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

[tinm]

Our plumber is a NM native and told me there’s nothing more delectable than roasted cicadas on a kebab...

Ok @Artful Dodger now I gotta know, you disagree with a silly post like that because...

• no, my plumber never said that to me? (he did, just last week, grossed me out)
• you're grossed out to the point of marking the post "disagree"
• you disagree because some other crunchy critter kebab is tastier, and you know from experience?

I think I speak for everyone here when I say, we really gotta know.

(This is the craziest weekend OT ever)

 

[BioSehnsucht]

Another recent tweet on Summon:

"Parking lots are a remarkably hard problem. Doing an in-depth engineering review of Enhanced Summon later today."
Elon Musk on Twitter

Also, last week:

"Intersections with complex traffic lights & shopping mall parking lots are the two biggest software challenges. Developer branch already mostly works in these scenarios, but massive effort required to get to 99.9999% safety."
Elon Musk on Twitter

At first glance, Summon seems really easy. And it is, for simple uncrowded parking lots. But in a very busy parking lot, with many moving cars and pedestrians, where there are not even lanes or traffic lights, Summon is probably the hardest of all Self-Driving problems. Human drivers may rely on non-standard or subtle cues (for example, direct human to human communication), which are not possible in a car without a human driver (of course, Tesla can train its NN to recognize, for example, human hand gestures from other vehicles and pedestrians).

So this is going to be very interesting. What will Tesla do here? Looks like Summon is a much harder problem than initially thought. It could be quite a while till it is fully solved. Or does Tesla limit summon to non-busy parking lots (and how does Tesla determine this)? In short (IMO), be prepared for either a significant delay and/or significantly limited functionality on initial release.

I would not be surprised if Enhanced Summon ends up being FSD computer only (after all, it's part of FSD feature set isn't it?), and the solution is simply to throw more hardware at it and anyone on HW2.x has to wait for their FSD computer upgrade.

Similarly would not be surprised for intersections - recognizing a stoplight/sign and stopping to hand control over to a human is probably solvable on HW2.x (a nice safety net for HW2.x AP/EAP cruising on highways in the middle of nowhere coming upon an intersection) but accounting for all the possible variables as needed for FSD would probably need the extra horsepower ...

 

[BioSehnsucht]

I wonder whether they are using a 3rd party cloud - much faster to scale up than 1st party data centers.

At some point Karpathy refers briefly to compute budget, and that would imply 3rd party cloud.

They might have some cloud based compute resources currently but from the way they spoke about the new video (vs still image) processing NN system at the Autonomy Day it sounded like they were custom building something to handle it.

 

[BioSehnsucht]

Yes, there are two budgets. One is the total number of networks (I guess layers) they can have in the car. The second is training budget - which could also be a time constraint rather than $ - in which case could be either 1st party or 3rd.

BTW, it doesn't make sense to target HW2 for public release of City NOA. No point in constraining the NN to HW2 after spending all that effort on HW3.

This is what EM tweeted on Jul 7th.

Production fully switched over ~3 months ago. Functionality won’t diverge until Q4, as it’s limited by software validation. Will be later for Europe compared to rest of world due to regulatory constraints that were put in place years ago by big ICE companies.
​

So let's parse it.

"Functionality won’t diverge until Q4" = Same functionality for HW2 and HW3 until Q4. So same NN.
"as it’s limited by software validation" = The new NN for HW3 will finish validation in Q4 and then it will be pushed out to HW3 fleet. At that time we will start upgrading the HW2 fleet to HW3.

Could also be that the extra FSD/HW3 functionality would be locked out but running in shadow mode on FSD computer equipped cars, which would mean different NNs, but available-to-the-driver functionality would be the same. Part of the software validation is likely to let FSD NN's run on FSD computers in shadow mode. Might even be able to run both the HW2.x sized NN and portions of the FSD NN at the same time (with FSD in shadow mode) for validation (sort of like how they wanted to run MobilEye and AP1 at the same time).

 

[Artful Dodger]

View attachment 429390

Ok @Artful Dodger now I gotta know, you disagree with a silly post like that because...
<snip>
(This is the craziest weekend OT ever)

I Disagree with you and others replying to this continuous off-topic.

[MODS] Please delete this comment along with all the other Insecta posts. Thanks.

 

[EVNow]

Could also be that the extra FSD/HW3 functionality would be locked out but running in shadow mode on FSD computer equipped cars, which would mean different NNs, but available-to-the-driver functionality would be the same. Part of the software validation is likely to let FSD NN's run on FSD computers in shadow mode. Might even be able to run both the HW2.x sized NN and portions of the FSD NN at the same time (with FSD in shadow mode) for validation (sort of like how they wanted to run MobilEye and AP1 at the same time).

