Something's Wrong at Tesla

[desy]

http://www.thetruthaboutcars.com/2018/05/time-revive-ttacs-tesla-deathwatch-tesla-run-cash-year/

Reporters Dana Hull and Hannah Recht did a deep dive into Teslas finances, both where their money has come from and where it’s been going. They came up with some interesting data. The company is going through cash at the rate of about $6,500 a minute, a bit more than $9 million a day. Free cash flow has been in the red for over a year. That’s how much money a firm generates after subtracting capital expenses.

The company’s cash situation is closely related to employment levels. In just three years, from 2014 until last year, Tesla went from about 13,000 employees to more than three times that number and they are still looking to hire thousands more as they try to ramp up production on their Model 3. The Model 3 is unarguably vital to the company’s success.

As the payroll increased, and with sales relatively flat before the Model 3 finally started shipping, revenue per worker dropped. Musk has often spoken of showing Detroit how to do things, but GM and Ford each take in more than twice as much revenue per employee as Tesla does. In the 4th quarter of 2017, Tesla paid out as much in interest on its debt as General Motors did, with Tesla earning one tenth of the revenue of GM. The revenue per employee problem is exacerbated by what Musk himself called a “Russian nesting doll” of contractors and subcontractors at the California assembly and Nevada battery plants.

At the start of 2018, Tesla had $9.4 billion in debt and $3.4 billion in cash. Moody’s says that $1.2 billion of that debt has to be repaid next year and over and above that, to stay operating it will need another $2 billion this year.

Tesla claims that it won’t require additional financing this year, save for existing credit lines, but that’s contingent on the positive cash flow that would result in the company actually hitting 5,000 units per week production level on the Model 3, something it hasn’t yet proven it can do.

Jeff Osborne of Cowen & Co. doesn’t believe Tesla’s claims, since Musk has made similar statements in the past and then gone out to raise more cash. Osborne thinks Tesla is going to try to raise $3 billion with a stock sale late this year, and another $2 billion late next year, to cover cash burn and debt retirement while still keeping more than $1 billion in cash reserves. I don’t believe that figure would include the $854 million in (mostly Model 3) deposits, but it likely includes the substantial interest earned on what amounts to interest free loans to the company.

 

[Kaido]

I imagine that there are tons. A lot of interstates that are primarily local or "business routes" routinely go from rural to "small town" over and over. IME, these are usually 55mph zones, but some are 60mph, but the speed often changes because the routes frequently go into commercial and residential areas and yeah, there are often traffic lights because you have local roads trying to cross these ~4 lane divided highways. But in those cases, you will almost always have speed limit warnings well in advance, bringing you down to 45mph or less, as well as warnings about an upcoming signal.

I can kinda see where an autopilot might fail on such a road, (it's not like it knows the posted speed, does it, especially if it drops quickly?), if coming around a long, blind corner and there is a signal at the other of the turn, autopilot probably wouldn't have time to see it. This of course means that the human driver had been ignoring all of the signage up to that point and essentially let this happen, so I can't imagine not faulting the driver--I just think there are situations where this current generation of automated driving could plow a car into the back of a stopped firetruck at 60mph, if left on its own.

Huh, interesting. Most of our stoplight roads here max out at like 40 MPH. I know of a turnpike that is basically a 2-lane road on each side (4 lines with a barrier in-between) that is 55 MPH. Maybe it's just because it's really hilly where I live.

Although when I was out west I did hit some 80 MPH highways, which was pretty awesome!

 

[breag]

"Tesla's Autopilot system uses radar, cameras with 360-degree visibility and sensors to detect nearby cars and objects. It's built so cars can automatically change lanes, steer, park and brake to help avoid collisions."

https://www.msn.com/en-us/news/tech...-crash/ar-AAxh9VQ?li=BBnbfcL&ocid=mailsignout

How in the world could something like that not detect a fire truck?

Also, loading up roadways with RF emtting self driving cars - just what we all need.

https://www.cancer.org/cancer/cancer-causes/radiation-exposure/radiofrequency-radiation.html

I can take some easy guesses at parts of the why.
First - the firetruck was red
Red may be the color we're all used to but a lot of departments have switched to yellow or yellow/green for a reason.
Red vehicles are harder to see at night and identify as a big truck even with the reflective tape.
If the firetruck was stopped at a light I suppose it's a possibility that the Tesla couldn't *see the light.

