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

[Intl Professor]

Wow I am such a bad driver! Got side swiped two years ago, got hit in the wheel well in a parking lot a year ago. Got hit by a bycle after came to a complete stop a year ago. got rear ended on the highway 4 months ago! None of those are my fault but man that was annoying.

Go to church. I think they're out to get you!

 

[oneday]

Go to church. I think they're out to get you!

Does that really work Professor?

 

[JRP3]

Didn't Elon just recently make FSD/robotaxis the main and most important part of the TSLA investment thesis? And wasn't the cap raise investor call largely about this as well? It seems like a relevant and timely topic for this thread as it has huge implications for TSLA both present and future.

Autonomy day is old news at this point. We have no new data to discuss, hence irrelevant. The minutia of FSD can be endlessly discussed elsewhere.

 

[Dreadnought]

It's actually an attempt to build in a cushion that accounts for the lunacy of everyone else on the road
<snip/>

The principle I picked up from my defensive driving courses was "Assume everyone else is a maniac and will do the stupidest, most dangerous possible thing".

Ha! I knew all along that at one point you'll admit that FSD is simple
Kidding aside, this is all the strategy you need: Who's going to cross my path next? What's my preferred evasive action? Who's the most likely to cross my path in that case?

And no, this is not only about simple physics. We're not navigating between planets that move along fix orbits. The other particles on the road accelerate, decelerate and change their direction at will of other drivers, albeit obeying the laws of physics (and traffic regulations to some extent). Like a dog needs no physics degree to predict the trajectory of a ball after it bounced off multiple walls, neither the human driver nor the FSD car needs any psychology to predict others' behavior. Just enough training data to predict the most likely next actions is all it takes. If unsure, increase distance. In Germany, some drivers even give you hints about their intentions through signalling.

Sensing is IMO still the crucial element and with HW 2 it's not solved. Possibly HW 3 and compatible NNs but I doubt it at this time. Accuracy and recall have to be really high to even match average human level. Several 9s here to be chased.

What about "surprises", e.g. a kid chasing a ball, deer jumping out of the woods or an accident behind a curve?
These are a function of braking distance (or whatever lead time other evasive action needs), sensor range and what we consider to be an acceptable risk.
The first two are not difficult for a machine to determine. The FSD developers will at some point need to define the acceptable maximum chance for hitting something that the car sensed too late while e.g. driving through a forest at night or going through a residential area in the afternoon. This will determine a maximum speed. Consequence might be that FSD cars will generally drive slower than humans. But who cares if one can use the time to catch up with this thread, right?

tl;dr: FSD is literally right around the corner and who needs insurance then?

 

[Intl Professor]

Does that really work Professor?

Some say it does but ElefIno. Praying for others does, at least once a day I try to be unselfish. Such prayer is really for ourselves, the forgetting of self. Buddha says good thoughts lead to good consequences. That doesn't mean bad consequences don't happen by themselves and bad acts do result in bad karma. I don't think the Tesla did anything bad and I like Remus' posts, evidence of good which has the effect of good karma, according to the enlightened one. So as Yul Brynner said in the King and I, "it is a puzzlement" he or his car attracts bad drivers. Maybe he should pray for bad drivers.

 

[Todd Burch]

I was thinking about this yesterday, I am really curious to analyze his body language whenever Musk talks boldy in the presentation. Anyone pick up on anything?

I’ve been watching Musk in just about everything since the late 2000s.

I have never seen him more assertive and confident (and more frustrated that others don’t see it) than at the autonomy event.

 

[Todd Burch]

I think you're assuming that FSD needs electric cars. But I'm not assuming that. Most of the cars that comma.ai puts their camera system on are ICE cars (albeit cars that already have factory-installed mobileye camera/system that comma.ai basically hacks into).

Tesla is definitely in the lead, but others will find a way to catch up and solve FSD.

FSD is basically requires three things:
1. Hardware on cars
2. Software
3. Data

Currently Tesla has a lead in all three areas.

But Tesla is far from a monopoly on those three areas.

Dave, I’m kinda shocked at your views on this.

If you are a robotaxi customer, and you have the choice between a noisy, vibrating robotaxi versus a quiet, smooth EV at several times cheaper per mile, which would you choose?

 

[Intl Professor]

Does that really work Professor?

There's an alternative response to this.

When he was ten I asked my oldest son if he believed in baptism. His answer, "Hell, Dad, I've seen it done."

