Software is eating the world, automobiles included.
I'm a huge skeptic when it comes to Tesla, but JLG raises a good point here. Apple had all sorts of problems in 2007, so many that established phone manufacturers largely brushed them off. A focus on software rewrote the mobile phone industry, and every day I get into my car and watch the onboard computer struggle to handle bluetooth properly.
I need to look beyond the obvious inanity of Musk and remember this shift: software will eat this part of the world too.
> and every day I get into my car and watch the onboard computer struggle to handle bluetooth properly.
I hate "me too" posts, but this gives me so much rage that I can't help it. My car has a touch screen interface and I am puzzled why car manufacturers even bother. Who thinks the interface they made is even remotely good. It actively degrades the otherwise fine experience of my car. It is ugly, poorly made, and fails at even basic functionality.
I am truly puzzled why these car manufacturers released such terrible user facing features. Maybe it's better these days (my car is ~5 years old), but it's just impressively bad in my car at least.
I totally agree with you, having also had to deal with rental cars with batshit crazy touchscreens that barely functioned, lagged like crazy, and were obviously not fit for the purpose for which they were built.
My last car was a 2006 G35 with no screens as all, and a nice analog clock. I loved that about it so much.
I never understood why manufacturers tried to reinvent the wheel and not just use the phone for all things entertainment and navigation. I would never want GPS in my car, because they all were so horrifically bad.
Until I got the Model 3.
It take less than 1 second, literally two taps in almost exactly the same location on the screen, to activate GPS to my ‘Work’ or ‘Home’ presents. And if I do need to dial in an address it has a full keyboard, full POIs, and autocomplete which is as fast as Google Maps on my iPhone X. Traffic aware navigation which appears to use the same feed as Google Maps (estimates and routes seem to always match). The GPS on Model 3 is a joy to use, actually, truly.
The best part is that it implements very good interpolation when GPS drops out using compass and velocity, so going through tunnels tracks almost perfectly compared to Google Maps on iPhone which gets hopelessly stuck.
The interface is very reminiscent of an iPad, just as quick and responsive, and has a bunch of cool touches.
For example you can drag the media player to be hidden, to be just showing key info, or to nearly full screen, based on the level of interaction you need in it. The drag is a quick flick of the finger and doesn’t require precise targeting at all. It’s amazingly natural to navigate. While driving all the media control you need (volume and skip) is on the steering wheel. You get free Slacker and free LTE and I’ve really enjoyed the streaming music and have actually been introduced to some new artists I didn’t know before.
I never wanted a touchscreen in my car. Full stop. Until now. The fact that it’s all touch with consistent OTA upgrades give me confidence that it has staying power to remain functional, relevant, and looking elegant.
Particularly with the way Tesla designs their cars, I would be surprised if Model 3 wasn’t essentially the same interior design 10 years from now. I know they designed 3 in particular to have as little as possible that might look “dated” in the future.
I just got a 2018 BMW 2 series and it's the best system I've personally used outside of a Tesla. It has a mobile connection so it can find POIs very quickly. It's surprised me many times when I've asked it to navigate somewhere.
I'm sure the recent crop of Audis and Mercedes are the same. The German brands seem to be finally getting on the voice assistant train with a "Hey Mercedes" or "Hey BMW" cue.
Your BMW is new, I have an old one, and the sofware updates made it worse actually. Tesla is making sure that the new processors are backward compatible (for example the new, v2.5 generation autopilot hardware fits in the v2.0 hardware, and doesn't increase the power usage substantially).
In return I'm a bit scared of buying a Tesla because I live in Europe and read quite bad stories about the service wait time here.
Did they see themselves as a "hardware company", or as a "computer company" at a time when the hardware and the software all came together, and there wasn't really a separate software industry?
Which would make them the Apple of their time, where they built and sold the hardware needed to deliver the software?
Did they see themselves as a "hardware company", or as a "computer company" at a time when the hardware and the software all came together, and there wasn't really a separate software industry?