I'm honestly shocked that a 60mph crash with a big heavy stationary object only resulted in a broken ankle.

 

[Kaido]

Except, no, they really aren't. I like Tesla's in general, if I had the cash, I'd probably get a Model S. But you're drastically over-hyping their upsides and dismissing their downsides.

Superchargers really only matter if you take road trips. Personally only one of the Tesla owners I know does and it's for business. The rest just drive theirs around town, and charge it at night. As a general marketing point, sure it's drastically faster charging than you get with competitors but 30 minutes is still a long time compared to the 2 minute it takes to pump gas. IF you're taking road trips with your EV (and that's a big IF), the value of that goes up. But the price gap (regardless how you try to spin it) between Tesla's and most the other EV's is VERY sizable.

Admittedly Tesla has nobody to blame but themselves, but every time somebody calls a Tesla self driving, god kills a kitten. Is it impressive? Yes. But the software isn't there yet to call it truly self driving even if the hardware is. There's a reason the car tells you to keep your hands on the wheel every time you use it.

"4x safer" is complete and total marketing BS. From Elon's own mouth, that's based on the NHTSA's stated one fatality every 86M miles vs Tesla's one in 320M. He's taking his one recent model and comparing it to the national average which includes a mountain of old cars. When it's regurgitated by his fan club, I'm willing to dismiss it as ignorance. Out of his mouth, it's flat out deliberately deceptive marketing. I haven't seen 2018 data but in 2017 IIHS published a report of the safest vehicles on the road, 11 of which had ZERO fatalities in a 4 year period. Spoiler alert, Tesla wasn't one of those 11. The size of the Model S and Model X are certainly a large factor in it's safety as well. There's also a variety of other numbers Tesla "massages" (or ignore) when doing their song and dance. Tesla doesn't report US auto sales. One might ask why not if things are as rosy as they make them out to be. But based on registration, Tesla's US sales last year looks to be about 55k. Which puts them behind Porsche, which means they're not selling like hot cakes.

You could counter that last point by saying it's not an interest issue, it's a production issue. But then you've just reached the major concern most finance people have. Tesla STILL isn't reaching the production numbers they said they would. For a company that's bleeding money at a shocking rate, that's a big problem.

Excellent points.

I don't remember the exact number, but I timed filling up my Jeep from empty a couple times, and it was around 2 minutes and 30 seconds. Even with the limited 250-mile range on my Renegade, refueling is ridiculously fast & gas stations are everywhere, so there is zero range anxiety associated with driving an ICE vehicle - no thinking or planning is required for anything outside of a normal daily commute & errands. tbh, I have a hard time seeing electric cars really taking off until they get longer ranges, due to the long charging times (or get faster charging...no one wants to hang around at a refueling station for more than ten minutes). Even 30 minutes at a Supercharger is an extremely long time compared to refilling a gas or diesel vehicle.

Well, that and the economics need to get better, like you said, to the point where it's cheaper to purchase & drive an electric car over a gas car. $35k is a reasonably attractive price point for a brand-new Model 3, but that's for the base 220-mile, RWD, non-performance, barebones version. Adding AWD, 300+ mile range, enhanced Autopilot, etc. reportedly drives the price up to nearly double that. Although $35k is a price point out of reach for a lot of people as well, so even billed as being "the people's" electric car, it's really only for the people who can afford a nearly $600/mo car payment for the base model (or $1,000/mo for the fully-loaded $60k Model 3 with zero down & zero interest over 60 months). And what you get, bottom-line, is a car with a limited range, a relatively long recharge time, and one that you still have to babysit while it "self-drives". But, at least they're working on it, so hopefully full self-driving pans out, the range increases, the charging time decreases, and the price comes down over time.

 

[Kaido]

Cause of firetruck crash determined:

http://www.thedrive.com/news/20912/...crash-into-fire-truck-cause-determined-report

The gist of it:

According to the release, 82 seconds prior to the crash, she again activated Autopilot and adaptive cruise control, and removed her hands from the wheel within two seconds of activation. For the next 80 seconds, the Tesla continued at her preselected speed of 60 mph, her hands still off the wheel. Against eyewitness reports, the car's computer did reportedly register brake pedal application, as well as her hands' return to the wheel, but the reaction was too little, too late.

