoting users on here isn't the best experience but I prefer trying that to dropping a bunch of posts in a row and taking up even more space with something that should be in a different forum to start with.

I refuse to get into a discussion about the SAE levels because it's just a classification system. It is actually irrelevant to determining whether a vehicle can really operate as a robotaxi. For that, you need a proof case.

it's 100% not irrelevant.

Most regulations and laws allowing self driving explicitly reference the SAE levels.

This is another on the increasingly long list of facts about this you refuse to accept or attempt to understand and your refusal to get into such a discussion is indicative of your unwillingness to have your erroneous assumptions challenged.

Again, Waymo is the proof case that robotaxi can be done. So point out something essential that Waymo can do but Tesla end-to-end with vision only can not do.

I already listed two specific things Waymo currently does that Teslas current system can not.

If you have no better argument than "Tesla will just figure out all the missing parts" without any understanding of the parts, what they are, what they do, and why you need them then we are back to your entire argument is "I want to believe it's a done deal because then I'll be rich and Hopium will let me ignore any information suggesting otherwise"

The "spillover" argument is a red herring because the need for node 2 was under a different software architecture. We don't know the onboard performance characteristics of the end-to-end system.

Which means we don't know if the new architecture did anything, at all, to fix the problem.

You, however, are already assuming the problem no longer exists.

There is zero evidence that the new system will not be able to utilize redundancy, even under HW3.

There is though. Nodes A and B don't have the same actual capabilities-- something Green noticed once the spillover began- and which you've had pointed out to you before and just keep dismissing or ignoring (long list of those).

The ability to clear sensors won't come into play any time soon.

You don't think cameras ever get dirty?

People get "camera obstructed" messages from FSD all the time--- it's in play now

Once there's no human to substitute for a blinded camera you're done. How can a vehicle in the middle lane of a 3 lane road "fail safely" if it can't see anything to one of its sides since there's only 1 camera with visibility there?

There is no evidence that occluded intersections would keep Tesla from starting a robotaxi service.

Except, there is. The inability to reliably handle such intersections on current HW for example.


Every time someone points out an actual, current, real-world limit-- you just hand waive it away as Tesla Will Magically Solve It Somehow.

Even physical limits software can't fix (like detecting objects not in the cameras field of view soon enough).

As for distance measurement, I'm not sure exactly what you are complaining about.

Are you unaware FSD v12 has already hit parked cars due to insufficient resolution on distance measurement (and lack of USS sensors to back up the vision distance guessing)?

But you seem to think it will not be a likely problem on HW4

I'll take "ideas you just made up and ascribed to me for $1000 Alex!"

. Requiring HW4 for a robotaxi trial would be just fine.

You just ranted for 5 paragraphs about how HW3 is fine for RT, now you're ok with requiring 4 (and stranding like 4 million current owners).


It's pretty clear dude.... RT=You Get Rich and you don't really care about the details or the reality of the situation.


Feel free to head back to the FSD forum for even more folks to explain this to you, but we're done in this thread.

 

[Zaddy Daddy]

I see, that's fair logic.

I'll try this another way...

SOURCE FOR BELOW:

https://twitter.com/x/status/1751963928542781758

View attachment 1029146
Teslas TOTAL REVENUE, which Cern calls INDICATIVE REVENUE(TR=IR=Reported Revenue + change in Deferred Revenue & Unsatisfied Performance Obligations) was nearly $10B in 2023 (with Lathrop ramped at half mast [20GWh] for only a few months of the year) and the difference in unreported revenue is nearly $4B!

So in 2024 IR should likely approach $20B as Lathrop will be fully ramped to 40GWh for about half of 2024 (as you say), so we should end up something like double the 2023 IR.

In 2025 Lathrop could do over $20B of IR and Lingang(Shanghai) should do about $10B of IR (like Lathrop did this year).
Let's assume this $30B of 2025 IR between MPactories only REALIZE $20B @ your 20% margin. That's only $4B or $1.3EPS. Maybe you are correct sir. OR maybe at 30% margins that's $6B INCOME or $2 EPS. So Tesla comes in at $1.3-$2 EPS in 2025.

Now if margins crush expectations at 35% and REALIZED revenue is $25B instead of $20B, energy alone does $3 EPS In 2025. This is entirely possible if efficiencies of scale come into play and COGS reducation occurs, say from 30% battery declines as has been predicted.

I still stand by my CONSERVATIVE $2EPS TE FY2025

Yeah. IMO 30% operating margins are insane to assume. Gross margins last 2 quarters were like 18%. I think those could max out at 30%.

Remember the CEO specifically stating they were targeting gross margins around the same as auto (at the time, 25-30%). Which would make operating margins lower.

No reason to think Musk was underselling Megapack margins, nor does it make sense given the current trend in margins for the product.

 

[aronth5]

Exactly. And robotaxis are still many years away yet, we need two things to happen first:

1) FSD solved to Level 5
2) Gen 3 (which the Robotaxi will be based on) car needs to begin production.

