2017 Investor Roundtable:General Discussion

[neroden]

In my experience as companies get bigger they are expected to have financial processes in place that allow them to close their books in less than 5 or 6 weeks. I am assuming that the shortening of the period in Q3 to less than 4 weeks was something that Jason Wheeler put in place and that it will be repeated in future quarters. If so we can expect earnings in the last week of January.

BUT on the other hand, integration of SolarCity may delay the earnings report. I've found that the immediate post-merger earnings reports are often delayed because of the complexity of integrating accounting systems and ironing out the problems with that. So we might not see earnings until, say, the last week of Feburary.

(That said I think we'll probably see them within 4 weeks for later quarters; this would be a one-off thing)

 

[Krugerrand]

BTW No, no one needs to buy a 6 package pack of toilet paper. But it is cheaper and does not spoil so I do buy toilet paper at Costco. Charmin toilet paper pack is $4.40 off this month at Costco USA.

Somebody install a SuperCharger at RobStark's, he's got a sugar ton of the softest toilet paper this side of the moon.

 

[neroden]

Two other things I didn't mention that are quite relevant:

1. When that Comma.AI dude wanted to work with Tesla on AI, Musk specifically said that what he is doing isn't that interesting for Tesla because having something working in a straightforward scenario isn't that difficult, it's handling the corner cases that will make it a usable product and that requires a lot of grunt work not some one genius. You can't possibly claim they arrogantly think they'll just show some videos to some neural network and call it good. They stated that explicitly a year ago.

2. On the narrow bridge scenario. There's a much more frequently occurring and a more difficult scenario: a narrow one lane city street with cars parked on both sides where you can go both directions. Not only AI will have to figure out how to arbitrage with an oncoming car, but it must be able to back up and let someone go to resolve a deadlock (possibly with another AI car). This they'll have to solve soon, if they haven't already. Your one way bridge scenario is actually a somewhat easier instance of the same generic problem of navigating a narrow passage where oncoming traffic might be present.

Thanks. Good example. I'm quite sure they haven't solved it already, and I'm quite sure it'll take them years to get it right. They'd have mentioned it as something they were working on if they were anywhere close to solving it; they do tend to brag a little when it comes to Autopilot.

To reiterate, I do believe they'll get it done eventually, and relatively soon in absolute terms, and critically, *ahead of the competition*. But this is simply not a done-by-2020 project, not when they haven't really started on most of the corner cases (and they clearly haven't, since they haven't gotten some of the extremely common cases done yet).

This might actually prove to be a lasting advantage for Tesla. It's looking like there will be real competition in the electric car market by 2019 or 2020. But Tesla has enough of a lead on everyone else that they might be solving the obscure corner cases in the 2022-2025 time range while their "competition" is still working on the more common cases....

 

[AlMc]

Somebody install a SuperCharger at RobStark's, he's got a sugar ton of the softest toilet paper this side of the moon.

Works for me as long as he does not have a ton of sugar.

J/k Rob

 

[Krugerrand]

Works for me as long as he does not have a ton of sugar.

Does Costco sell that? From organic produce to tires...they probably do.

 

[Krugerrand]

This is entirely true, if you amend it to "Realistically, an untrained computer program isn't going to behave any worse than average drivers do in these situations".

The thing is that public perception and acceptance of automated stuff is a bit weird. People expect it to be better than very good drivers. Musk's even commented on this. Public perception is the main reason we don't have automated trains everywhere; the tech was working in the 1970s and perfected in the 1990s. Psychology is weird.

Hitting the mark where they're better than 90% of drivers is going to take a while.

That's a fair point.

 

[neroden]

Thoughts on your Price target for selling?

I've been holding Tesla for a long time. Seen it hit $260 2 previous times now only to see it drop to low $180s and every time I bang my head against the wall asking myself why didn't I sell back at $260 so I could rebuy at the $180 mark. Buying stock is easy... knowing when to sell is the hard part.

I bought 1000 shares of Nvidia at $26 and sold at $28 thinking I was awesome... It's now $103...

So question for all you investors... what's your game plan?

Sell at $1000/share ! Actually, I'm being a bit facetious, because I may end up reevaluating Tesla when it hits $1000 and decide that it's actually worth more than that. But if it hits $1000 this year, I'm out.

 

[Krugerrand]

So question for all you investors... what's your game plan?

End game.

 

[hockeythug]

'Bye then.

You wish. Not going anywhere.

 

[larmor]

That's a fair point.