I doubt they will do all this - because making something work for HW2 & HW3 has significant costs. Unless, the NN is still too small, so can run on HW2 (but it sounded like there was a budget which they allocate to tasks - so they are worried about NN size). Listening to the talk - it is clear that optimization, training etc are quite complex and resource intensive (both computation and manual time). So, difficult to see why they would do it twice - once for HW3 and once for HW2.

I know they have talked about shadow mode, but not sure how much of it they use.

Bottomline, if they are using the same NN for both HW2 & HW3, then HW2 cars will get some of the new functionality. Otherwise only HW3 cars will get the new functionality.

 

[StealthP3D]

Excellent interview:

John McElroy doesn't seem to get it. In response to more knowledgable industry analysts who insisted all the big OEMs are scrambling to get on the electrification bandwagon, he kept saying, look, they make money at making gas vehicles, no one is making money with this electrification thing, so he doesn't think they will switch. He didn't seem to understand that demand for old school vehicles will decline so quickly, in the face of more desirable EV's with lower total cost of ownership and better performance, that they won't be able to sell enough gas cars to stay profitable anyway.

He acted like it's their choice to make: Profitable gas cars or unprofitable EV's. He made this point forcibly at least three or four times. Totally insane. He thinks like a dinosaur.

 

[StealthP3D]

I wonder whether they are using a 3rd party cloud - much faster to scale up than 1st party data centers.

At some point Karpathy refers briefly to compute budget, and that would imply 3rd party cloud.

I didn't listen to it but, in this context, "compute budget" generally doesn't refer to a financial budget, it's the finite amount of processing power you have to work with. It could include cloud computing (based upon the wireless bandwidth and latency limitations) but I doubt he was referring to the cloud.

 

[kengchang]

John McElroy doesn't seem to get it. In response to more knowledgable industry analysts who insisted all the big OEMs are scrambling to get on the electrification bandwagon, he kept saying, look, they make money at making gas vehicles, no one is making money with this electrification thing, so he doesn't think they will switch. He didn't seem to understand that demand for old school vehicles will decline so quickly, in the face of more desirable EV's with lower total cost of ownership and better performance, that they won't be able to sell enough gas cars to stay profitable anyway.

He acted like it's their choice to make: Profitable gas cars or unprofitable EV's. He made this point forcibly at least three or four times. Totally insane. He thinks like a dinosaur.

Then you are going to love his dismal Autopilot review video....

 

[StealthP3D]

Then you are going to love his dismal Autopilot review video....

I doubt I will waste my time watching it. John McElroy is an industry dinosaur who doesn't want to see the "glory days" of the American auto industry be displaced. What he doesn't seem to realize is that Tesla is America's best hope to regain industry leadership (that was lost decades ago)..

 

[FrankSG]

Trump administration freezing fuel efficiency penalties - Reuters

Bad for the world, good for Tesla.

Why do car manufacturers not understand that by lobbying for these things and delaying the inevitable, it'll only get worse for them? Nobody's going to stop electrification at this point, and by transitioning slower these car companies are only going to give Tesla (and Chinese car companies) a bigger lead. I think the Autoline show from this week made that painfully obvious.

 

[EinSV]

John McElroy doesn't seem to get it. In response to more knowledgable industry analysts who insisted all the big OEMs are scrambling to get on the electrification bandwagon, he kept saying, look, they make money at making gas vehicles, no one is making money with this electrification thing, so he doesn't think they will switch. He didn't seem to understand that demand for old school vehicles will decline so quickly, in the face of more desirable EV's with lower total cost of ownership and better performance, that they won't be able to sell enough gas cars to stay profitable anyway.

He acted like it's their choice to make: Profitable gas cars or unprofitable EV's. He made this point forcibly at least three or four times. Totally insane. He thinks like a dinosaur.

I really enjoyed watching this and I think the statements you mention still reflect the thinking of the vast majority of the legacy auto industry.

We see it in the recent comments of the BMW head of development that there is no demand for EVs — when the Model 3 is eating their lunch.

We see it in Daimler execs still trying to promote diesel despite repeated emissions cheating scandals.

We see it GM selling a grand total of ~8000 EVs in the first half of 2019.

The ICE dinosaurs are being dragged kicking and screaming into EVs because they can’t make money in the short run. We’ll see how they adapt.

 

[MarcusMaximus]

Trump administration freezing fuel efficiency penalties - Reuters

Bad for the world, good for Tesla.

Why do car manufacturers not understand that by lobbying for these things and delaying the inevitable, it'll only get worse for them? Nobody's going to stop electrification at this point, and by transitioning slower these car companies are only going to give Tesla (and Chinese car companies) a bigger lead. I think the Autoline show from this week made that painfully obvious.