I think they did see themselves as a hardware company. Software was something they had to make to make their hardware work, but they didn't think of it as a major part of their business. I happened to work for the "world's first software company" which was founded by a star IBM salesman. (Cincom Systems Inc. in 1968.) The Smithsonian has a write up on it. It was the first time that a company was intended to make all of its money from the sale of software, and succeeded in doing so.
My response to that is to say that IBM belongs in a third group—the fundamentally systems-oriented company. Systems-oriented companies are capable of understanding and delivering the whole product without either the hardware or software seen as more important. This would include the likes of 1970s IBM, Apple, Sun, Atari, Commodore, etc.
It's not any better. This past year I've driven something like 2 dozen rentals from a variety of manufactures, they all suck. The only saving grace is sometimes I can use CarPlay and I'm much happier for it.
The fact that I have to struggle to adjust the climate control while driving to click through their terrible interface is just insane.
> My car has a touch screen interface and I am puzzled why car manufacturers even bother.
IMHO, touch screen controls should be banned from cars, no matter how good. Tactile physical control based interfaces allow drivers to keep their eyes on the road. Touch screens inevitably take driver's eyes off the road and distract them, since they need to use those same eyes to figure out where on the screen they need to touch.
>Maybe it's better these days (my car is ~5 years old), but it's just impressively bad in my car at least.
I honestly don't think it's getting any better. At best, it's just getting different - but maybe worse as manufacturers try to add more "smart" features.
Was a passenger in a <1 year old BMW 5 series for about a week this year for a roadtrip. There were so many stupid behaviours with the software that we effectively gave up using it for anything except bluetooth music. Trying to use the built in GPS navigation was awful.
Also a passenger in a new SUV style vehicle for two weeks last year, I forget what brand. It too had utterly bizarre behaviours that it wouldn't explain anywhere, and it took a variety of experimentation to figure out. For instance, Bluetooth pairing is disabled once the car is in put into gear, and is not re-enabled until you turn the car off and turn it back on again.
However it doesn't tell you this, it just doesn't show bluetooth options in the menu anywhere. So I spent hours going through every menu item (as a passenger, mind) over a few days until I finally figured it out.
In every case, the number of either unlabelled or poorly labelled buttons is overwhelming.
My Toyota "grays out" all Bluetooth options other than "choose from paired devices list" while the car is in motion, deliberately. Coming to a stop re-enables the other options, though, at least making the cause obvious. BMW is absolutely making the right decision here, but apparently communicating their reasoning poorly.
> For instance, Bluetooth pairing is disabled once the car is in put into gear, and is not re-enabled until you turn the car off and turn it back on again.
That's not a bug, that's intentional, and saves lives.
If cars only contained drivers it wouldn't be that bad.
If only there were some way that someone besides the driver could manipulate the in car system.
My Chevrolet handles this well, using the passenger seat occupancy sensor. If someone is sitting in a passenger seat, it gives the option "I am the passenger" to allow access to normally blocked-while-moving screens.
And the behavior the parent described is just annoying. The device has the capability to know if the car is in motion or not. Permanently locking out the function until the car is turned off and on is dumb.
I agree with you on the touch screen interfaces, especially considering how long people keep their vehicles nowadays. I have had my Tacoma for 11 years. Had I gotten one with some sort of touchscreen or other display interface, it would seem like utter dog shit now using a touchscreen and computer system from 2008.
That's the first thing I think about when I see new cars with their fancy displays. How will this look and perform in five years? In ten?
“That's the first thing I think about when I see new cars with their fancy displays. How will this look and perform in five years? In ten?”
I wonder about the same. I think there is a good chance that cars will go the way of the computer where it’s almost inconceivable to use something that’s ten years old. So people may buy cars every five years and used cars will be recycled instead of being sold as used cars.
