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

[mejojo]

...
This is a HUGE step forward for the company's valuation. I don't really care that much about autopilot as an investor, but others do, and there is a massive disruption happening here. If a huge chunk of lyft and uber stock doesnt get sold and converted to tsla stock soon, its just proof nobody is paying attention.

Tempted to short both.

What I'm confused about is what will a robotaxi do when a situation arises that even a human does not know what to do. e.g. crash ahead in the rain and police signalling to go some other direction than planned. Will they have someone in the mothership connect to the car and resolve as a remote driver? Can someone explain the plan?

 

[Fred42]

P.S. Mark my words: I predict that in 20 years Tesla will be building autonomous mining equipment to operate in the Asteroid belt. With Tesla A.I. hardware, and boost services purchased from SpaceX. There's a whole lotta nickel to be found on near Earth asteroids.

It's also possible that in 20 years Elon will be living on Mars and Tesla management will be maximizing return on what Tesla accomplished in the past and minimizing expenditure on innovation.

 

[Fred42]

At 6:35 he says "I just want to share the experience as Tesla has asked me to do." I was unaware that Tesla asked owners to share their experiences.

There's a perception issue with these videos. Tesla owners try the new release and are impressed with the improvements. Others view the videos and note that it doesn't drive as well as a human and have a genuine negative impression.

 

[j6Lpi429@3j]

What I'm confused about is what will a robotaxi do when a situation arises that even a human does not know what to do. e.g. crash ahead in the rain and police signalling to go some other direction than planned. Will they have someone in the mothership connect to the car and resolve as a remote driver? Can someone explain the plan?

as an interim measure, before you had actual robotaxis, you could still get a LOT of value from a system where, for example an uber driver had an FSD tesla, and was basically just a passive custodian sat in the drivers seat. They could happily read a book, play games on their phone, even watch a movie on headphones while 'almost-robotaxi' did its thing, just for those 1 in 50 trips where the car would play a beeping noise to the 'driver' to tell them they need to handle whatvere just caused them to stop.
That would also be the case for flat tires, and when another car crashed into the taxi, etc, all those 1/1000 possibilities...

Obviously that means you dont vaguely get the revenue of a true robotaxi... but given that driving is tiring, what uber/lyft driver is going to drive a non tesla (drive all day) when they can nap/read/chill all day in a tesla? And if you *are* driving a tesla anyway why not sign up to teslas own ride-hailing service instead of uber?

FWIW I think true global FSD is years and years off. We recently nearly hit a cow in a dark lane on a bend in heavy rain. No road markings, narrow road, hedges both sides... basically FSD hell...
...but I think for *certain markets* like taxi trips in western city centers, tesla will crush the market quite soon.

 

[dakh]

What I'm confused about is what will a robotaxi do when a situation arises that even a human does not know what to do. e.g. crash ahead in the rain and police signalling to go some other direction than planned. Will they have someone in the mothership connect to the car and resolve as a remote driver? Can someone explain the plan?

There are a lot of whacky things... Like if there's a deadlock with a string of cars, humans can see that and if the first car wants to back up, the others can understand what's up and the whole line backs up. Or if a long truck tries to turn but doesn't fit because there's someone on the line at a stoplight, they usually are aware enough to back up or do something to unblock the truck.

But don't underestimate everyone's potential to adapt. If police needs cars to go a certain way, I'm sure they'll figure out how to get Teslas with no driver to do what they want.

 

[OrthoSurg]

This sort of driving may trigger a training video upload.

The Tesla solution will want to know why your behavior varied from what Autopilot would have done.

Perhaps humans need to label potholes, the NN needs to be trained to recognise potholes and they need to adjust the rules.

E.g. Aviod potholes when there is sufficient clear driveable space. That is no different to avoiding other stuff. If there is debris on the road FSD also wants to avoid that.

To avoid roads with numerous potholes by maintaining the suggested speed, you drive very differently depending on how many cars are following you and which lanes behind you the cars are following. To drive securely by avoiding potholes you need to have good cameras behind, which are often obstructed with the bad weather during spring when all the brownish ice and snow on the road is melting

 

[StealthP3D]

I don't have any argument about the reality of Apple's attempt to explore the EV space. Perhaps it was very serious. Perhaps they even spent billions on it (although I doubt it got that far).

That's right. They kept it real quiet. It's almost like they knew from the get-go they didn't have the required technical expertise or the amount of innovation required to develop their own EV inhouse (even with hiring outside help from the industry). At least they were smart enough to pull the plug (so to speak) before things got more out of hand. The number of people who argued Apples' foray into EV's would be a huge success was mind-boggling to me. It's as if they thought EV's were easy, you just hook the red wire to "+" and the black wire to "-" and off you go! LOL!

We will see how competitive their autonomous driving software becomes (or if they ever release anything). My guess is it will never amount to anything that could be considered even remotely competitive with Tesla and it too will die a quiet death. In fact, the latest FSD beta release from Tesla might save Apple billions if it causes them to give up sooner rather than later. It would make more sense for Tesla to make a move into the smartphone market than for Apple to have aspirations about being competitive in the EV/Autonomy space.