 

[Unpilot]

Some say it does but ElefIno. Praying for others does, at least once a day I try to be unselfish. Such prayer is really for ourselves, the forgetting of self. Buddha says good thoughts lead to good consequences. That doesn't mean bad consequences don't happen by themselves and bad acts do result in bad karma. I don't think the Tesla did anything bad and I like Remus' posts, evidence of good which has the effect of good karma, according to the enlightened one. So as Yul Brynner said in the King and I, "it is a puzzlement" he or his car attracts bad drivers. Maybe he should pray for bad drivers.

OT/
Well I think happy thoughts make a difference...to you and those around you which can have a ripple effect.
But praying in the sense the vast majority of people in the USA do it...um no......no effect.
Many study's into the effect of prayer show no effect.

But beers with a friend and some jokes work wonders for me.

 

[Unpilot]

There's an alternative response to this.

When he was ten I asked my oldest son if he believed in baptism. His answer, "Hell, Dad, I've seen it done."

You mean he saw a person have H2O poured over someones head?

 

[JRP3]

You mean he saw a person have H2O poured over someones head?

You can see that every day in my shower. That's not an invitation however.

 

[M3Rider]

If you look closely at the video you'll see how the Tesla curiously centered on the middle lane at 0:08 before it went back into the original right side lane again. There's very few drivers who'll have the skills (and interest) to perform two evasive maneuvers in 3 seconds the whole incident took and center on the lane they are spending about 1 second in ...

But to the FSD code driving the car it's just another workday in the computer: there were ~800 milliseconds spent in the middle lane which is plenty of time to center on the lane.

It also takes significant skill and steel nerves not to overreact here, to not steer to the left too much, to not oversteer and fishtail. Instead all three legs of the driving maneuver were performed almost perfectly, taking the dynamic properties of such a violent maneuver perfectly into account.

Based on that I'd guess there's a significant chance that this was a variant of the "Emergency Lane Keep Assist" FSD feature, which was activated last week IIRC.

Honestly, after an accident this becomes a gray area where I'm not sure about what should happen, because there could be a damage to sensors and misinterpretation of what the car "sees".

Maybe it can diagnose them within some milliseconds and decide they are still good and then figure out what's next. Or navigate based on partial sensors still remaining. This is really mind boggling.

There was another poster in the reddit thread claiming the AP wrestled control away from him and changed 2 lanes to avoid side collision.

That would be a total of 2 claims to-date where AP automatically leaves your current lane.
I am still not sure this has really happened, because it is not documented.

I think this is freakish somewhat if you have to encounter a scenario where AP takes away your control and drives all over the place - especially if you didn't notice where the danger comes from. This kinda needs documented, so you know what to expect.

 

[AcesDealt]

Pretty good and reasonable back and forth on robotaxis today. Mods have to decide when to toss such things as OT fyi.

Tesla Network will start with a large number of constraints, and only over time and iteration will those constraints get lifted. From a consumer perspective this looks like 'Most of the time I want to use the network but I can't because it says 'no' or a car is unavailable'. If Tesla pushes too hard then a large number of trips will end with 'We are sorry but we can can't continue. Please get out of the vehicle. An Uber or Tesla Network Associate has been called to pick you up. Your flight leaves in 15 minutes but we now have acquired a more tolerant airline who understands our level of innovation. Please consider it on your next trip. '.

 

[Zhelko Dimic]

A vision self driving system needs a supercomputer, redundancy built into the car control systems, loads of wiring for the sensors etc. Retrofit cost for a real self driving solution will be extremely expensive so very hard to get into consumer's cars. Comma.AI is 12 people working on retrofit driver assist, they do not seem to have any credible path to full self driving - i'm not sure this is even their ambition.
Waymo could choose to admit their Lidar first approach is wrong and drop everything to try to copy Tesla, but this is still extremely difficult to do if you are not a vertically integrated EV manufacturer where all systems have already been designed for redundancy and are set up to be software controlled.
Traditional auto manufacturers are also extremely slow to change their car designs, taking many many years to go from design to production. I think if Waymo/other Lidar companies were to pivot to copying Tesla's approach, it would be several years before they got the hardware rolled out into consumer's cars in significant numbers, and then they would also be very far behind in building the fleet data collection & filtering systems and other software 2.0 infrastructure Tesla has already spent years building and improving, as well as behind on building up annotated datasets/training cases for 100k+ driving scenarios.