In more detail:

https://www.teslarati.com/tesla-autopilot-crash-details-nhtsa-investigation/

TL;DR: Autopilot crashed her car straight into a giant firetruck at 60 MPH.

So there are 3 high-profile Autopilot crashes:

1. The semi-truck in Florida (fatality involved)
2. The barrier in California (fatality involved)
3. The firetruck in Utah (broken ankle)

In each case, the driver was not using Autopilot as intended. They were distracted and were not watching the road and were not babysitting the system or their situation as required. The software is in beta and has reminders about proper usage.

Again, on the flip side, I still don't like Tesla's "wink wink" approach to Autopilot usage -
'yeah yeah it's in beta, but it's a self-driving car, man!'. I don't like the name Autopilot in terms of its currently available feature set. And I really don't like how they brush off accidents. Their response to the California crash was pretty crummy - they tried every angle they could to blame the driver, even going so far as to cite a quote by the victim's family that he had had prior issues on that stretch of road, so why was he not paying attention? And Elon Musk's response to the latest crash where Autopilot drove the user straight into the back of a giant firetruck at 60 miles per hour was this:

It’s super messed up that a Tesla crash resulting in a broken ankle is front page news and the ~40,000 people who died in US auto accidents alone in past year get almost no coverage

Is he wrong? No. Autopilot is supposedly statistically safer than all other methods of driving at the present time. But the software literally drove a car directly into another vehicle. And Tesla does whatever they can to avoid saying that directly! Yes, technically, per the letter of the law, it's labeled as beta, has some warning chimes & lights, and you're supposed to babysit it while it's self-driving. In practice, people don't buy a Tesla with Autopilot so they can babysit it, they want the self-driving features, obviously!

I don't think it's realistic to expect to NOT fall into the trap of distracted driving when you have a (mostly) reliable self-driving car on the highway. imo Tesla could exercise more responsibility in how they handle alerts & in how they respond to accidents where Autopilot caused death or injury. We understand the Autopilot is a safer way to drive, but that doesn't mean it's perfect, and I certainly don't like how they brush it under the rug & point fingers instead of accepting any level of blame for their system running people's cars into stuff & hurting or killing them.

And to be clear, I like what Tesla is doing, both in terms of electric car progress & in improving automotive safety, both through the design of the car & the Autopilot system, and I don't think they should stop, because we'll never get to where we need to be if we try to be perfect first - it's going to be incremental improvements along the way, and negative things like these Autopilot crashes will continue to happen because the world is complex & it's a risk. For now, at least, I'd just like them to publicly shoulder more of the responsibility for Autopilot crashes & also improve the alerts & alarms when the system is not being used properly. I just don't like how they try to hide Autopilot as the root cause of the crashes.

 

[Kaido]

There's some good discussion on the Utah crash in this thread:

https://www.reddit.com/r/teslamotors/comments/8jz5tl/tesla_reports_from_utah_firetruck_crash/

One user mentions that the steering wheel torque sensor does a poor job of detecting attentiveness:

Ok so here's what bugs the shit out of me: The overwhelming majority of the time that my car nags me to put my hand on the wheel, my hand's already on the wheel, I'm just not applying enough torque for the car to "see" that because my left elbow is resting on the window sill and my hand is resting on the 9oclock position. So I give it a quick wiggle and then we're friends again. This means the car logged a "no hands" event

So the reasonable belief here is that Tesla is using possibly false negatives to publicly shame people who very well might have been following the...guidelines.

Another user makes an excellent point about who the blame ultimate lies with:

I think the more conclusive evidence is the accident itself. Attentive drivers do not plow into a fire truck stopped at a red light in good driving conditions.

I 100% agree with this. Driving literally puts your life on the line because you are piloting a multi-thousand-pound vehicle at high speeds along with other people doing the same thing. Even though Autopilot was technically the root cause of the accident (because it didn't "see" the firetruck, for whatever reason), there's no excuse for what happened, in terms of the driver not paying attention.