As someone with a car from 2015 that lacks both touchscreens and infotainment options due to going for the "value edition" I dread the day I have to move on from it. Of course it still has a terribly slow bluetooth connection time but I'm gonna miss real gauges and physical controls when it comes time to replace it. I also fear that it will be the last manual transmission car I've owned if those vanish from the market.
Mercedes-Benz is certainly egregious in this regard. Their screens are non-touch screen and require you to move a weird joystick to move about major functions. Their radio channel display is a tired version of the iPod/Finder song carousel view and direct access to premium channels is a chore.
Their Apple Carplay adaptation is a least free as opposed to BMW charging some bogus subscription.
Big fancy screens were an effective way to sell cars 10 years ago. Now, technology is so cheap, it's just cheaper to stuff those fancy screens into every car.
I think for SUV's (the new ones) backup cameras are pretty much a standard safety feature because you certainly can't see behind them whatsoever due to their shape.
Part of the problem is Bluetooth itself. The protocol is an over-engineered abomination full of totally unnecessary statefulness and special-casing and other nonsense. It's really hard to get Bluetooth to work reliably.
My Model S has a worse Bluetooth implementation than my Chevy Volt did.
Tesla’s software is objectively pretty terrible.
Sometimes my center console reboots randomly while I’m driving, which means I lose all audio, and the climate control shuts completely off for several minutes.
Tesla has not yet achieved Apple quality polish on their software.
This is a good observation, and since I don't own a Tesla, I appreciate your perspective.
I definitely don't love Tesla, but I guess I think of it like this: I suspect Tesla can improve their software (and deliver it to all existing customers) a lot more quickly that their non-software-focused competitors can. I still don't think Tesla will ever have full auto-driving, but I know my car's Bluetooth handling in my 2014 Toyota sucks, and will always suck. If a Tesla's Bluetooth handling suck, there's a good chance it won't in another 5 years.
And Tesla isn't even a company I'm eager to defend, but I think JLG is right: software-focused companies will eventually beat out those which aren't. It'll take a while, but it'll happen.
I have a 2018 Leaf. Drove a model 3 for a day last month. The software of both vehicles is not even comparable. It's like going from a Nokia 3310 to a modern iPhone.
Interesting analogy because there are those who would consider the 3310 the better phone in terms of build quality and having a device that minimizes distraction due to how limited its functionality is.
My 2015 model S has appallingly bad voice recognition and turn by turn directions compared to my phone (1st gen Android Pixel) from the same year. Both have software updates but the car is really, really bad when it comes to those things.
I'm happy with the bluetooth though, and I've never had a random reboot.
I've actually had cars with fantastic voice recognition, but they have always been when using Android Auto in the car (so it's not really the car doing the voice recog)
I used AA a TON in an Audi TT convertible and it would correctly understand me even while driving on the highway with the top down!
Funnily enough, the built-in voice recognition in the Audi was horrible and was hard to get working right even when sitting still.
I have a model 3 and the Blutooth seems about average. All the rest of the software seems way better than any other car I've ever been in though. Wonder what your complaints are?
That statement has underlying meaning. It implies that it has not worked perfectly and lag-free every single time. It implies you have identified ways it could be better.
Yet it also implies you are happy with its level of buggyness. You have come to terms with it. You don't complain when occasionally it won't connect without cycling Bluetooth off and on again, or when there's a slight glitch.
If I design or develop a technology, I would consider it a 100 percent failure if as well as there being bugs with it, users stop complaining about the bugs and they just become 'normal'.
Often I have to hold my phone to my ear for a call because the audio doesn’t come out over the speakers. It’s random, though - Bluetooth sometimes works great.
I love the car because it’s fast and electric. But the software sucks. I was expecting better.
The moment Tesla cares more about the software experience of existing owners, than things they can brag about to reporters and observers (who don't actually own/drive their cars everyday), that will hopefully change. When that'll happen, I have no idea.
