TSLA Market Action: 2018 Investor Roundtable

[Cosmacelf]

Tesla's Autopilot engaged during Utah crash

Autopilot was on, but Utah driver was using phone and is likely to be charged.

I’m probably going to get creamed for this, but Tesla really should address the issue where people are using AP on streets that are inappropriate for it. I’ll bet if you polled Tesla owners, you’d find that the majority of them assume that AP would detect and stop for a parked car in front of it. It does not, and this lack of knowledge is getting people hurt.

 

[Menifeer]

I’m probably going to get creamed for this, but Tesla really should address the issue where people are using AP on streets that are inappropriate for it. I’ll bet if you polled Tesla owners, you’d find that the majority of them assume that AP would detect and stop for a parked car in front of it. It does not, and this lack of knowledge is getting people hurt.

What's getting people hurt is the fact that the car doesn't stop when an obstacle is in its path and the driver isn't paying attention.

The issue that should be addressed is the failure of the autopilot system to respond to an obvious obstruction.

 

[Reciprocity]

I honestly never understood how being 4 times safer than the average car in use is supposed to be especially good. I'd assume the current car fleet includes a huge amount of older, cheaper and smaller cars which usually also means they have less safety features. No, that doesn't mean i say or want to imply that Teslas are especially unsafe. But i can't wrap my head around the figures and bring myself to believe it's the safest car around. At least when it comes to deadly accidents this simply doesn't seem to be true. I mean it's completely understandable that more accidents happen, given Teslas are damn quick and fun to drive. The following link contains a list of cars that seem to be relativly safe in that regard, if you scroll down to the part about "Lowest rates of driver deaths".

http://www.iihs.org/iihs/sr/statusreport/article/52/3/1

It shows several cars having 0 deaths per "million registered vehicle years". I believe Tesla has sold something around 300.000 cars over the last years. More in the last 2 or 3 years than before, so altogether registered vehicle years should be in the range of 1 million years. There have not been 0 deaths.

So how does building the safest cars on the road (citing crash test data) translate into being 4 times safer, when it comes to deadly crashes and another dataset shows others cars are safer in that regard? Either a piece of the puzzle is missing or one of the datasets is flawed, but that doesn't really make sense. Probably i'm just stupid again and missing something ...

Well first of all, Tesla doesn't have a gas tank full of explosives. Second, the battery pack can be volatile, but it takes massive damage to compromise the pack. The kinds of accidents you see where Tesla's burn are high speed collisions with solid objects. The battery pack ads rigidity to the frame of the car and basically turns the side of the car at the b Piller into a tank. Which helps to mitigate side collisions. The massive crumple zone in the front is alone enough to save your life when you hit a fire truck at 60, but at 80 your aorta will tear and no one is saving you.

 

[JRP3]

The issue that should be addressed is the failure of the autopilot system to respond to an obvious obstruction.

So what's your theory, that Tesla has that technology available but just decided not to use it?

 

[dc_h]

CNBC Now on Twitter

Tesla to undergo management reorganization as CEO Elon Musk aims to flatten structure - Dow Jones, citing memo

And... Fred sure types fast, doesnt he?

Elon Musk announces ‘a thorough reorganization’ at Tesla, will be ‘flattening the management structure’

CNBC barely gives the reorg a paragraph and jumps into Tesla’s challenges. Even when they have a bull on the focus on negative questions, but but but Chanos zero, fire ...

 

[Mkorpal]

I’m probably going to get creamed for this, but Tesla really should address the issue where people are using AP on streets that are inappropriate for it. I’ll bet if you polled Tesla owners, you’d find that the majority of them assume that AP would detect and stop for a parked car in front of it. It does not, and this lack of knowledge is getting people hurt.

Tesla, in my opinion, hasn't made too many major mistakes. But the whole Autopilot, FSD thing has been a huge misstep. That should never have been something public facing. It's still a long way away, but for some stupid reason, it is on the build screen for a new Model S. They really should have kept the tech private until mature and marketed Autopilot like everyone else does, as driver assist. Now we seem to have too many confused drivers who think the car drives itself.

 

[Skryll]

Tesla, in my opinion, hasn't made too many major mistakes. But the whole Autopilot, FSD thing has been a huge misstep. That should never have been something public facing. It's still a long way away, but for some stupid reason, it is on the build screen for a new Model S. They really should have kept the tech private until mature and marketed Autopilot like everyone else does, as driver assist. Now we seem to have too many confused drivers who think the car drives itself.