 

[BigPeePee]

That's right. They kept it real quiet. It's almost like they knew from the get-go they didn't have the required technical expertise or the amount of innovation required to develop their own EV inhouse (even with hiring outside help from the industry). At least they were smart enough to pull the plug (so to speak) before things got more out of hand. The number of people who argued Apples' foray into EV's would be a huge success was mind-boggling to me. It's as if they thought EV's were easy, you just hook the red wire to "+" and the black wire to "-" and off you go! LOL!

We will see how competitive their autonomous driving software becomes (or if they ever release anything). My guess is it will never amount to anything that could be considered even remotely competitive with Tesla and it too will die a quiet death. In fact, the latest FSD beta release from Tesla might save Apple billions if it causes them to give up sooner rather than later. It would make more sense for Tesla to make a move into the smartphone market than for Apple to have aspirations about being competitive in the EV/Autonomy space.

I agree. Bleeding-edge engineering and manufacturing have never been Apple's forte, which is why Apple went to traditional car part suppliers. They knew they could never make an affordable EV with current car part suppliers, which was why the project was canned. The thought of designing and manufacturing their own parts never crossed their mind.

I wouldn't say Apple is gone for good in the EV industry. They may be waiting for costs to come down before trying again.

 

[Sanny]

Is the FSD price going up tomorrow? Or is it Thursday like someone said? I’m seriously thinking about pulling the trigger. I think it would add at least $4000 of value to my car, so it’s really only like spending $4000 instead of $8000. As an investor I need to be able to watch firsthand how Tesla is progressing, and to me that seems like a good enough reason to blow some money I’ll probably never miss. I especially won’t miss $8,000 if the stock price is at $40,000 in ten years.

Yeah, one of the reasons I'm using when barganing with myself. Honestly still can not fully justify to shell out $8k on something I will not really use.
Just to keep as a robotaxi play? But robotaxi scenarious are really super hard to predict currently, who knows how it will be played out even with a successful fsd implementation.

 

[OrthoSurg]

Why?
FSD is a VISION based system. You can see the potholes. So will FSD. It's just another thing to recognize and react to. I would change your timeline from 10 years to maybe 10 months.

Dan

Because the center console keeps repeating that the cameras and sensors are obstructed when I drive in the melting brownish snow during January, February, March, April. Even if a clean them before driving. So we will have to agree to disagree that this kind of driving relying on cameras getting constantly dirt on them while driving in a mixture of sand, brown snow, salt. The day there are camera washer apertures to clean the back up cameras I will totally agree it is feasible. But after 2 years driving my Tesla with constant blocked sensors and cameras, there are 4 months during the year which it will not be 100% reliable.

 

[Dan Detweiler]

Because the center console keeps repeating that the cameras and sensors are obstructed when I drive in the melting brownish snow during January, February, March, April. Even if a clean them before driving. So we will have to agree to disagree that this kind of driving relying on cameras getting constantly dirt on them while driving in a mixture of sand, brown snow, salt. The day there are camera washer apertures to clean the back up cameras I will totally agree it is feasible. But after 2 years driving my Tesla with constant blocked sensors and cameras, there are 4 months during the year which it will not be 100% reliable.

...and you don't think Tesla is aware of this? After what they have accomplished over the last decade or so I am willing to give their engineers the benefit of the doubt that they will figure all this out, if they haven't already and just not released it.

Ye of such little faith.

Dan

 

[StealthP3D]

Yeah, one of the reasons I'm using when barganing with myself. Honestly still can not fully justify to shell out $8k on something I will not really use.
Just to keep as a robotaxi play? But robotaxi scenarious are really super hard to predict currently, who knows how it will be played out even with a successful fsd implementation.

Yeah, I thought it was dumb to shell out over $60 bucks for a bottle of whiskey until I started experimenting with different Scotches. Now it seems dumb to shell out $30 for a bottle of bourbon.

Likewise, I bought AP/FSD originally because I wanted to experiment with it (and I liked the idea of supporting Tesla during one of their most trying times) - I didn't think the functionality would be worth thousands of dollars. But I didn't know what I was missing until I started using it and now, I wouldn't be caught dead without Autopilot/FSD because I learned what a useful driver aid it is. Of course, exactly how much each is "worth" will vary with different individuals depending upon how much discretionary capital they have. And that's no different than the different costs of single malt scotches vs. a typical bourbon. Value is very variable from individual to individual. Some people NEVER buy a new car for this very reason (and we NEED people like that to reduce wastefulness of people who like new cars). Some people would never eat out (too expensive) yet they might think nothing of plunking down $2000 for a TV. Another person might eat out all the time yet feel it was wasteful to spend ANY money on a TV. With enough discretional capital, most people will think nothing of spending plenty on both.

As to your particular dilemma, just give Tesla your money if you have it. It's one of the most environmentally friendly ways to spend a nice chunk of change and feel like you've splurged a bit.

 

[OrthoSurg]

...and you don't think Tesla is aware of this? After what they have accomplished over the last decade or so I am willing to give their engineers the benefit of the doubt that they will figure all this out, if they haven't already and just not released it.