Tesla does try to produce a better experience, in theory, than everyone else. It's the small details and bugs that you'd likely only notice as an owner, which they seem to struggle with (or simply not care about).
Why is the theory that they try to produce a better experience? Where does that come from? Probably from their self-promotion and tech journalists eating it up, and nothing else.
Not sure what you're talking about in terms of Apple's 2007 problems. They were highly profitable, had successfully migrated their computers to Intel chips (an incredibly thorny technical platform hurdle) and the iPod was a global cultural phenomenon that never faced a true competitor. All of their products were arguably best of breed. Apple stores were already the centerpiece of most upscale American malls. They were firing on all cylinders.
Specifically with the original iPhone: no 3G, no MMS, semi-frequent crashes. I owned an original iPhone in 2007, so clearly the problems weren't enough to deter me, but it wasn't a perfect device.
It was, however, in an entirely different conceptual category from its competitors.
The same could be said of Tesla, I guess, but I'm still very much a skeptic on Tesla. My point is just that being in a different conceptual category is a huge advantage. Perhaps Tesla will capitalize on that and come out on top (as Apple did), or perhaps it will be some other company (as Google did with Android), but I think it will happen.
- Selling a $1k phone or a $2k computer is very different from selling a $80k luxury car
- A smartphone is a full fledge computer, which is a massive change from a dumb-phone (which capabilities were limited to calling and SMS). An electric car has the same features as a gas car, with the exception that its drive chain has a different source of power.
- Apple was an established business for 30 years before 2007, and navigated a software/hardware development environment (90's) that was very immature prior.
> software will eat this part of the world too.
Yet we're talking about a company that sells a big hunk of steel. The quote "software is eating the world" has to do with the replacement of human beings by software, not with "shiny new thing that still needs to be powered by energy".
PS - Tesla's fully autonomous (Level 5) solution is unproven. Until then, let's take the "software is eating the world" talk out the question.
>A smartphone is a full fledge computer, which is a massive change from a dumb-phone (which capabilities were limited to calling and SMS).
I'm sorry, but this is the very blindness the old smartphone manufacturers exhibited. Nokia, Sony-Ericsson, RIM, they all had well-established offerings of smartphones with the on-paper capabilities matching, or rivaling, that of iPhone. Unified messaging, camera, keyboard, handwritten recognition, web browser, you name it. Even apps ecosystems. All good, right? And yet the iPhone won via software. And via the hype. And via the app store in the longer run. And via tight integration with other Apple offering in the longer run. And via upmarket, "cool" brand.
The customers were sold "the experience of future, today". Cringe? Sure, but it worked.
>An electric car has the same features as a gas car, with the exception that its drive chain has a different source of power.
Wrong, very very wrong. A gasoline car is goods sold via dealership network at about zero margins, with the tacit understanding the customer will be back frequently to change oil & filter & timing belt, service the engine (almost every new generation of ICE comes with serious deficiencies), upgrade firmware, and all the jazz. The money for the dealership is in the servicing, not in the initial sale.
An electric car is, for better or worse, mostly absent those services. You rotate tires, replace brake pads & shock absorbers once every several years, and that's that. Any money to be made is in the charging network, if at all.
There is a dual shift - software/perception/hype, and also the business model. The traditional vendors have already mostly figured out how to put electricity in cars (if at low efficiency); but they desperately need to un-learn how to sell cars via the legacy business model. Right now they aren't ready to ditch the dealerships, while at the same time they want to sell ever more of the dealership un-friendly electrics. Talk about stuck between rock and hard place.
Tesla does not have SAE Level 5. It is an absolutely ridiculous claim. At most Level 3, but the SAE levels are kind of fuzzy between 2 and 3, so you could even argue level 2. [1]
Tesla Autopilot cannot and will not be able to deal with ALL situations that can possibly occur - which would be SAE Level 5. They are spouting marketing bullshit in face of their poor quarterly results. Most engineers in the space will agree that level 4 will be the highest level reached for decades - if ever. Showing an Autopilot in heavily trained-for area during perfect weather and road conditions with infinite retries for getting the video right means nothing.