My feeling is the stock won't reflect any true robotaxi potential until both of those are realities, and that's still two or more years away at the earliest. A physical robotaxi service is probably more like 4-5 years away at the earliest. We'll likely have Optimus selling to customers before then.

It is coming, its a matter of WHEN not IF, but its probably still quite a ways off.

Tesla doesn't need FSD solved to Level 5 to bump the take rate and stock value. Level 3 with a graceful FSD to person in the driver seat handoff would be a significant improvement. Owners could then do pretty much anything from the driver seat like watch video, read, text etc. Pretty much anything but taking a nap. Sure L4/L5 is the final goal but doesn't have to be the only goal. Do I believe Tesla will do this, nope which is too bad.

 

[cliff harris]

Also Tesla wants to drop the price per mile vs uber, so profit margin wouldn't be as super high.

Ideally sure, but not 100% required. I'd pay extra to be driven by a robot that I dont have to talk to, and has no danger of talking to me. And thats me just as an autistic 54 year old dude. I can imagine young women would, assuming its technically superior, be happier to be driven home by a robot car than some random dude.
Also, if there is no driver, there is one extra seat. A 4 seat robotaxi carries 4, A 4 seat uber carries 3, and one poor sod has to talk to the driver.

 

[UncaNed]

My own experiences. For real, the wife was highly impressed on multiple instances. (I can't imagine how scary it must be the whole time sitting in the passenger seat - that would freak me out I think.)

I have noticed it doesn't seem to like turning right on red in places it used to before, but it's early still. So I really only needed to add some pedal here and there, but with that, there is no way anyone would be able to tell this is driving on FSD.

My settings were on aggressive. Very light rain at times. It's that good.

Is V12 FSD update coming to all Beta folks now? Wondering why I haven't seen any sign of it yet. I'd love my skepticism for imminent TSLA price boost from V12 to be proven wrong.

 

[UncaNed]

The best day of owning a business is when you first realize that you can fire customers.

Indeed. And then there's the sudden realization that you could/should have been doing that all along.

 

[SOULPEDL]

I just got a letter from someone that I met several years ago at a conservation meetup and thought I'd share for the Az folks here. We practically have the most sunshine on the entire planet and yet I'm still the odd one out with solar in our Chandler home.

He's urging me to sign a petition for people who support clean initiatives. He also suggest we vote on SRP utilities election (and to register ASAP if you haven't, but I've already voted).

FYI, SRP is one of our utility providers who recently pushed forward legislation (battle from hell) to finish up the Coolidge Burner plant off in some poor neighborhood, and pays us as little as 2.5 cents/kWh as a DC during peak hours. There is legal history between this State and Tesla on Net Metering etc. There was even a period of blackout for years where I couldn't get Tesla Solar due to that legal mess. (So politics aside, I think this is fair game to share. But I'd understand if maybe it needs to change or be removed.)

Seems the ACC (Arizona Corporation Commission) has a couple seats open as well. I've signed the petitions and this is who he recommended in case others in Az want to add to the score.

• Ylenia Aguilar, a current Central Arizona Water Conservation District Board Member
• Joshua Polacheck, a retired Foreign Service officer and former Executive Director of the Pima County Democratic Party
• Jonathon Hill, an engineer and scientist at the ASU Mars Space Flight Facility

Edit: SRP Clean Energy link.

 

[aronth5]

I made an appointment to see my broker in person. Way harder to lie to me or dispense bs. For starters, I’ll be expressing my desire that the company stop spamming commercials using high dollar athletes and actors to sell their products. Animated talking reptiles and cartoon characters seem to be fairly effective for others.

I’d rather not have to shop around. I’ve been quite satisfied to this point, so I’ll give them an opportunity to make it right first. They tell me I’m already getting all their discounts. Guess they’ll need to make up a new one because I won’t pay the inflated prices, especially with another vehicle coming on board shortly.

So if the car insurance companies are gouging the public, which they are, one has to wonder why Tesla hasn't added a state to their insurance program for a long time or made money underwriting? Anybody care to speculate?

 

[SOULPEDL]

Is V12 FSD update coming to all Beta folks now? Wondering why I haven't seen any sign of it yet. I'd love my skepticism for imminent TSLA price boost from V12 to be proven wrong.

You're East coast, ya maybe that's why. Past has gone West to East. There's a lot of cars to update.

 

[Usain]

Corrected via edit- I have now ascribed the earlier quote to the right poster, and only the one that's actually you is cited to you now-- thanks for catching this, mutli-quoting users on here isn't the best experience but I prefer trying that to dropping a bunch of posts in a row and taking up even more space with something that should be in a different forum to start with.





it's 100% not irrelevant.

Most regulations and laws allowing self driving explicitly reference the SAE levels.

This is another on the increasingly long list of facts about this you refuse to accept or attempt to understand and your refusal to get into such a discussion is indicative of your unwillingness to have your erroneous assumptions challenged.







I already listed two specific things Waymo currently does that Teslas current system can not.