Software teams often have few teams working on different generations of software, up to two generations ahead. It's likely, almost certain that Waymo already has teams that are working on pure vision based systems. It would be inconceivable that they're not hedging.
I should qualify this approach isn't useful for all problem area, and is not always used, but this one is perfect fit.

 

[Artful Dodger]

Meanwhile, (in the actual Market) another huge trade:

18-11 hrs 577,197 @ 254.13

This is over 5.6% of today's total volume (including this trade).

This is 2.5x the volume trade at close:
226,794 @ 16-00 hrs

We had similar large trades after hrs last week. I think someone is accumulating.

 

[Krugerrand]

My guess is that when people say Tesla’s lead is insurmountable, what they really mean is that with a two to three-year headstart in the industry, they will get a ton of money early and become the de facto standard in the business. It will be incredibly difficult for anyone else to overcome. Eventually though, I think competition will come and start a price war down.

Kinda like all the EV competition that was coming that would cause an EV price war and put Tesla out of business?

Nobody has managed to even come close to a Tesla vehicle to date and none on the horizon. Wait...what? Was that Tesla that reduced prices so much earlier this year it caused an all out panic? So who’s pricing whom out of the EV market? Oh, right. There’s a demand problem - but the competition is coming and it’s coming into a market with no demand.

Point being, when someone actually shows up with a fleet of connected cars that can out perform, out range, out charge, and out FSD Tesla, then and only then will you get me out of Mother’s basement. Until then, talk is cheap and I’ve pizza to eat and zombies to blow up.

 

[Unpilot]

Meanwhile, (in the actual Market) another huge trade:

18-11 hrs 577,197 @ 254.13

This is over 5.6% of today's total volume (including this trade).

This is 2.5x the volume trade at close:
226,794 @ 16-00 hrs

We had similar large trades after hrs last week. I think someone is accumulating.

Would that large of an after hours trade not move the needle more?

 

[Todd Burch]

The extended warranty & extended service plan were originally comically mispriced. Anyone who knew very little (including me) could see it was too cheap. They doubled the price a couple of months later. It was still a bit too cheap.

I did not get the extended warranty on my S. I spent a total of $700 on out-of-warranty repairs (failed door handle) between 50,000 and 100,000 miles.

In my case at least, Tesla would have come out well ahead had I bought the extended warranty.

 

[Carl Raymond]

“Current investors should find this news alarming for several reasons. The idea that a car will miraculously increase fivefold or more in value based on the ability to use it to generate income is nonsense. Tell that to the millions of Uber and Lyft drivers who use their cars for revenue generation every day. I cringe every time I hear Musk make that comment.“

Lol, this is a quote from Donn Bailey, Seeking Alpha.
Tesla: Elon Musk Moves The Goal Posts, Again - Tesla, Inc. (NASDAQ:TSLA) | Seeking Alpha

Such a dumb deception. Regardless of the actual timeline to FSD, there is no comparison between a driven car and a driverless car when calculating earning potential.

After reading that, I figure every paragraph in the article is written with the objective to deceive. Some FUD is well written. It must be embarrassing to be a writer of poor quality FUD.

 

[Reciprocity]

I think economics will require it before regulators do. There will certainly be room for ICE vehicles in autonomous fleets for 2-5 years— autonomy is a massive price advantage over fleets with drivers— but the $/mile advantage of EVs will push fleets to adopt a higher & higher % of EVs until they hit 100%.

If Google is confident in their software, they should be buying the largest stake they can in either Tesla or GM (or maybe Nissan?)— they’ll want to pump out as many EVs as possible as quickly as possible, even if ICE vehicles are initially part of their fleet. The aesthetic deficiencies of the Bolt and Leaf will be much less of a problem for commercial adoption. I think Tesla would be the best choice, but I could see them opting for GM or Nissan if Musk won’t give Google as much control as they would want (presumably because of his confidence in Tesla’s progress in autonomy).

If Google isn’t confident in their software, they should be buying Tesla AND GM. They have the market cap & cash to do it, and it would guarantee they achieve robotaxis first (whether from Waymo or Tesla or Cruise), and the manufacturing capacity to crank out EVs quickly. Tesla’s access to batteries would be particularly vital.

Tesla isn't for sale and Tesla doesn't need Google, Google needs Tesla. For software and hardware and cars and batteries. What does Tesla need from Google? Cash? They can get so much as they want or are willing to go get.