As I recall, they were actually lobbying against this. Not because they want to be forced to be more fuel efficient, but because CA and a few other states are forcing them to be anyway. So, on any given vehicle, they can either abide minimally by the federal standard and be locked out of several of the most populated states or hit CA standards and be undercut on price everywhere else.

 

[Buckminster]

Just....I hope Musk does not experiment on himself. Using experimental Neuralink tech, fine, but using it on Musk is too big a risk.

Oops....

Okay, I got this:

(This is the craziest weekend OT ever)

Yup, my preference is for using the sub threads. Either way, I wouldn't want to miss out on all the Cockroach milk discussion for all the Cicadas in the world!

I love you people

 

[Artful Dodger]

he kept saying, look, they make money at making gas vehicles, no one is making money with this electrification thing, so he doesn't think they will switch

This is the single most compelling reason for Tesla to temper growth (just a small amount) and strive to include some baseline profits on an ongoing basis, while still growing at an enviable rate.

Other OEMs are NOT going to switch if they see that making only EVs isn't profitable, which is a common opinion in Autoline roundtables (not to mention in Detroit, Stuttgart, Tokyo and Yokohama). And ICEmakers WILL lose money during the switchover (hence the footdragging in ICE boardrooms).

But Telsa's mission is to ACCELERATE the transition. That means getting help making more EVs (and to convice the ICEmakers there's a better way). That means providing leadership on the business case: 'Yes, EVs are profitable once you've sunk in the investments required. Don't be afraid.'

My personal benchmark for leadership success is S&P 500 inclusion. IMHO, when TSLA achieves that, doubters can pound sand while they watch large swaths of institutions buying Tesla common stock. ICEmakers will switch. They'll be joiners, not disruptors. Their stogey old BoDs will like that.

Cheers!

P.S. This is not my newly held opinion. Indeed, it was my first post on joining TMC 11 months ago: (a topic which gets relitigated in this thread about once a month)

S&P500 listing and moderator actions (out of Market Action)

 

[MarcusMaximus]

This is the single most compeling reason for Tesla to temper growth and strive for profits on an ongoing basis. Other OEMs are NOT going to switch if they can't make money. They really can not make money while doing the switch to EVs (hence the footdragging in the boardrooms).

But Telsa's mission is to ACCELERATE the transition. That means getting help making more EVs. And that means providing leadership on the business case: Yes, EVs are very profitable once you've sunk in the investments required.

My benchmark is S&P 500 inclusion. IMHO, when TSLA achieves that, doubters can pound sand while they watch large swaths of institutions buying Tesla common stock.

P.S. This is not my newly held opinion. Indeed, it was my first post on joining TMC 11 months ago: (a topic which gets relitigated in this thread about once a month)

S&P500 listing and moderator actions (out of Market Action)

Alternatively, they could let the old guard sink and take over the world along side the BYD’s and Rivians.

 

[Singuy]

Excellent interview:

Man that quote about how there's an over saturation of Teslas in Cali and the minute someone drives a Taycan down the road, all these Tesla owners would be like "boy I need to get myself one of those!".

I find this statement to be hi-freaken-larious! Not one Tesla owner would think that. Every Tesla owner would think "why the F did this dumbass spend 85k on a car with low range, poor software, no charging stations, and no service?". Why no service?..because no Porsche dealership is going to train or hire an EV specialist when there are only 5 cars on the road. I have yet to hear one Tesla owner say "boy I wish I bought that iPace". Doesn't abysmal iPace sales give these guys a clue that until these legacy takes EV as seriously as Tesla, no one is going to buy them?

 

[RobStark]

As I recall, they were actually lobbying against this. Not because they want to be forced to be more fuel efficient, but because CA and a few other states are forcing them to be anyway. So, on any given vehicle, they can either abide minimally by the federal standard and be locked out of several of the most populated states or hit CA standards and be undercut on price everywhere else.

What the legacy OEMs want most of all is regulatory certainty. They want Trump and CA to negotiate a settlement. They probably prefer something closer to Trump but want certainty so they can plan out their product portfolio and budgets.

They can set a product portfolio for Central-Southern States and then have an international portfolio everywhere else. China,EU, CA and CARB states too. Squeeze the last cent of profit from ICE patents and also get ready for future.

GM and Nissan are also lobbying for extension of the Federal EV credit.

 

[Krugerrand]

Alternatively, they could let the old guard sink and take over the world along side the BYD’s and Rivians.

I vote for this. They’ve had almost a decade. How much longer do they deserve to have to drag their feet and spew their crap? Enough. Die already.

 

[KSilver2000]

Then a few days later, I had a serious technical problem where the car would not drive and I was stranded. I got through to a live human within 10 minutes,

You’re lucky the car would not drive. Otherwise, the automated menu prompts would’ve just put you straight to a voicemail.