If you have no better argument than "Tesla will just figure out all the missing parts" without any understanding of the parts, what they are, what they do, and why you need them then we are back to your entire argument is "I want to believe it's a done deal because then I'll be rich and Hopium will let me ignore any information suggesting otherwise"





Which means we don't know if the new architecture did anything, at all, to fix the problem.

You, however, are already assuming the problem no longer exists.




There is though. Nodes A and B don't have the same actual capabilities-- something Green noticed once the spillover began- and which you've had pointed out to you before and just keep dismissing or ignoring (long list of those).





You don't think cameras ever get dirty?

People get "camera obstructed" messages from FSD all the time--- it's in play now

Once there's no human to substitute for a blinded camera you're done. How can a vehicle in the middle lane of a 3 lane road "fail safely" if it can't see anything to one of its sides since there's only 1 camera with visibility there?




Except, there is. The inability to reliably handle such intersections on current HW for example.


Every time someone points out an actual, current, real-world limit-- you just hand waive it away as Tesla Will Magically Solve It Somehow.

Even physical limits software can't fix (like detecting objects not in the cameras field of view soon enough).




Are you unaware FSD v12 has already hit parked cars due to insufficient resolution on distance measurement (and lack of USS sensors to back up the vision distance guessing)?




I'll take "ideas you just made up and ascribed to me for $1000 Alex!"





You just ranted for 5 paragraphs about how HW3 is fine for RT, now you're ok with requiring 4 (and stranding like 4 million current owners).


It's pretty clear dude.... RT=You Get Rich and you don't really care about the details or the reality of the situation.


Feel free to head back to the FSD forum for even more folks to explain this to you, but we're done in this thread.

If Tesla can do everything Waymo can do then it will satisfy whatever regulations are out there, regardless of your interpretation of SAE levels. Again, this makes SAE discussion irrelevant.

You have now pointed out some problems, but not limits. Solving problems are not done by magic. It's done by engineering. And I don't see any that are particularly hard to solve now that Tesla has gone end-to-end.

I never said HW3 was a requirement for a robotaxi trial. HW3 could very well be enough though.

Dirty cameras are not a problem that will keep Tesla from rolling out an initial trial of robotaxi. One blinded camera does not mean the car can not fail safely. Longer term, there are both hardware and software solutions to deal with the problem.

So now I know what you were referring to with the distance measurement problem. We don't even have confirmation that FSD was enabled in that little accident. But we do know that actually smart summon is on the way soon. So Tesla must have confidence that their vehicles are not going to go around bumping into parked cars.

 

[Knightshade]

So if the car insurance companies are gouging the public, which they are, one has to wonder why Tesla hasn't added a state to their insurance program for a long time or made money underwriting? Anybody care to speculate?

From the data posted I dunno maybe a week ago, we know at least on the policies Tesla is underwriting themselves they're losing money- could be they don't want to scale further until they've figured out how to make the business profitable.

 

[Snerruc]

Sorry to interrupt the insurance companies are crap thread, but I’m wondering if Tesla has reached the limits of increasing demand by lowering prices. Conquest sales often are based on things other than price. Tesla may be reaching the limits of what it can do with limited models and limited variations on them. Lots of people like choice, lots of people have brand loyalty, and so on. I suspect Tesla may be studying this problem. I had to drive a Nissan Rogue for a couple or days. Yep software and much else was crap, but it had those wonderful mirror warnings and uss with 360 view. When I got back in my Tesla I really missed them. And that was on a $30,000 car. Believe it or not, they are more important to some people than being an electric car. All my wife cared about was that when she pushed the pedal, the wheels went around. Lots of people like that.

 

[M|S|M MYP]

So....is $TSLA up today because:

1. It's been oversold and value is coming back into play
2. Shortzes made enough and are covering
3. People realizing FSD is farther along than it should be
4. Report saying that Elon's use of prescription drugs is good for the company
5. Elon saying 'flying car' is possible in Don Lemon interview
6. MM's met again with Tesla IR and now have new delivery info

Anything else?

 

[Buckminster]

So....is $TSLA up today because:

1. It's been oversold and value is coming back into play
2. Shortzes made enough and are covering
3. People realizing FSD is farther along than it should be
4. Report saying that Elon's use of prescription drugs is good for the company
5. Elon saying 'flying car' is possible in Don Lemon interview
6. MM's met again with Tesla IR and now have new delivery info

Anything else?

It's because of Apple buying Gemini usage. I read all bout it.

 

[SOULPEDL]

So....is $TSLA up today because:

1. It's been oversold and value is coming back into play
2. Shortzes made enough and are covering
3. People realizing FSD is farther along than it should be
4. Report saying that Elon's use of prescription drugs is good for the company
5. Elon saying 'flying car' is possible in Don Lemon interview
6. MM's met again with Tesla IR and now have new delivery info

Anything else?

7. Above.
8. Tesla Pushed Dan O'dowd V12 update, now he can't even fake the accidents.
9. Jim Cramer saying the opposite. (Assumed)
10. Optimus has learned to sleep and now wants something to eat!

Edit... but I do think it's #3. FSD has been out of MSM, so Joe investor really didn't know what to believe. Heck, people on this forum are not convinced.