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

[Mike Ambler]

Let's assume a conservative by the car every 2 minutes as of today. One accident every 10 years should be good enough to count as mission accomplished (fatal accidents would be a small fraction of that). 10 years = 5.256 million minutes. This equals to 5.256/2 improvement over current capabilities or log(2.63M) of '9's, resulting in 37 iterations.
TL;DR: FSD in six months maybe, one year definitely.

Wow! I wonder if the capture & "solve" rate would slow, as cases get edgier and edgier...i.e. it's easier to pick the low hanging fruit, so to speak. So could 6 months turn into a year or two?

When they get a single odd case, could the system get good enough to train for it once and be confident to deal with it...if it ever happens again.

Second though bubble / question. What will be the SP blow out trigger? Something like...

After Y million/ billion miles driven without disengagement, Governor X of State Z is announcing that as of tomorrow, FSD Teslas can legally drive state roads without driver...

 

[JBRR]

Tesla to build lithium hydroxide refinery in Texas to feed Terafactory; first automaker to enter lithium (Exclusive) | Benchmark Mineral Intelligence

 

[Tim S]

I wonder if they will change the way software updates work on the cars then. Right now a software update takes 20+ minutes where you can't move the car at all. This is fine for weekly or monthly updates but if they plan on pushing new FSD code daily they need to come up with an update method which doesn't disable the whole car for half an hour at a time.

You can just set the software update for a time when you don't need the car, e.g. when you are asleep.

 

[Tim S]

Tesla to build lithium hydroxide refinery in Texas to feed Terafactory; first automaker to enter lithium (Exclusive) | Benchmark Mineral Intelligence

FYI that story is a month old (Sep 28).

 

[PeterJA]

Completely disagree on that one. Short-haul flights need to disappear, yes. But: Short-haul mass transit (flights are, Teslas are not) needs to be transitioned to high-speed rail. See most of Europe, China, Japan, etc.

It's a bandwidth problem: the need to efficiently move 200 people that all want to get from A to B cannot be solved by 100 - 200 EVs with FSD going the same direction.

View attachment 603734
Source: International Sustainable Solutions

High-speed rail is also a great infrastructure investment for a country transitioning to sustainability....

And yes: I get the irony of suggesting "cramming" 200 people into a metal tube during a pandemic. But this too shall pass...

No, rail is doomed.

As usual, it's not easy to foresee all the implications of revolutionary new technology, but I think I see a few.

1) When computers finish learning to drive, they will not drive like humans. A robo-car/truck that can see in all directions and react much faster than humans will safely drive much faster and/or with less space between moving vehicles. Therefore, robo-autos that react 10 times faster will increase a highway's "bandwidth" (potential throughput) by 10 times, through some combination of faster speed and closer spacing. Every traffic flow during rush hours will be a high-speed convoy.

2) Human driving of cars will not disappear, just as horseback riding has not disappeared, but it will be restricted to "human-speed" roads. Horses are currently prohibited on interstates. Eventually human-driven cars will be too.

3) Other current forms of land transportation will gradually disappear because robo-autos will be cheaper, faster and/or more convenient. Why load freight containers on-and-off a train (and wait for the train to depart) when robotrucks will take them directly from loading dock to loading dock (and depart immediately when loaded)? Why load people on-and-off buses and light rail when robocars will take them from door to door on their own schedules?

4) Future forms of land transportation (tunnels, hyperloop) will still have some advantages, such as protection from weather and very high speed. But the need for them will be reduced, because highway speed and bandwidth will be increased and people will sleep while traveling overnight.

5) Future challenges for Tesla's material engineers will include making roads more durable and tire dust less toxic. Tesla will not run out of things to innovate.

 

[wipster]

No, rail is doomed.

As usual, it's not easy to foresee all the implications of revolutionary new technology, but I think I see a few.

1) When computers finish learning to drive, they will not drive like humans. A robo-car/truck that can see in all directions and react much faster than humans will safely drive much faster and/or with less space between moving vehicles. Therefore, robo-autos that react 10 times faster will increase a highway's "bandwidth" (potential throughput) by 10 times, through some combination of faster speed and closer spacing. Every traffic flow during rush hours will be a high-speed convoy.

2) Human driving of cars will not disappear, just as horseback riding has not disappeared, but it will be restricted to "human-speed" roads. Horses are currently prohibited on interstates. Eventually human-driven cars will be too.