This has to be viewed separate from some real progress and contributions that Tesla has made in the space. However, that doesn't mean they get to throw around the "Level 5" term yet. And they likely won't within the next decades. Level 5 is a whole other beast than Level 4 and nobody is even remotely close to get there.
Point 2)
Have you ever sat in a modern car beyond 1995? They ARE computers. And in most recent years, user access is widened. 10 years ago an onboard computer would calculate your air-fuel mix and you wouldn't ever interact directly with it. Today there are huge and ever growing interfaces and you can use voice control to make your car perform certain tasks.
So yes, software is eating up the industry. Doesn't mean the steel hunks disappear.
Tesla is presenting a new model for automobiles where your car isn't outdated every year by a new trim design or every 5-7 years by a generational uplift. Instead your car evolves as you own it and enhancements are made incrementally both to the software and hardware rather than being held back for annual or generational releases.
In the less than 6 months I've had my Tesla Model 3 it has had software updates to:
- improve braking performance
- increase surpercharger performance
- increase power output, speed, and acceleration
- increase range
- enhanced automatic electronic braking
- enhanced lane departure warnings
- add dashcam recording capabilities
- add deterrence, video monitoring, and logging security features
- add climate control management for pets
With a traditional car company all of those things would be added incrementally in subsequent years or not at all.
Or, especially the first one, would already have been there from the very beginning.
But I see the advantage of Tesla's approach. On the other hand I can sign up with our family Peugeot for regular updates for free for, it was, 2 years or so. Not over the air the registration is a pain. But it's not that it isn't there.
No. That braking problem would have plagued ANY car manufacture. The only difference it there would have been a mad rush of people setting up service appointments at dealerships all over the country to get it fix -- costing time and money for everyone involved. Tesla did OTA updates, no fuss.
My Chevy Volt has had two software updates in the past 6 years.. not user noticeable improvements. I love my car but the car I was sold in 2013 is the same (problems and all) car I have now. My friends who have had their TM3s for over a year have drastically improved cars in almost every aspect.
On the flip side, my Nissan's connectivity module was 2G based and deprecated by the carrier shortly after the car was built. Nissan offered a free upgrade and I was eligible but I only knew about it because I'd read something on a forum referencing a service bulletin that I never received despite being the registered owner of the vehicle. The excuse being that the bulletin must have been lost in the mail.
There's more to your second point: aren't smartphones still a computer connected to an almost-completely separate cell phone? Haven't the phone producers repeatedly had problems with things like antenna placement and design? Google has repeatedly released their own physical phone, but it's never gotten much market traction.
I think it is very tempting for software guys to think that software will indeed eat the world.
Fundamentally, until we reach actual self-driving cars (which I believe to be far enough away that it does not merit much discussion of the current car market), cars need to drive first, software second. Given the shit show that is software updates on my workstation, I'm pretty sure I don't ever want my car to be that focused on software. There is something to be said for being careful -- I think the legacy manufacturers ought to embrace over-the-air updates because that just makes sense, but I hope they think more like aviation (but not Boeing ;-)) than silicon valley when they plan their release schedule.
I can't speak for anyone else, but I'm not even talking about the full-self-driving pipe dream. In general, despite real technological shortcomings, the iPhone approach to software beat out approaches of Nokia, Motorola, et al. It's easy to see how a software-focused company--maybe Tesla, maybe not--could do the same with the existing car companies.
For a tiny example of how I'm thinking: Apple introduced CarPlay, and it took years for car companies to be ready on their end. Super-long lead times seem like a good idea when you're talking about airline safety, but for bluetooth audio? For touch-screen user interface stuff? Think about how much iOS and Android have changed just in the last five years, and think about having everything delayed by five years because "we're a car company, not a computer company."
