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

[neroden]

For customer sitting in a Robotaxi it is not important if it is geofenced or not

Uh, yeah, it is. I don't know what you use taxis for, but I use them to go to many places. When I call a taxi to take me from Ithaca to Syracuse, if that's outside the geofence, well, that company is not getting a phone call.

 

[sundaymorning]

Waymo was at 0.64 disengagements per 1000 miles in 2015. Tesla is at best 10 disengagements per 1000 miles according to the investor day. And Waymo is still not done with the fully self driving product even in Phoenix. How could Tesla go from 10 disengagements per thousand miles to 0 in one year or even 2-4 years when Waymo has taken 4 years to go from 0.64 disengagements to 0.09 disengagements.

PS. I know that Waymo is geofenced. For customer sitting in a Robotaxi it is not important if it is geofenced or not.

Geofence data is different than real world data. Take Waymo outside of the geofence and it’s likely to be driving blind.

 

[Artful Dodger]

Oh. OK, yes, "it's a possibility more payments will follow". It's a *likely* possibility. All the insider leaks indicated the FCA contract was for more than $140 million, therefore I believe it is for more than $140 million.

Jefferies forecasts that FCA faces fines of over 2B Euro by 2021.

So what's more likely:

• Tesla pooled with FCA for a one-time payment of 93M Euro (4.5% of fines), or
• Tesla pooled with FCA for ongoing quarterly payments of 93M Euro.

A two yr deal would still be about just 35% of the fine. More likely I think is a 3 yr deal, which would be about 50% of the fine.

Fiat Chrysler pools fleet with Tesla to avoid EU emissions fines | Financial Times

 

[EVNow]

Waymo was at 0.64 disengagements per 1000 miles in 2015. Tesla is at best 10 disengagements per 1000 miles according to the investor day. And Waymo is still not done with the fully self driving product even in Phoenix. How could Tesla go from 10 disengagements per thousand miles to 0 in one year or even 2-4 years when Waymo has taken 4 years to go from 0.64 disengagements to 0.09 disengagements.

Two things - training & probability.

Tesla is betting their varied and large base of data will help them get better much faster.

# of disengagements is also a question of how aggressive you are. You get less aggressive - you get less engagements. That's why Tesla can go to near zero disengagements in slow & steady summon / park modes. They don't have to be otherwise, because there will be a driver.

How long did Tesla take to go from bad lane changes to good ones ?

 

[Alexander Grea]

Uh, yeah, it is. I don't know what you use taxis for, but I use them to go to many places. When I call a taxi to take me from Ithaca to Syracuse, if that's outside the geofence, well, that company is not getting a phone call.

Waymo has a product and needs to monetize it. If they can geofence entire LA or entire NYC and achieve 0 disengagements then they can start providing the service there. If the customer puts another destination in another city then they cannot order it. In those cities I guess 95% of the rides at least are within the same city.

Those top tier cities would be already a huge disruption to Uber. They do not need an universal solution. Covering city by city is completely fine and geofencing the areas. At least for the purpose of ridesharing service like Uber without the driver.

 

[neroden]

OT

Waymo has a product and needs to monetize it. If they can geofence entire LA or entire NYC and achieve 0 disengagements then they can start providing the service there.

Yeah. They can't. They are extremely far from even covering all of Phoenix, which is an easy environment compared to LA or NYC.

 

[scaesare]

Regarding Autonomy...

A. It appears that LIDAR(+HD Map) vs Cameras System is a media hype. Both solution are feasible today, Industry leaders recognize both, and there are working examples for both systems (Waymo/Cadillac vs MobileEye/Tesla). Industry players will probably choose their preferred stacks without consulting CNBC analysts.

While Lidar may be technically capable (up to a point*), is it really a 'feasible" solution if it adds $5,000-$10,000 the cost of manufacturing the car?


* hehe

 

[MarcusMaximus]

I don't know man.
When the competition can put a 10 000+ FSD hardware on a 15 000 ICE car. A tesla at 35 000 still cost more to produce.

What I think most are not thinking about is the willingness of rivals to burn cash in order to kill off the competition. It was quite an eye opening experience to watch Uber and DIDI fight over the past few years.

Hindsight, everyone probably agrees that it is stupid to go up against a gov owned corp with near infinite resources. But before such a duel took place, nobody knew whether attractive tech from Silicon valley will win against gov backed corp.

Probably a good idea that TSLA is such a Chinese Gov darling. One rule of thumb for Elon's twitter about China. If it is not a praise, do not tweet it. Chinese "Face" can make or break your sales.

I think you're forgetting that Tesla has a very cheap source of robotaxis for their eventual service: leases. Every Model 3 a customer leases right now has to be returned to Tesla at the end of the lease, for this express purpose. For a black, no options SR+, a customer will make $21,144 in lease payments before returning the car to Tesla, essentially lowering Tesla's manufacture cost for that robotaxi by that amount. Even if the competitors scrape the bottom of the barrel beater cars to throw their hardware on, they're going to struggle to compete on cost with that.