Add to that equation: drivers' ability is stable to declining with the usage of phones. Super unscientific observation and broad sweeping generalization (that means I should get paid for this observation) is that people are on their phone at nearly all lights. Make your own judgement.

The progress of machine computing power, coherence with other machines and machine learning is happening fast and will tilt the balance of driving ability to machines sooner than we think. In the linked ted talk by Sam Harris, there is a picture of a human stepping on an ant. Will we be the ant next?

Sam Harris: Can we build AI without losing control over it? | TED Talk | TED.com

 

[Krugerrand]

No my point is I'm tired of all the personal attacks on Trump. Does that really make people feel better mocking him?

I feel like I should lie.

 

[MitchJi]

Thinking about it, the aging is probably the reason the Gigafactory is so huge.

No it is not.

 

[neroden]

From what I can tell they are only talking about safety. Accidents per mile, that's it. "Better" driving is a highly contested topic and I doubt there will ever be an agreement on that. Drive from A to B in reasonable time, don't have tickets or accidents. Simple (doesn't mean easy).

Automated trains have had close to zero accidents per mile. There have been a large number of suicides, which are very hard to prevent -- there have been a few people who accidentally fell onto tracks in front of an oncoming train. Over two decades of experience.

By contrast, manually driven trains still have accidents where the trains actually go where they're not supposed to. There was one just last week on the LIRR.

And yet it's still hard to get public acceptance of fully automated trains. Frustrating, isn't it?

Do you really expect fully automated cars will be accepted by the public just because they're safer than manually driven cars? If so, why? Didn't work for trains.

 

[larmor]

Add to that equation: drivers' ability is stable to declining with the usage of phones. Super unscientific observation and broad sweeping generalization (that means I should get paid for this observation) is that people are on their phone at nearly all lights. Make your own judgement.

The progress of machine computing power, coherence with other machines and machine learning is happening fast and will tilt the balance of driving ability to machines sooner than we think. In the linked ted talk by Sam Harris, there is a picture of a human stepping on an ant. Will we be the ant next?

Sam Harris: Can we build AI without losing control over it? | TED Talk | TED.com

One more thought regarding AI for the weekend: what if an AI engine starts trading-- really trading with high freq, options, etc. The AI engine could soon have all the money in the world, and humans would have to contract and subcontract out to machines in order to get any work or sustenance...

 

[JGInnovation]

So question for all you investors... what's your game plan?

Sell when Elon is on the rocket to Mars!

 

[Waiting4M3]

I've backed up the truck on TSLA and holding.

One thing that may make me think twice is if Elon starts using the word "imminent" again. If it's really imminent then he needs to just give a date, or keep his mouth shut until it's done.

 

[EinSV]

Thoughts on your Price target for selling?

I've been holding Tesla for a long time. Seen it hit $260 2 previous times now only to see it drop to low $180s and every time I bang my head against the wall asking myself why didn't I sell back at $260 so I could rebuy at the $180 mark. Buying stock is easy... knowing when to sell is the hard part.

I bought 1000 shares of Nvidia at $26 and sold at $28 thinking I was awesome... It's now $103...

So question for all you investors... what's your game plan?

I think all the pieces are now in place for Tesla to become a $700B to $1T company if it executes well. Unless there is a dramatic change I plan to hold tight for a long time to let it play out.

 

[3Victoria]

Do you really expect fully automated cars will be accepted by the public just because they're safer than manually driven cars? If so, why? Didn't work for trains.

Possibly because people are much more familiar with cars, and how they operate.

 

[neroden]

No it is not.

OK, Mitch, seriously, I'm going to challenge you on this. Evidence?

Musk is speeding up every segment of the production process to be as fast as possible -- which means that the production lines are "spitting out cells at the speed of bullets". That's a very space-efficient thing to do. All the production lines are set up to be very space efficient and very fast.

And then they have to have all the cells sit still for *how* long? You can't reduce the space usage there. Not one bit. The higher the production volume, the more aging space you need. Even if you make the machines work faster, you need yet more aging space....

 

[neroden]

One more thought regarding AI for the weekend: what if an AI engine starts trading-- really trading with high freq, options, etc. The AI engine could soon have all the money in the world, and humans would have to contract and subcontract out to machines in order to get any work or sustenance...

Some would point out that this is already happening with algobot hedge funds; the best are very consistently better at short-term trading than any human. Obviously due to the legal system the owners of the "AI" are the ones who have the money.

Long term trading is another matter.