Probably true. Hopefully they will get their breakthrough to suddenly fix this, but I am not holding my breath. Autopilot driver assist is super valuable to me as it is today, and I am looking forward to even more improvements, but probably should not have advertised it as aggressively as self driving capable in the very near future...

 

[Steph]

I have 357 shares.
I'm back baby!

 

[uselesslogin]

I have 357 shares.
I'm back baby!

Hi Steph, when you last posted in 2015 I don't know if anyone noticed your post. So this time I will say welcome back our long term very occasional poster!

 

[ls7corvete]

Tesla, in my opinion, hasn't made too many major mistakes. But the whole Autopilot, FSD thing has been a huge misstep. That should never have been something public facing. It's still a long way away, but for some stupid reason, it is on the build screen for a new Model S. They really should have kept the tech private until mature and marketed Autopilot like everyone else does, as driver assist. Now we seem to have too many confused drivers who think the car drives itself.

As Elon pointed out on the cc, all the deaths so far have been with seasoned drivers well aware of the limitations. I doubt a name change will save lives. Though I do think limitations on where it can be used are in order, looks like a few have ruined this for us all.

 

[dmckinstry]

I have 357 shares.
I'm back baby!

Are you the same Steph who was an uber bull prior to when the stock took off in 2013 and then left to get into politics, I am glad to see you back.

 

[bdy0627]

I’m probably going to get creamed for this, but Tesla really should address the issue where people are using AP on streets that are inappropriate for it. I’ll bet if you polled Tesla owners, you’d find that the majority of them assume that AP would detect and stop for a parked car in front of it. It does not, and this lack of knowledge is getting people hurt.

And this is just the tip of the iceberg if AP continues to have this limitation as model 3 hits the streets by the hundreds of thousands over the next year. I love using mine, but it may be wise for Tesla to do something about this until they are able to improve the technology regarding stationary objects. They need to figure this out ASAP. Three people have died over-relying on the technology. That could easily triple over the next 12 months. Yes, it's distracted driving. That's the reality but humans are proving that they think this technology is so good that they can virtually ignore the road and it's costing them dearly since it doesn't even recognize stationary objects.

Tesla, in my opinion, hasn't made too many major mistakes. But the whole Autopilot, FSD thing has been a huge misstep. That should never have been something public facing. It's still a long way away, but for some stupid reason, it is on the build screen for a new Model S. They really should have kept the tech private until mature and marketed Autopilot like everyone else does, as driver assist. Now we seem to have too many confused drivers who think the car drives itself.

Totally agree. We just ordered a Model X and, sure enough, FSD is still listed there as an option to purchase. Those of us who follow Tesla closely know that Musk recently said he thinks they should finally have the technology in place by the end of next year, which likely means quite a bit later. People who don't know the situation well could easily think they are purchasing something that will be made available to them soon. It seems odd to have it available as an option to purchase right now. Things were different when they first made it available to purchase about a year and a half ago I think. It seemed like it was coming soon. We now know that was extremely optimistic.

 

[Skryll]

Though I do think limitations on where it can be used are in order, looks like a few have ruined this for us all.

I disagree. People without autopilot also crash when they text and drive. Are we going to impose limitations on their car functions too ? At some point it comes down to personal responsibility.

 

[bambam4171]

Thing is that the (very) few crashes give AP a bad reputation and I think it could be so far as legislation doing something about the issue. The public opinion can switch very fast and if that happens we'll see politicians loudly voicing their opinions. Waymo got it right: People are lousy monitoring systems, we tend to get overconfident and then bad things happen. I think a system like the Lexus or the Cadillac one that monitors the eyes and stops AP if the eyes wanders off to say a phone or are closed for longer than a blink is needed for the stages 1-3 of autonomy. And something like this could easily be requested by legislation.

I know Elon does not like Eye-tracking but the system seems to be working?

 

[dc_h]

They're actually much more resilient than a lot of longs : I very rarely hear about shorts covering. But longs selling, even here among fans, it happens often.

The bear theme was rampant today and the last week. I think some CNBC people really like tony saragachi. They were upset that anyone would be disrespectful to a guy in the upper 75% of analysts and who’s been calling for apple’s demise since iPhone 4. I assume some people pushing this are hoping it scares away customers.

