Articles/megaposts by DaveT

[tlo]

Not sure about funds. But you can buy on the secondary market (ie., http://www.sharespost.com). However, it's a hassle (lots of paperwork and large minimum investment requirements).

ps., if anyone has bought shares on sharespost.com, please PM me and let me know how it went.

I've bought shares on sharespost before.

I think it's OK for me to share some details here:
- paperwork not bad at all, just a few e-signatures, a couple of phone calls, and a bank wire
- minimum transaction usually $100K, varies by deal
- one-time fee around 4-7%
- deal usually priced close to the public figure of the most recent funding round -- this is the main complication because you don't know the details of that funding round (i.e., Uber's valuation is said to be >$60B, but that number is inflated because there are special clauses that protect private investors)
- they transfer the shares to your broker as soon as they can, which is probably about one year after IPO when the lockup period ends

 

[generalenthu]

Wanted to throw in a wild guess here. Uber will shortly make an acquisition announcement for Volvo cars. Really no basis for this and less than 5% chance of this happening, but was thinking what Kalanick would do looking at the pace at which Tesla is moving. This is a bet the firm kind of moment for Uber.

 

[palmer_md]

Why buy when you can just set up a partnership to get a volume discount on the cars.

 

[generalenthu]

Same reason as why Tesla is buying SCTY. Also this years Nobel

Hart has investigated how best to write contracts when the possible outcomes are hazy. He changed the way that economists think about corporations, highlighting how firms can increase efficiency not only through competition, but also through cooperation — by contracting with each other. And sometimes, he concluded, it is more efficient for companies to simply merge because a single owner would make better overall decisions.

 

[SteveG3]

Do autonomous cars essentially need to be BEVs? How do gas based autonomous vehicles refuel themselves? Could Tesla's moat be autonomy + battery + data collection advantage?

there is something immensely more advantageous to EVs vs. ICE for this market than the fact that Tesla is working on having charging stations that autonomously recharge the vehicle.

 

[Alketi]

Confounding the Experts

Here's an absolutely fascinating illustration of how Tesla has utterly confounded even the so-called investing experts.

Kevin O'Leary, aka Mr. Wonderful from Shark Tank, opining that Tesla will fall due to gravity eventually, because --

"Tesla makes ONE thing. A car."

O’Leary: Gravity will strike Tesla eventually

Absolutely stunning. This implies that he either stopped listening or has no concept of Tesla's long-term vision, nor is he apparently aware of any impending disruptions to the car / transportation / driving / energy industries.

And, if you listen to Bears or doubters with any frequency, arguments reappear over and over. A non-scientific sampling:

1. Elon is a con-man. A modern-day PT Barnum. (the red Model 3 was made of wood, the self-driving video had camera cuts, etc)
2. Tesla loses money on every car sold.
3. Tesla made money but <fill-in-the-blanks> (i.e. cooked the books, didn't pay vendors, cashed in EV credits, etc)
4. Tesla won't tell us how many Model 3 orders they have. They're clearly dropping.
5. Self-driving cars will never happen.
6. Self-driving cars are not achievable with <fill-in-the blank> (i.e. Neural Networks, Camera/Radar, etc)
7. What happens if a self-driving car encounters a potential head-on collision, and all the escape routes are blocked? One sidewalk contains a family singing Christmas carols, and the other sidewalk contains a child on a tricycle carrying a basket full of puppies. How does it know what to do???

I suspect when you combine:

1. New technology
2. The implication of future disruption

With:

1. Bold claims
2. Losses due to expansion
3. Production issues
4. Mis-steps

That some people literally lose their minds. It's an amazing dichotomy.

As a side note, I thought it was interesting that Elon said the car owner would be keeping the majority of the Robotaxi profit. I didn't expect that, actually.

Spitballing numbers you quickly get to very large returns that probably shouldn't be believed at this point --

500K mile robotaxi
$1/mile
50/50 split
= $250K

 

[AudubonB]

====>I believe all the $ #s I have seen on this thread as well as the other ones in the Investor sector regarding Robotaxi per-miles....


...have shown me a disturbing tendency of their authors to, umm...unqualified greed<====

I'm trying to believe it's all right, though, because market forces ****even just WITHIN the Robotaxi milieu*** will conspire to bring the fare down to a market-clearing point, which in my way of thinking is closer to $0.18-.22 per mile.

You can - and clearly will - take this point as you will, but you might consider it and, I would suggest, back-test it by asking "Why would Robotaxi owner #2 not undercut #1 by just enough to grab that ride - as long as he is covering his marginal cost of operation?"

And that will mean, as I see it, an end to starry-eyed uber- super-profits.

No?

 

[techmaven]

====>I believe all the $ #s I have seen on this thread as well as the other ones in the Investor sector regarding Robotaxi per-miles....