And again, Tesla needs to recognize that their current system creates a distraction trap for people & thus needs to improve their system for babysitting the driver's attentiveness. This isn't an uncommon issue. Pre-meme days, Tide had a lot of issues with little kids eating their laundry pods because they look like candy & then getting really sick. On one hand, you could argue that the parents should do a better job babysitting their kids & putting stuff out of reach, but life happens - but kids are tricky & get into stuff no matter what. The same thing with the AR Pokemon game - just google "pokemon go accidents". The developers weren't careful to avoid public roads, especially when young, distracted children were playing the game who sometimes fail to look both ways when crossing the street, and multiple people have died & scores more have been injured.

So while Tesla may place the blame squarely on the driver, and I do believe the bottom line is that the driver is ultimately responsible for the accident if they are not paying attention, they've created a trap for people to get distracted in, in addition to Autopilot literally driving people straight into other cars & into barriers.

 

[Midwayman]

I think Telsa is just proving the idea that advanced auto pilot features are a bad idea until they get more reliable. The fail on the firetruck should have been the car stopping and pulling off to the median if the driver wasn't paying attention. Sure she should have been paying attention. She was using it wrong. However the question of why it didn't see the firetruck is a very good one. Like the uber accident this seems like it should have been very easy to avoid. Something is wrong with their software.

 

[local]

And this is why I say we will never have fully autonomous vehicles outside of restricted areas. No company is going to be willing to take the blame instead of pointing the finger at the distracted driver.

 

[SearchMaster]

I think Telsa is just proving the idea that advanced auto pilot features are a bad idea until they get more reliable. The fail on the firetruck should have been the car stopping and pulling off to the median if the driver wasn't paying attention. Sure she should have been paying attention. She was using it wrong. However the question of why it didn't see the firetruck is a very good one. Like the uber accident this seems like it should have been very easy to avoid. Something is wrong with their software.

I don't know that they'll ever be 100% reliable until all cars are "autopilot" and communicating with each other. Tesla's current system is pretty remarkable but there will always be fringe cases like the ones mentioned above that cause the system to fail - just as there are fringe cases with humans driving that cause the human system to fail. And as long as humans can override the automation, then the automation will be more prone to failure.

I can't personally imagine that in my lifetime I would ever feel confident enough in an automated driving environment to allow it to completely take over - at most for a few seconds on a relatively empty road - but I guess others have let technology take over so much of the mundane parts of their life that they allow this to do the same.

 

[bradly1101]

They have a fleet of 55 cars just in California:

https://techcrunch.com/2018/05/14/apples-self-driving-car-fleet-grows-to-55-in-california/

Hmmm. Nvidia.

 

[bradly1101]

I think it's going to win out in the long haul:

https://www.cnbc.com/2018/05/14/ron-baron-were-going-to-make-20-times-our-money-in-tesla.html

Musk is right about the short sellers imo:

https://www.cnbc.com/2018/02/07/ark-chief-catherine-wood-sees-tesla-stock-going-to-4000.html

Their Supercharger network is getting ridiculous:

https://electrek.co/2018/05/12/tesla-supercharger-2018-growth/

Electric cars are cool. However, most of the them only have an 80-mile range. Newer offerings like the Bolt & upcoming Kona have better ranges, but still have slow charging. Tesla has a nationwide Supercharger network now...30 minutes on a Supercharger vs. 3 hours for a regular EV is just no contest. Plus, nobody else is offering anything remotely close to Autopilot, including the built-in future capabilities. Sure, Cadillac has Supercruise & Nissan has ProPilot, but long-term, they're not going to match what Tesla cars can do in terms of city-street self-driving. It's like no one else is even trying...both with making full self-driving cars & with rolling out a high-speed nation-wide charging network.

Eventually the Model 3 production will ramp up, the back-orders will get filled, and they'll be ready to deliver to new customers in pretty short order. And then the Model Y will come out, which will probably be an even bigger seller than the Model 3. And then the Roadster 2.0 will come out, and I'd imagine they'll find a way to setup their other vehicles (S & X, at least) with the 620-mile battery, which will further distance Tesla from their competitors. Although I am a Tesla fanboy, these are just the facts: No one else has a Gigafactory. No one else has built-in future FSD capabilities. No one else has a Supercharger network.

I think Tesla is going to kill it in the long game.

It's a bitch though, faster charging = lower battery lifespan.

 

[Midwayman]

And this is why I say we will never have fully autonomous vehicles outside of restricted areas. No company is going to be willing to take the blame instead of pointing the finger at the distracted driver.