 

[Cloxxki]

I'm with Elon Musk. If, god forbid, the price went to zero I'd be there with him buying all the way. The ultimate privilege of having a few dollars to invest is being able to take a stand with the one person trying to fix what's wrong.

You do know I hope that only if you buy during a share emission, you actually help the company? If you buy from the market, you just made a more prudent investor's day ;-)
I used to be all in on something I believed in and still do. Lost everything because of it anyway. Retail investors rarely beat inflation.

 

[anthonyj]

@anthonyj
____after the 2:1 stock split?

LOL that sounds like something I would do actually

 

[Alexander Grea]

Um, where are you getting this?

Also, disengagement rates are plenty easy to rig. One, you only have to report safety-related disengagements, so if the car is going to miss a turn because it can't change lanes, or whatnot... you just let it rather than disengaging. And two, if there's anywhere that you know the car struggles with, you just exclude that from your geofence until you've gotten good at it in your simulator of that location.

Based on the reports by Galileo, Ross Gerber and Adam Jones. All of them reported at least 1 disengagement during their ride on the investor day. The ride was about 20 miles each.

 

[MP3Mike]

Based on the reports by Galileo, Ross Gerber and Adam Jones. All of them reported at least 1 disengagement during their ride on the investor day. The ride was about 20 miles each.

No, of the two rides they talked about only one of them had a disengagement.

 

[EVMeister]

Based on the reports by Galileo, Ross Gerber and Adam Jones. All of them reported at least 1 disengagement during their ride on the investor day. The ride was about 20 miles each.

Nice extrapolation and sample size there. I take it you're not a scientist of statistician?

 

[EVNow]

Geofence data is different than real world data. Take Waymo outside of the geofence and it’s likely to be driving blind.

I think the problem is not geo-fenced as such, but, the time it is taking Waymo to expand. If they can add a geo-fenced city every moth, they can do great. But if they take years just in one suburb - that too with ideal weather, traffic and road conditions - when will they cover top 100 markets ?

 

[Alexander Grea]

No, of the two rides they talked about only one of them had a disengagement.

This means for these 2 it is 1 disengagement per 40 miles. Adam Jones reported a disengagement in his report.

I conservatively put the disengagement at 1 per 100 although it looked to be much higher at the event. And the event was in perfect conditions on their home turf.

 

[EVNow]

Just 53k shares to push the SP below 240 this time.

 

[KarenRei]

Based on the reports by Galileo, Ross Gerber and Adam Jones. All of them reported at least 1 disengagement during their ride on the investor day. The ride was about 20 miles each.

Well, then that's a great job comparing apples to chestnut-necklaced hill partridges.

Did any of them report a safety-related disengagement? Because those are the only ones that get reported, e.g. by Waymo.

 

[lklundin]

Agree and it questions their approach on how they optimize (or not) the sensor usage. Seems Lidar is an overkill if only used front facing, let alone self driving usage.

During the Investor Autonomy Day Elon Musk usefully remarked that if you are going to be actively sending out photons from the car, then why do it in the visible spectrum, where there are either already photons to be sensed or where they are useless due to e.g. fog.

Hence Tesla's forward looking radar, rather than lidar.

I wish so much I could have been present at that talk, there are so many questions that I would have liked to ask.

PS. While we are bashing Audi's poor technology choices, I will point out that I updated the Wikipedia article on the e-tron. Now that the vehicle can actually be purchased (in small numbers) in Germany, they have had to put out the actual specs on the car, which are abysmal: 265 kW delivered in total by the two motors, 0 - 100 km/h in 6.6 seconds, the 417 km WLTP range is actually the upper limit, with the lower range (in combined driving) is only 358 km. As for the acceleration, my 2008 Audi A8 TDI had a comparable 240 kW engine but a 0 - 100 km/h in 5.9 seconds. Game, set & match...

 

[Alexander Grea]

Nice extrapolation and sample size there. And even then you're got it wrong. I take it you're not a scientist of statistician?

This is the data we have as TSLA is hiding the actual data they have based on their testing.

Due to the small sample size I put the disengagement rate at a much lower rate than the investor day rides show as a benefit of doubt for TSLA.

Market action:
Super important to climb back over 240 and finish the day near the highs to set the tone for the rest of the week.

 

[Nocturnal]

I think the problem is not geo-fenced as such, but, the time it is taking Waymo to expand. If they can add a geo-fenced city every moth, they can do great. But if they take years just in one suburb - that too with ideal weather, traffic and road conditions - when will they cover top 100 markets ?

And how do they actually deploy to those markets? Will consumers be satisified with imaginary boundaries in their towns that they can't cross? "ok, I don't need a car because ride sharing is cheap, but I can't go anywhere past 50th Ave."