Live data from Hacker News

Tesla Q2 2019 Letter

ir.tesla.com

101–110 of 189 posts

Re: Tesla Q2 2019 Letter

#101

Tesla's stock has massively under-performed over the past 5 years (will be up 8% over that period if the after hours price action holds). The opportunity cost of owning it has been huge while the rest of the market ripped up.

As a non owner of a vehicle or stock... I don’t see a reason why the average interested watcher should care. There’s significant speculation and huge financial forces putting unusually large efforts to push this stock up or down.

Personally I don’t think it matters or changes the likelihood that they’ll eventually succeed or not succeed.

Re: Tesla Q2 2019 Letter

#103
post #88

Earlier quoted context omitted.

Well, maybe. Since most of this data will be very repetitive, I don't think we can know for sure how much it helps?

You (and neural nets) learn the most in the smaller number of cases when a boring repetition is violated. The error function was activated and that's worth feeding back into learning. Ask any kid, they probably couldn't say what happened at school all week. But if they do remember something, it would have been something outside of their routine.

Yes, exactly! We don't know the amount or quality of interesting data from the total amount of data.

Re: Tesla Q2 2019 Letter

#104

Sidenote: I can't stand the ipad in the middle of Tesla cars (and other carmakers who choose this path). I got a BMW because of the sheer usefulness of physical buttons while driving / keeping eyes on the road.

Guess people like fiddling through an ipad while driving a 2 ton machine at highway speeds? ¯\_(ツ)_/¯. Looks like Tesla drivers are becoming the "Toyota Camry/Prius" drivers of the past.

Is there some historical problem with camry/prius drivers?

Re: Tesla Q2 2019 Letter

#105
post #68

Being able to test your code out in the real world and dark launch a feature gives Tesla a ridiculous advantage. No other company has the infrastructure to be able to do that. Plus they can even tell if the decision the neural net made was wrong because I assume even Tesla knows you got into an accident. And this over thousands of cars. The speed at which they are collecting data is unprecedented.

I'm so tired of this unapologetic Tesla hand holding.

> No other company has the infrastructure to be able to do that.

Because they don't want it. They don't want AP regressions we've already seen happen. They don't want engineers testing code out on real people's lives.

You really think other companies just don't know how to send OTA updates to an ECU or other car components?

I could literally put that together with scraps in my garage, my car's ECU can be flashed over OBD, it'd be trivial to write firmware for a microcontroller to repeat a set of OBD commands downloaded from a webserver.

Obviously an auto manufacturer would do more than that, but there is literally nothing but choice stopping other manufacturers from doing this.

Why is everything Tesla does always suddenly some sort of competitive advantage. Next people will be saying the fact they have cars that spell out "S3XY" is a competitive advantage.

Re: Tesla Q2 2019 Letter

#106

anyone care to talk about this part "This feature is currently operating in “shadow mode” in the fleet, which compares our software algorithm to real-world driver behavior across tens of millions of instances around the world." that seems very juicy? edit: it seems people put a disclaimer on their ownership of tesla stock on hn about tesla threads. i dont own any.

They run autopilot in all the cars whether it is turned on or not.

If it is not turned on, they still send the inputs to autopilot, without controlling the car. I think they can also compare what the driver does against what the autopilot would have done.

They can also ask their "fleet" to match specific conditions, and collect data if they are met.

The autonomy day video showed some interesting images, for example "animals" the fleet had seen.

I think this is the video https://youtu.be/tbgtGQIygZQ

Re: Tesla Q2 2019 Letter

#107
Looking healthy to move along for Chinese factory, Model Y and pickup truck later this year.

I like that they don't stop moving even if it scares a few investors with heavy capex spending on their next products.

Re: Tesla Q2 2019 Letter

#108
"GAAP net loss of $408M" - the number that matters.

Tesla isn't doing that badly. But now that they have some production volume, they have to be evaluated as a car company, not a startup. Can they make lots of cars at a profit? They got the production quantity up by throwing people, money, and a big tent at the Fremont plant. That ran up the cost per car. Although it's better than their previous state of low production with a full staff. Can they get to a smooth running production plant with labor hours per car comparable to Detroit? (About 30-50 labor hours per car is typical in the industry.)

"Model 3 average selling price was stable at approximately $50,000."

That's nice, but the Model 3 was supposed to be $35,000. Tesla has caught up with their backlog, and you can now buy a Tesla Model 3 off the lot, just like regular car dealers.[1] That's good; they're now like a normal car company. But it also means they've saturated the market at their high price point, which is about where BMW is.

They're also at the point where all their models need a refresh or a replacement. Tesla is slow at new models compared to the competition.

[1] https://www.extremetech.com/extreme/289911-what-shortage-tes...

Re: Tesla Q2 2019 Letter

#109
post #47

Earlier quoted context omitted.

This report is bad news for Tesla, but I think it disproves the "demand problem" narrative. Tesla delivered more cars than they could produce this quarter. Both delivery and production numbers were at record highs. The average sale price of the Model 3 remained stable. Meanwhile they are nearing production of the Model Y which is likely depressing demand of both the Model 3 and Model X. Once the Model Y is available,…

Agree with most of the above except for the "Model Y likely depressing demand"- is there any indication that the Model Y is in demand at all?

People with families strongly prefer crossover models and they've mostly pushed sedans out of the market. It would stand that an electric crossover would outsell a 4 door sedan.

Re: Tesla Q2 2019 Letter

#110
post #46

Earlier quoted context omitted.

That's how I took it. They can then compare to what the human driver actually did, and find false positives (driver continued when the AI would've stopped) and false negatives (driver stopped when the AI would've continued).

ugh! the amount of supervised training data that would generate is staggering.

Competitors buying hundreds of cars and continuously driving over mannequins.
Post reply on HN