 

[jhm]

What's getting people hurt is the fact that the car doesn't stop when an obstacle is in its path and the driver isn't paying attention.

The issue that should be addressed is the failure of the autopilot system to respond to an obvious obstruction.

The obvious solution is for drivers to pay attention and safely conduct their car.

 

[SteveG3]

Thing is that the (very) few crashes give AP a bad reputation and I think it could be so far as legislation doing something about the issue. The public opinion can switch very fast and if that happens we'll see politicians loudly voicing their opinions. Waymo got it right: People are lousy monitoring systems, we tend to get overconfident and then bad things happen. I think a system like the Lexus or the Cadillac one that monitors the eyes and stops AP if the eyes wanders off to say a phone or are closed for longer than a blink is needed for the stages 1-3 of autonomy. And something like this could easily be requested by legislation.

I know Elon does not like Eye-tracking but the system seems to be working?

hmm, you got me thinking...

from various discussions here on TMC, and IIRC comments directly from Tesla, my understanding is that the reason stationary objects when traveling at considerable speed may not be picked up is that having the software designed to do so currently would result in false positives for the likes of highway overpasses to the point Tesla feels it would be a net reduction in utility to have AP react to this bucket of events.

if this is the reason such events are not currently detected by AP, how about pairing two changes,

1) use the eye tracking to determine when the driver is paying attention to the field of view ahead,

2) have AP respond with a vigorous alert, but not a driving reaction, for this bucket of events only when the eye tracking determines the driver is not looking forward.

in other words, both eliminate potential accident risk of AP having an agressive driving reaction to false positives AND greatly lower the nuisance factor of false positives by only having this aggressive alert go off when eyes are not on the road.

for those using AP as intended (eyes on the road nearly 100% of the time), this alert would almost never go off, so false positive nuisance issue reduced to nil. for those playing loose with AP instructions, lots of eyes off the road time somewhat deterred by more of these false positive agressive alerts (something of a nuisance), and, potentially fatal accidents considerably reduced in the instances when the alert is for a real stationary object ahead (not a false positive) when the driver needs to be alerted (and is) to have their eyes on the road.

 

[Tslynk67]

I honestly never understood how being 4 times safer than the average car in use is supposed to be especially good. I'd assume the current car fleet includes a huge amount of older, cheaper and smaller cars which usually also means they have less safety features.

I had exactly the same reflection, I wonder how the stats are skewed by deaths in less developed countries with really crummy cars and poor infrastructure.

The injury/fatality rate per accident would be an interesting stat, as would the number of vehicle fires per high-speed collision/model.

The issue that should be addressed is the failure of the autopilot system to respond to an obvious obstruction.

It's not a failure of the AP system as it's not designed to work like that, at best you can argue it's a missing feature and I will admit that I cannot personally see how autonomy can be increased without it, so I assume it will come.

In any case, I don't know what people are thinking when they're approaching a stationary object at 100kmph waiting/hoping for the car to stop for them... The only thing I can think of is that they think it should react the same way as adaptive cruise. Maybe the distinction between the two could be emphasised to the owners?

 

[CyberDutchie]

Brace for a good drop tomorrow—target cut by Morgan Stanley

 

[CyberDutchie]

Cash Consumption vs. Strategic Value: Target Cut to $291
MORGAN STANLEY & CO. LLC
Adam Jonas, CFA

Snippet:

Following 1Q18 results, we are making significant cuts to our near-term and long-term auto margin forecasts and allow for marginally greater equity dilution. We see Tesla as trading near fair value with a balanced risk-reward. We remain EW.

It seems the reduction in price target was mostly based on forecasting lower gross margins (27% vs. 34%), and the capital raise estimate increase (from $2.5B to $3B) which Adam Jonas believes will be in Q3 2018 (despite Elon saying no capital raise in 2018).

Most interesting question in the note though, I think is this:

In our view, the number one issue at the heart of the Tesla investment debate is whether the company is 1) currently operating at a low level of utilization of a very large industrial complex, where incremental revenue can bring large incremental gross margins and improvement of cash consumption, or 2) the next 50-100% of growth in revenue brings forth other calls on cash and does not materially move the company in the direction of a fully self-funding existence.