I don't just mean everything that happened in 2012 happens in 2017 instead, I mean that the first release is delayed by five years, which means that it isn't until that release hits the market that developers can see what direction to go in based on user feedback, and then there's another five years before that hits the market, and then competitors are also doing things, but it's another five years before that has any effect, and so on.
I'm as terrified as anyone ought to be about the idea that Tesla can over-the-air release a regression that drives someone into a concrete barrier it used to navigate properly, but the details of Tesla aside, the whole idea of focusing on software seems like a good move, and I think the car industry will (much more slowly) follow the phone industry overall pattern.
Well, I'd say it has more to do with having billions of dollars to throw at the effort, and a small but enthusiastic niche of tech folks with good salaries and an appreciation for pretty touchscreens and no real concern about driving dynamics. Followed by some image-seekers who hope that a Model 3 in their driveway could be mistaken for a Model S.
Sales are already faltering and money is getting very tight over at Tesla; one of the reasons I opted out of buying a P3D and instead am waiting for a Model Y is because by the time that hits the markets we'll have a better idea on whether Tesla has a real future or not.
Well that's the beauty about Tesla - it's the most fun car to drive. I feel like 99% of people who don't like Tesla have never takes a test drive - it's just very different from all the other cars out there. Software is just the icing on the cake (but also a strong competitive advantage in the future).
The iPhone has a user interface and a very strong network effect - more users means more apps, which means more users.
Self-driving software is fundamentally different. A self-driving Tesla would have no significant UI or network effect, so users could easily switch and Tesla won't have a durable competitive advantage. As soon as there are two robo-taxi services, people will simply choose whichever one is cheaper. It will be a commodity service similar to air travel.
Which, one could argue, opens up a new front for Tesla on configuration management. This is bad enough for hardware and poses already a problem regarding spare parts. Add the additional layer of software that might or might not work with a certain hardware config and this problem increases exponentially.
I'd live to see the Tesla announcement that from today on they will focus exclusively on X because of Y and the measureable and benefits will A, B and C til date D.
But all we get is "full self-driving next year" and "fleet of autonomous robo-taxis".
Mentioning configuration management, that's something Apple is really good at... it will be fascinating to see what comes out of their efforts eventually.
I almost prefer their approach of keeping things under wraps until they are ready, rather than pre-teasing stuff that might happen later than predicted.
To be fair to Tesla, they are not saying we will have full self driving next year... it is Elon making a prediction, not a promise, and he couched it many ways. Full self driving for Tesla development cars, basically, and robo-taxis possibly by the end of next year, with safety drivers, if you listen to what he said in response to an audience question... so his actual statements are pretty far removed from what you have dressed up as "quotes" right there.
I'm still puzzled by the meat being transported along packed freeways every day. Other than the segments of the economy that require face to face interaction, it is stupid, wasteful, dangerous and destructive.
What is your alternative that the world can switch to without an enormous cost? Utopias are fun to think about, but I am skeptical if flawed humans could ever produce one.
It's worth emphasizing that Tesla is strong at software relative to other automakers; in absolute terms, the (non-AP) parts of Tesla's software frequently have user-facing problems [0] and appear to have been built on a Babelian tower of hacks [1]... but it's still better than the rest of the market, and it keeps improving month over month.
If there's one thing that fills me with dread is buying a car knowing that there's going to be a dozen obviously awful and easily fixable aspects of the software that will bug me every single day—every single time I drive it, until it is replaced by the next car with a dozen different obviously awful things.
Tesla, for all its faults, offers its owners hope that of the dozen obvious and easily fixable aspects of the software, some of them might actually get fixed in that car's lifetime.
(And sometimes you get fart jokes. Fix all the damn bugs; only then do I give you permission to fart.)
Love my TM3 but as I have posted before, Tesla needs to focus, Musk needs to focus. Tesla cars have some great user facing features but also are woefully behind in other areas.