...have shown me a disturbing tendency of their authors to, umm...unqualified greed<====

I'm trying to believe it's all right, though, because market forces ****even just WITHIN the Robotaxi milieu*** will conspire to bring the fare down to a market-clearing point, which in my way of thinking is closer to $0.18-.22 per mile.

You can - and clearly will - take this point as you will, but you might consider it and, I would suggest, back-test it by asking "Why would Robotaxi owner #2 not undercut #1 by just enough to grab that ride - as long as he is covering his marginal cost of operation?"

And that will mean, as I see it, an end to starry-eyed uber- super-profits.

No?

I think it will be a while before robo-taxi market saturation is to a point where such a thing exists. And obviously not everyone wants to participate, and there will still be surge in demand issues. So, yeah, some hours of some days it will be bare bones, but others would not be.

 

[ggies07]

IMO these are the reasons why Tesla will dominate in the public eye. Tesla has in place right now responsible/safe ways to collect the data. I would not want to get into an Uber car that had some 'kit' installed on a random normal car and in which I do not know how well their data network is in play.

Here we go, one down, how many more to go? This just reiterates what I said above. How is Uber going to show the public it's self-driving car/software is safe? Where will they get the data?

George Hotz cancels his Tesla Autopilot-like ‘comma one’ after request from NHTSA

 

[Ovulation]

Here we go, one down, how many more to go? This just reiterates what I said above. How is Uber going to show the public it's self-driving car/software is safe? Where will they get the data?

George Hotz cancels his Tesla Autopilot-like ‘comma one’ after request from NHTSA

Yeah its quite a smart move from Tesla to imply that do you need 10B miles to prove the car is saver than a human ...

 

[Johan]

Yeah its quite a smart move from Tesla to imply that do you need 10B miles to prove the car is saver than a human ...

Agreed, this "appeal to common sense" and "pulling a big number out of thin air" could be a tactic to block others. There are scientific methods by which you could calculate how much data you would need in order to show that automated driving is safer than humand driving, if you know the accident/event frequency in the human driving data set and given a desired confidence interval on the data. This is called calculating Statistical Power and is not at all difficult mathematically. Since accidents, even serious and fatal accidents, are quite common I would suppose (now it's me pulling numbers out of thin air) you could get very far with only a few hundred million miles of comparable driving data (if you accept someting like a 95% confidence interval, or a p<0.05 as significant).

 

[NOLA_Mike]

RE: Uber

I think another possibly underrated risk for Uber going forward is the current trend in court rulings that Uber drivers are employees versus independent contractors.

 

[Alketi]

Elon's number of 6 billion miles is a bit of an approximation, simply being 100x the world-wide fatality rate of one every 60 million miles.

He's approximating that you need a sample size of 100x your occurrence rate.

Of note, the Rand Corporation took a deep look at this and concluded that for statistical accuracy, you'd need to acquire anywhere from 8 billion to tens or hundreds of billions of miles, depending on how low the fatality rate got. [e.g., if the fatality rate was one every billion miles, you'd again need a sample size of ~100x that to feel confident about it]

Click the "Read Online" link to read the full report
Driving to Safety: How Many Miles of Driving Would It Take to Demonstrate Autonomous Vehicle Reliability? | RAND

I presume Tesla will try to prove (at least) some of this statistically to NHTSA through miles driven in shadow mode, logging both false positives and negatives.

Another interesting note in the CC was that Elon said they're acquiring 1.5 million miles per day on AP 1.0 hardware. That means every 2 days they eclipse Google's entire lifetime of fleet learning. The two systems are obviously not equivalent, but it does illustrate the advantage of fleet deployment vs test cars.

 

[GoTslaGo]

RE: Uber

I think another possibly underrated risk for Uber going forward is the current trend in court rulings that Uber drivers are employees versus independent contractors.

Here is one:

Europe|UK Uber Drivers Win Case to Get Paid Vacation, Minimum Wage

New York Times - 26m ago


LONDON - Uber drivers in Britain should get paid vacation days and be guaranteed a minimum wage, a tribunal said Friday, in a ruling that tests the limits of the so-called "gig economy," in which companies rely on individual to work as self-employed ...

http://www.nytimes.com/aponline/2016/10/28/world/europe/ap-eu-britain-uber.html

 

[MitchJi]

This left me scratching my head a bit. This moat is precisely what disappears in an autonomous world as one side of the market becomes moot. With full autonomy, it's just consumers buying a service from a single service provider--no value add in connecting them with one of many people currently available to drive.

Exactly. The only barrier for Tesla is the effort required to download a Tesla app.

@W84M3 In your cost factors, you're missing the actual cost of the car. Tesla's cars are expensive, and even with the Model 3 the ASP is likely going to be over $40k. Compare that to a $10k ICE car, and I think you probably will have lower cost of ownership (let's say over 5 years) with the ICE car, including fuel and maintenance.

That sounds like an argument for ICE cars vs EV's.