3) Other current forms of land transportation will gradually disappear because robo-autos will be cheaper, faster and/or more convenient. Why load freight containers on-and-off a train (and wait for the train to depart) when robotrucks will take them directly from loading dock to loading dock (and depart immediately when loaded)? Why load people on-and-off buses and light rail when robocars will take them from door to door on their own schedules?

4) Future forms of land transportation (tunnels, hyperloop) will still have some advantages, such as protection from weather and very high speed. But the need for them will be reduced, because highway speed and bandwidth will be increased and people will sleep while traveling overnight.

5) Future challenges for Tesla's material engineers will include making roads more durable and tire dust less toxic. Tesla will not run out of things to innovate.

Mercy sakes alive, looks like we're going to have a convoy...

 

[EVortex]

Mercy sakes alive, looks like we're going to have a convoy...

As long as we are not headed for bear, we should be good.

 

[RichardL]

No, rail is doomed.

As usual, it's not easy to foresee all the implications of revolutionary new technology, but I think I see a few.

1) When computers finish learning to drive, they will not drive like humans. A robo-car/truck that can see in all directions and react much faster than humans will safely drive much faster and/or with less space between moving vehicles. Therefore, robo-autos that react 10 times faster will increase a highway's "bandwidth" (potential throughput) by 10 times, through some combination of faster speed and closer spacing. Every traffic flow during rush hours will be a high-speed convoy.

2) Human driving of cars will not disappear, just as horseback riding has not disappeared, but it will be restricted to "human-speed" roads. Horses are currently prohibited on interstates. Eventually human-driven cars will be too.

3) Other current forms of land transportation will gradually disappear because robo-autos will be cheaper, faster and/or more convenient. Why load freight containers on-and-off a train (and wait for the train to depart) when robotrucks will take them directly from loading dock to loading dock (and depart immediately when loaded)? Why load people on-and-off buses and light rail when robocars will take them from door to door on their own schedules?

4) Future forms of land transportation (tunnels, hyperloop) will still have some advantages, such as protection from weather and very high speed. But the need for them will be reduced, because highway speed and bandwidth will be increased and people will sleep while traveling overnight.

5) Future challenges for Tesla's material engineers will include making roads more durable and tire dust less toxic. Tesla will not run out of things to innovate.

This implies (to me at least) that the vehicles will need some way to identify a robo vs human driver in other vehicles. If it’s robo, you can convoy closely. Otherwise give them a wider berth

 

[Lessmog]

This implies (to me at least) that the vehicles will need some way to identify a robo vs human driver in other vehicles. If it’s robo, you can convoy closely. Otherwise give them a wider berth

Easy. Stick a Lidar detector on the roof. Done.

Late edit (slow today):
Why not call it Liar Mc Lieface?

 

[Dreadnought]

Wow! I wonder if the F-U capture & "solve" rate would slow, as cases get edgier and edgier...i.e. it's easier to pick the low hanging fruit, so to speak. So could 6 months turn into a year or two?

When they get a single odd case, could the system get good enough to train for it once and be confident to deal with it...if it ever happens again.

Second though bubble / question. What will be the SP blow out trigger? Something like...

After Y million/ billion miles driven without disengagement, Governor X of State Z is announcing that as of tomorrow, FSD Teslas can legally drive state roads without driver...

Please don't bet any money on these figures. It was simply a mental excercise triggered by your question and my curiosity about where the ballpark numbers provided by Elon would lead to.

The rare edge cases, by definition, occur less frequently, causing fewer interventions. Having enough training data is a matter of fleet size, so not an issue. Still, I expect them to become incrementally harder to resolve as the simple perception and planning tasks will have been solved already. The car will over time rather get stuck / misbehave in the same kind of complex situations that are challenging to unexperienced human drivers, too.

There has been some dabate upthread about whether driverless cars are allowed already. I believe that this is the case at least in Florida but it doesn't matter. Once the software is ready, legislation will follow suit. Tesla flipping the switch won't come totally out of the blue, though. The capabilities will gradually improve and many owners will see stretches of thousands of miles without taking over once. A couple of iterations before robust statistics prove that it is safer than humans, many will be yelling to release FSD already.