Geographic restriction will probably be a thing for a long time anyways. I'd be happy for a highway only mode anyways. As for as insurance, I think once we get past the 'advanced driver assist' phase and into level 4 cars you're going to see insurance bundled in your purchase price or lease. There won't be much of a finger to point if humans aren't assumed at the control. Another reason most companies are avoiding what telsa is doing.

I don't know that they'll ever be 100% reliable until all cars are "autopilot" and communicating with each other. Tesla's current system is pretty remarkable but there will always be fringe cases like the ones mentioned above that cause the system to fail - just as there are fringe cases with humans driving that cause the human system to fail. And as long as humans can override the automation, then the automation will be more prone to failure.

I can't personally imagine that in my lifetime I would ever feel confident enough in an automated driving environment to allow it to completely take over - at most for a few seconds on a relatively empty road - but I guess others have let technology take over so much of the mundane parts of their life that they allow this to do the same.

Well I think we're 10-20 years away from 100%. Meaning any road, any weather condition. Still a truck stopped at a light is hardly a fringe case. I think what we need to be looking at is when it is undeniably safer to take humans off the road. We're pretty terrible at a lot of things driving related. Limited range of vision, distracted, slow reflexes, only poorly understand vehicle dynamics. Aside from the interpretation of data the machines are better in every way. I'm sort of the other way. I fully expect humans to be limited from driving on roads during certain hours or limited geographic regions. Rush hour and downtown specifically. Once you prove the cars safer after a couple years of release there is going to be a huge push for them.

 

[Kaido]

I don't know that they'll ever be 100% reliable until all cars are "autopilot" and communicating with each other. Tesla's current system is pretty remarkable but there will always be fringe cases like the ones mentioned above that cause the system to fail - just as there are fringe cases with humans driving that cause the human system to fail. And as long as humans can override the automation, then the automation will be more prone to failure.

I can't personally imagine that in my lifetime I would ever feel confident enough in an automated driving environment to allow it to completely take over - at most for a few seconds on a relatively empty road - but I guess others have let technology take over so much of the mundane parts of their life that they allow this to do the same.

Volvo is working on the car-to-car communications feature right now:

https://techcrunch.com/2018/05/07/volvo-cars-and-trucks-can-now-share-real-time-traffic-information/

I've taken Autopilot out for a test-drive, it's actually pretty dang good. For the first minute, you're just like AHHHHHH! And then you realize it can react faster than you can & probably drive, generally-speaking, better than you can, because it's constantly scanning the road & making adjustments & whatnot.

Personally I plan on waiting until the technology matures more (and the price comes down, the Model X I test-drove cost more than my last house haha) before I jump. For now, without FSD enabled, it's basically just a glorified adaptive cruise control & automatic lane-keep assistant system, which you still have to babysit, especially because if you get hurt or die, Tesla will blame you anyway lol.

 

[Kaido]

Well I think we're 10-20 years away from 100%. Meaning any road, any weather condition. Still a truck stopped at a light is hardly a fringe case. I think what we need to be looking at is when it is undeniably safer to take humans off the road. We're pretty terrible at a lot of things driving related. Limited range of vision, distracted, slow reflexes, only poorly understand vehicle dynamics. Aside from the interpretation of data the machines are better in every way. I'm sort of the other way. I fully expect humans to be limited from driving on roads during certain hours or limited geographic regions. Rush hour and downtown specifically. Once you prove the cars safer after a couple years of release there is going to be a huge push for them.

Even so, I think Autopilot is probably a better driver than most people on the road. I drive 50 to 100 miles daily and almost get in an accident every single day from other people acting crazy.

 

[XavierMace]

I think Tesla calling their system "auto-pilot" is a reasonably accurate name. I think the problem is people's assumptions about auto-pilot in planes is wrong and that carries over to Tesla. Auto pilot takes care of the long boring cruise. The pilots still generally take care of the taxi-ing around the airport, the takeoff, and the landing. Other events can trigger the pilot taking back over in flight as well. There's a clearly defined difference between a driver assist (which is what autopilot is) and self driving (like Waymo).