Ye of such little faith.

Dan

I have seen a liquid washer cleaning aperture on the rear camera on the Porsche Taycan.
I will be really happy when I see it on the future Teslas too, however it is not there on my Model 3 and might have trouble robotaxiing during these months of the year. And might be problematic for potholes. But we could Ben surprised.
Do you have faith it can be retrofitted?

 

[lafrisbee]

Thursday

HA! to hell you won't $8k in stock today will be $700k in 2030. But yeah, get the FSD you'll get more giggles in your life than $700k can give you in 10 years

 

[BigPeePee]

You guys have to be realistic.

$40,000 stock price? LOL... A market cap of $40 Trillion by 2030? C'mon guys.

 

[amolina]

here is the latest from Brandon on 2020.40.8.11 FSD BETA:

Gist:

-The driver is overly cautious and pulls the car out of autopilot quite often. It is difficult to be 100 % sure if he really had to disengage all the times. Haven't been inside the car so cannot be certain.
-Claims that the new FSD is jerkier compared to the regular autopilot.
-Swerves a lot as well and is more aggressive.

 

[Intl Professor]

The previous owners trashed it, and didn't have a security deposit.

We bought TSLA, a down payment on a different future available today.

 

[MC3OZ]

What I'm confused about is what will a robotaxi do when a situation arises that even a human does not know what to do. e.g. crash ahead in the rain and police signalling to go some other direction than planned. Will they have someone in the mothership connect to the car and resolve as a remote driver? Can someone explain the plan?

Mostly the plan is simply 'Do what a human driver does in this situation". They car needs to understand the police signal.
If it gets confused (and it is safe to do so) a Robotaxi may simply pull off the road.
A remote driver make take control... Perhaps immediately or (in the case of car pulling over) sometime later.
A remote diver needs internet connectivity to the car, then a small team of remote drivers can support a large fleet as interventions are rare.

As soon as the first Robotaxi encounters the problem, the network will attempt to route al other cars around it.

The other thing that is mostly likely is that speed will based based on the perceived certainty / safety of the situation.
When a Robotaxi encounters an unusual situation it is likely to slow down. IMO that is what good drivers naturally do.
Bad drivers can become more aggressive in tricky situations, humans underestimate the extent to which emotions or a lack of attention case them to make bad decisions. A Robotaxi is always attentive and never emotional.

 

[Singuy]

here is the latest from Brandon on 2020.40.8.11 FSD BETA:

Gist:

-The driver is overly cautious and pulls the car out of autopilot quite often. It is difficult to be 100 % sure if he really had to disengage all the times. Haven't been inside the car so cannot be certain.
-Claims that the new FSD is jerkier compared to the regular autopilot.
-Swerves a lot as well and is more aggressive.

This guy's fsd fails a lot. Would love for another person attempt the same things on the same streets just to see if it's his car.

 

[StealthP3D]

Because the center console keeps repeating that the cameras and sensors are obstructed when I drive in the melting brownish snow during January, February, March, April. Even if a clean them before driving. So we will have to agree to disagree that this kind of driving relying on cameras getting constantly dirt on them while driving in a mixture of sand, brown snow, salt. The day there are camera washer apertures to clean the back up cameras I will totally agree it is feasible. But after 2 years driving my Tesla with constant blocked sensors and cameras, there are 4 months during the year which it will not be 100% reliable.

Handy tips from a snow skier:

While dirt/slush/snow etc. can and will obscure the sensors when conditions are messy, if it's happening more than you think conditions would warrant, it might be water intrusion (into the camera area) causing fog. I've had this problem on one of my B pillar cameras as the car warms up and causes any moisture in there to condense. Water intrusion is covered under warranty. Also, the windshield behind the front camera cluster can fog if it's really cold and humid and the windshield defroster has already turned off. If that's not your problem, keeping a good coat of old-fashioned carnauba auto wax on the camera areas can help keep the cameras cleaner than they would otherwise be. I've also got in the habit of having a microfiber rag handy and give the rear camera a very light swipe before or after I've been driving in the winter slop. This is the same rag I use to keep the road film from making my headlights dimmer. I've found it won't scratch the lenses, even when done frequently, as long as a light touch is used. Don't rub back and forth with pressure, just lightly swipe the headlights and camera lenses surfaces to clean off the film.

The radar mounted behind the front bumper can become obscured by wet snow or slush sticking to the bumper. Rust-oleum makes a product called "NeverWet" that will pretty much eliminate this problem. It's a hydrophobic spray and the only downside is it's not invisible - it has a frosty translucent appearance that prevents it's use on windows and camera lenses. But it will wash off your bumper with warm water and soap. Don't use the base coat (part A) on your car (just the hydrophobic part B).

I'm sure as FSD becomes a reality all new cars will be equipped with automatic sensor cleaning systems. I think Tesla has decided to let the cost curve of EV's come down to better compete with ICE cars before they start adding sensor cleaning systems to their vehicles. We are still in the very early days of mass-production of EV's so some manual cleaning is still required.