Somewhere along the way between early Beta testers constantly on the edge and a car that can be trusted to do its job will be a twilight zone of not-quite-there-yet and complacent drivers that consider themselves passengers. I have no idea how this will pan out.

 

[insaneoctane]

I've been wondering how well the current beta FSD knows the dimensional boundaries of the Tesla vehicles. In really tight spaces, you need to know the exact limits of your specific vehicle. Models S, X, 3 and Y all have unique proportions. I know that using the current autopark feature and the ultrasonics feedback while pulling into my garage do NOT promote confidence today. For FSD to be the next level, the vehicles will have to know the dimensions better than most IMO

 

[bkp_duke]

Easy. Stick a Lidar detector on the roof. Done.

Elon has fired people for less . . .

 

[OrthoSurg]

If I had more capital, I'd be deploying some here for sure. As it is, I need to gamble on when I deploy it if I'm going to have a hope of meeting my long-term share goal, and events might provide a better opportunity in just a few days. I'll feel like a winner either way.

I’m a bit on margin to be honest.
But I’ll pay that back in the next 2 months

 

[OrthoSurg]

I used some dry powder today, nothing crazy and sold some misc junk to make some room. If it dips lower than around here I'll margin up a little.

I was a little on margin already. Now a little more. If it drops to 300 I go a little more and it drop to 200 I go full margin

 

[kanweg]

1. Using Dojo as opposed to the existing trainer they may be moving from running training in 72h to running [the same] training in 7h, but
2. They are not running the same training, if only because they are shifting from '2.5D' to '4D'. and we are unsure if the 72h > 7h is a like-for-like comparison or also takes the 2.5>4D shift into account, and

while 4D is more complex, it may also be better at solving problems that 2.5D struggles with and requires tons of hours in an attempt to achieve marginal improvement. Whereas with 4D the issue To resolve has a somewhat Independent additional parameter to exploit. So, 4D may speed things up even without better hardware.

But what do I know.

 

[PeterJA]

This implies (to me at least) that the vehicles will need some way to identify a robo vs human driver in other vehicles. If it’s robo, you can convoy closely. Otherwise give them a wider berth

Human drivers on a robo-only highway will be identified by crashing. If they survive, they will be jailed or heavily fined.

Maybe a transition step to robo-only highways will be robo-only lanes on human highways -- like a carpooling lane, but with closely-spaced robocars whipping past the human traffic at twice the speed. That will be good advertising for robocars and will teach even cocky human drivers that they really can't survive on a robo-only highway.

 

[LTC_RRR]

I'm pretty sure it doesn't drive. In fact, in the Doug DeMuro review, you can clearly see that it is plugged in with an extension cord. I would bet the cord is "shore power" and that there's no functional battery powering the lights and systems.

I wonder if it comes with that extension cord or is it extra.

 

[Lessmog]

This implies (to me at least) that the vehicles will need some way to identify a robo vs human driver in other vehicles. If it’s robo, you can convoy closely. Otherwise give them a wider berth

Easy. Stick a Lidar detector on the roof. Done.

Late edit (slow today):
Why not call it Liar Mc Lieface?

Elon has fired people for less . . .

Only suggesting a partial detection method. Algorithm for differential diagnosis is another matter ... ... and also preferred remedy.

 

[OrthoSurg]

I am predicting a bounce next week. Tech sold off today due to massive record profits reported and great guidance...lol.

All my tech stocks are down almost 10% this week.

SHOP blew earnings out of the water
Fvrr beat expectations and raised guidance.
AMD had a massive beat and raised guidance
Tesla we all know

And from where I am sitting.
Amazon reported record profits
Apple is expected to sell record numbers of 5g cell phones with their service division exploding in growth
FB also blew up expectations.

This was the reason why tech saw a V shape recovery from March..and if it wasn't for...nonsense..we should see a check mark, not a V with these earnings.

When the markets are uncertain, it’s time to buy.

They are uncertain if Bidden wins if he will impose lockdown and at what degree he will raise taxes.
They are uncertain if bidden wins with low margin if Trump contests
They are uncertain if Trump wins if Covid will get more out of control than it is out of control.
But I am certain TSLA will be higher in 2025 than it is now so I am not selling to pay crazy taxes on my gains.

 

[Lessmog]

I wonder if it comes with that extension cord or is it extra.

Does it limit range? Will it ... entangle the entire ...a area ?