Waymo just hit the 5 million autonomous miles mark and I believe a good portion of that is here in AZ. They've started public trials of taking passengers (https://waymo.com/apply/). There's a buttload of Waymo vans in the Tempe/Chandler/Scottsdale area. The only accident I'm aware of with a Waymo vehicle was 100% not at fault.

https://www.youtube.com/watch?v=KhcyTOaHdv4

 

[Midwayman]

Even so, I think Autopilot is probably a better driver than most people on the road. I drive 50 to 100 miles daily and almost get in an accident every single day from other people acting crazy.

On a restricted access highway or other simple situations during good conditions, probably. Certainly true for the google cars. I wouldn't trust it for downtown traffic, or in the snow or something.

 

[Zorba]

Huh, interesting. Most of our stoplight roads here max out at like 40 MPH. I know of a turnpike that is basically a 2-lane road on each side (4 lines with a barrier in-between) that is 55 MPH. Maybe it's just because it's really hilly where I live.

Although when I was out west I did hit some 80 MPH highways, which was pretty awesome!

There are a lot of roads around here that are 65 or 70 with stop lights or even stop signs. No reduction in posted speed limit either. Basically all of our two lane highways have a 65 mph speed limit and lights or stop signs.

 

[Zorba]

Cause of firetruck crash determined:

http://www.thedrive.com/news/20912/...crash-into-fire-truck-cause-determined-report

The gist of it:


In more detail:

https://www.teslarati.com/tesla-autopilot-crash-details-nhtsa-investigation/

TL;DR: Autopilot crashed her car straight into a giant firetruck at 60 MPH.

So there are 3 high-profile Autopilot crashes:

1. The semi-truck in Florida (fatality involved)
2. The barrier in California (fatality involved)
3. The firetruck in Utah (broken ankle)

In each case, the driver was not using Autopilot as intended. They were distracted and were not watching the road and were not babysitting the system or their situation as required. The software is in beta and has reminders about proper usage.

Again, on the flip side, I still don't like Tesla's "wink wink" approach to Autopilot usage -
'yeah yeah it's in beta, but it's a self-driving car, man!'. I don't like the name Autopilot in terms of its currently available feature set. And I really don't like how they brush off accidents. Their response to the California crash was pretty crummy - they tried every angle they could to blame the driver, even going so far as to cite a quote by the victim's family that he had had prior issues on that stretch of road, so why was he not paying attention? And Elon Musk's response to the latest crash where Autopilot drove the user straight into the back of a giant firetruck at 60 miles per hour was this:


Is he wrong? No. Autopilot is supposedly statistically safer than all other methods of driving at the present time. But the software literally drove a car directly into another vehicle. And Tesla does whatever they can to avoid saying that directly! Yes, technically, per the letter of the law, it's labeled as beta, has some warning chimes & lights, and you're supposed to babysit it while it's self-driving. In practice, people don't buy a Tesla with Autopilot so they can babysit it, they want the self-driving features, obviously!

I don't think it's realistic to expect to NOT fall into the trap of distracted driving when you have a (mostly) reliable self-driving car on the highway. imo Tesla could exercise more responsibility in how they handle alerts & in how they respond to accidents where Autopilot caused death or injury. We understand the Autopilot is a safer way to drive, but that doesn't mean it's perfect, and I certainly don't like how they brush it under the rug & point fingers instead of accepting any level of blame for their system running people's cars into stuff & hurting or killing them.

And to be clear, I like what Tesla is doing, both in terms of electric car progress & in improving automotive safety, both through the design of the car & the Autopilot system, and I don't think they should stop, because we'll never get to where we need to be if we try to be perfect first - it's going to be incremental improvements along the way, and negative things like these Autopilot crashes will continue to happen because the world is complex & it's a risk. For now, at least, I'd just like them to publicly shoulder more of the responsibility for Autopilot crashes & also improve the alerts & alarms when the system is not being used properly. I just don't like how they try to hide Autopilot as the root cause of the crashes.

Meanwhile, there automatic braking on my Subaru would've stopped me from hitting that firetruck... Obviously autopilot isn't actually road ready, but Elon it's prefectly happy using the public as beta testers for a potentially deadly system.

Also, this is the second high speed rear ending of a fire truck in 3 months. The first was quickly over shadowed by the fatality.