@dha Yes, I agree with you in that autonomous driving changes the equation. But whoever has the most cars (whether autonomous or not) resulting in the lowest wait times and the cheapest service, still has some advantage in providing a better service than the next competitor and that lead to more usage, which leads to increase in supply (since there's more revenue to purchase more cars), which leads to a better service. So, there's natural network effects still in place as more usage improves the service. Same goes for Tesla Network, if they can get it up and running successfully. The more people who use Tesla Network, will lead to more cars on their network, which will reduce wait times and reduce price... all a virtuous cycle.

Tesla should be able to be a lot less expensive (the cost advantages of EV's will increase), there are already cities in Europe and China where ICE's are prohibited or restricted, some people prefer to deal with green and ethical companies. M3 will be widespread in many areas. I might put two apps on my phone, and if a Tesla M3 isn't available in a convenient time frame check for an uber.

Uber's business model is screw the drivers, and Elon has said that he plans to pay more to the car owners. So uber will need to either start buying a ton of cars, like Bolts or compete with Tesla on price, and Tesla can afford to make almost nothing compared to uber, and still make a substantial profit, because for Tesla it's a feature that helps them sell more cars.

I think it's basically over and uber is already either severely wounded or lost.

 

[tlo]

All due respect to Uber, but I have to think they are fairly scared of Tesla right now.

Elon's quotes "most revenue will go to owners" and "it's not Uber vs. Tesla, it's Uber vs. people" is quite revealing.

Ride-sharing has always been a land grab where companies lose money to gain market share. Elon basically confirmed Tesla Network strategy is not to make money initially. And an advantage that hasn't been talked about is that Tesla can afford to do this because they are already making money on the cars (firstly from car sales and secondly from the full autonomy option).

Also, the virtuous cycle starts with people buying the cars. As long as Tesla makes the best cars, it'll be the only cars on the market people want to buy (especially the millennials who are also the most likely to participate in ride-sharing).

 

[DaveT]

====>I believe all the $ #s I have seen on this thread as well as the other ones in the Investor sector regarding Robotaxi per-miles....


...have shown me a disturbing tendency of their authors to, umm...unqualified greed<====

I'm trying to believe it's all right, though, because market forces ****even just WITHIN the Robotaxi milieu*** will conspire to bring the fare down to a market-clearing point, which in my way of thinking is closer to $0.18-.22 per mile.

You can - and clearly will - take this point as you will, but you might consider it and, I would suggest, back-test it by asking "Why would Robotaxi owner #2 not undercut #1 by just enough to grab that ride - as long as he is covering his marginal cost of operation?"

And that will mean, as I see it, an end to starry-eyed uber- super-profits.

No?

I generally agree that over time the ride-sharing/hailing market will move toward commodity pricing, especially when true autonomous (driver-absent) cars are abundant. However, there's still value in the actual network, logistics in organizing and directing thousands of cars and routes, and how the user interfaces with that network. In other words, just having a Level 5 car does nothing unless you've got the network, both in the sense of the cars and the riders.

 

[DaveT]

Here we go, one down, how many more to go? This just reiterates what I said above. How is Uber going to show the public it's self-driving car/software is safe? Where will they get the data?

George Hotz cancels his Tesla Autopilot-like ‘comma one’ after request from NHTSA

Yep, Tesla's advantage is that they're got a lead in integrating the hardware and also deploying the software that will lead to autonomous driving. In order for Uber to compete, they need to take some drastic measures. I'm expecting a big, bold move by Uber by the end of next year... perhaps partnering with an auto maker to create an "Uber" car that has Level 4/5 hardware, OTA software updates, etc. so they can get miles on the road.

 

[DaveT]

Yeah its quite a smart move from Tesla to imply that do you need 10B miles to prove the car is saver than a human ...

Yes, interesting take. I think it was 6B miles Elon referred to. But correct me if I'm wrong.

 

[DaveT]

Tesla should be able to be a lot less expensive (the cost advantages of EV's will increase), there are already cities in Europe and China where ICE's are prohibited or restricted, some people prefer to deal with green and ethical companies. M3 will be widespread in many areas. I might put two apps on my phone, and if a Tesla M3 isn't available in a convenient time frame check for an uber.

Uber's business model is screw the drivers, and Elon has said that he plans to pay more to the car owners. So uber will need to either start buying a ton of cars, like Bolts or compete with Tesla on price, and Tesla can afford to make almost nothing compared to uber, and still make a substantial profit, because for Tesla it's a feature that helps them sell more cars.

I think it's basically over and uber is already either severely wounded or lost.

I wouldn't count out Uber yet. They've got $9 billion in the bank, and they could easily raise several more billion. With that much money, they've got options to lessen the gap that Tesla has with self-driving. But Uber does need to act fast.

Perhaps it's even possible for us to see two or three healthy ride-sharing networks in the future (ie., Uber, Tesla Network, etc).