 

[rstrohkirch]

Stuff like this feels so minute to argue over sometimes. It's like the recent articles that around two dozen deaths over the past 5 years have been contributed to keyless cars being left on. Drunk driving kills around 30 people a day on average. I wonder how many people die every year because they don't look before entering a cross walk. More effort on bigger problems would seem ideal.

 

[zinfamous]

Stuff like this feels so minute to argue over sometimes. It's like the recent articles that around two dozen deaths over the past 5 years have been contributed to keyless cars being left on. Drunk driving kills around 30 people a day on average. I wonder how many people die every year because they don't look before entering a cross walk. More effort on bigger problems would seem ideal.

Humans are irrational. Statistically meaningless deaths by unthinking robots are monumentally more terrifying than the vast number of daily deaths by other means that we have long come to accept as a way of life, today.

 

[Kaido]

There are a lot of roads around here that are 65 or 70 with stop lights or even stop signs. No reduction in posted speed limit either. Basically all of our two lane highways have a 65 mph speed limit and lights or stop signs.

Wow that's weird, I guess the terrain around here just doesn't accommodate for those speeds with a stop sign or stop light.

 

[Kaido]

Meanwhile, there automatic braking on my Subaru would've stopped me from hitting that firetruck... Obviously autopilot isn't actually road ready, but Elon it's prefectly happy using the public as beta testers for a potentially deadly system.

Also, this is the second high speed rear ending of a fire truck in 3 months. The first was quickly over shadowed by the fatality.

Yeah...it's pretty ridiculous. And while I agree with what Musk said about media companies needing to focus on the 30k+ people killed annually in the United States, he was purposely deflecting from the bottom line: Autopilot drove that car straight into the back of a firetruck. Yes, it was the driver's fault for not paying attention, but like you said, the AEB on your Subaru would have prevented that, or eat least reduced the speed from 60mph to reduce the impact of the crash.

 

[Kaido]

Stuff like this feels so minute to argue over sometimes. It's like the recent articles that around two dozen deaths over the past 5 years have been contributed to keyless cars being left on. Drunk driving kills around 30 people a day on average. I wonder how many people die every year because they don't look before entering a cross walk. More effort on bigger problems would seem ideal.

I don't disagree from a large perspective, but self-driving cars is something new & something that will have broad impact in the coming years worldwide, so it's worth paying attention to the individual happenings because this is not only a technology we need, but also a technology that we need to improve. The sooner it rolls out & the better it gets, the less number of people will die in the future.

The entire American annual automotive death rate could be dropped to zero if we limited all cars to a top-speed of 20mph & required that your car have seatbelts, crumple zones, and airbags to be road-worthy, but economics dictate our situation, so it is what it is. And I do applaud Tesla for pushing this technology forward, despite it not being perfect...it's never going to be perfect, and it needs to be tested en masse for the eventual automotive AI systems to learn all of the different types of roads & driving situations, so I don't necessarily disagree with their approach of releasing beta software, either. I just think they should have waited to call it Autopilot until it was out of beta, and that their responses should be better in terms of accepting blame for the crashes instead of just pointing the finger solely at the driver.

 

[bradly1101]

In my opinion autonomous will never work until all cars, roads, and pedestrians are electronically talking to each other.

 

[zinfamous]

In my opinion autonomous will never work until all cars, roads, and pedestrians are electronically talking to each other.

To me, that is the ideal, utopian, perfectly-realized automated transport system. It can very likely work perfectly--but this would require a generation or two of humans that really only know such a world and the expectations of living in that world (i.e.: jaywalking? lol, you prolly gonna die and no one will gaf in that world and/or well-gated pedestrian crossing zones that strictly limit crossing to those intersections. The only traffic signals in this world are really for timed pedestrian crossings, because a fully automated road system needs no traffic signals. Think: ant colony or bee hive in perfect synchrony) and, as you say, some serious, serious, tech advancement, cash, and public will to get there.

This is the one tech sector of life that I really want to see happen, tbh, and I am both a rather analogue person (I have very limited need in my life for smartphones, no social media, etc. I prefer walking and old-fashioned cabs when appropriate)...and I would miss the "freedom" of driving....but the reality is that all of us have actually been missing that for decades, some of us are just too stubborn to admit that we don't live on those desolate roads that only exist in car ads, or the experience of such is only 0.02% of our actual driving lives