The most flagrant is the awful support for playing music from blue tooth or attached devices. It has been common for years to be able to use the touch screen or menu to select playlists, artists, or more, from the car's UI. Not so with Telsa, want to do that drag your phone out and change what is play. You can only navigate within a playing playlist. Seeing how its illegal to use your phone in your hand in many states Tesla needs to remedy it.
We have none of our promised voice commands, no improvement since when I bought mine in August. Our energy meter is near useless, does not track when car is not moving, and only tracks overall use and does not break out HVAC, regen,or drive train use.
But we got fart humor, animated rain drops, and MAME. Their priorities suck, its no wonder why they aren't profitable, they cannot get done what they started before going off doing something else or wasting resources on stuff that does not matter
Ever get a flat tire in your Tesla? I did. No shop would touch it and I had to have it towed to Tesla where it took 2 days and $400 for a simple tire change.
My blood boils when I see Musk tweet about putting more games in the car. They need to be reaching out to mechanics with training information so that they're not so spooked that they won't even change a tire.
I've gotten several flats over the past 6 years on my Model S. I take it to Discount Tire almost every time. Only time I really to take it to Tesla is when I hit a pothole so hard I had to replace a rim. Even then, I could take it to any Tesla-approved body shop if I wanted to.
Well that gives me hope that my experience was an anomaly. I was 200 miles from home with a disabled car so after calling several places I felt like I was out of options.
This is what the joy of OTA updates feels like Shivetya!
There is no other auto manufacturer on the planet where you could complain about wanting news features XYZ instead of the delivered ABC and actually expect something might be done about it!
> But we got fart humor, animated rain drops, and MAME.
This is unfair. Aside from the humor and slight touches, which sometimes is delightful, we also got Dashcam, Sentry, Dog Mode, updated Climate controls, Blind Spot display, zoom-in speed award guidance display, improved auto-park, improved Summon, and probably another two dozen features I’m forgetting.
That’s aside from all the core AP enhancements which are almost too numerous to list.
Media is hard because they probably refuse to make the interface work for just a single model of phone, or even a narrow range of software versions of a single model of phone. They don’t want to write and maintain features that need to be tested across 100 models and versions of phones from a dozen different countries.
I would just laugh at Elon's crazy proclamations, but there are plenty of people buying Teslas who believe them. At this point, Musk's claims verge on the fraudulent. Fully autonomous driving is never coming to current Teslas, sorry folks.
As for the frequent OTA software updates, I wouldn't consider that a feature. While it's true that at this point software is a critical piece of the puzzle and automakers who haven't woken up to the idea that owning your software is as important as any other aspect of the design may eventually be left behind. But I don't think Tesla offers a good model to emulate.
As for JLG's specific points, providing a single, centralized method of updating automobile software over the Internet--or even a specialized wireless connection--should be seen as a security risk, not a powerful feature. Most critical systems in cars should be isolated from each other and from the Internet entirely. Updates should not be haphazardly distributed, but should be carefully tested and rolled out as rarely as possible and in a controlled manner.
I have no hope that car companies will actually heed advice like this. But it's disappointing to see tech veterans buy into the delusion that the equivalent of Windows Update, but for your car, would be a good idea.
The OTA updates for car are more important that for any device IMHO. Phones are 'cheap' enough where you can upgrade every two years w/o breaking a sweat. Cars OTOH cost much more and getting software refreshes provides tons of value. Think about it, 2012 Teslas are STILL getting OTA software updates. How many devices do you own that are 6+ years old that are getting updates on a regular basis?
To me the primary value of OTA updates like Tesla’s is that the dash systems won’t become worthless in a few years. I don’t own a car myself, but it always grates me just how bad and outdated the dash systems in cars even just a handful of years old are when taking ubers or riding with friends. It’s generally not found acceptable when major features of any other expensive purchase are phoned-in at best, so why is it tolerated with cars?
At minimum, I think car manufacturers should decide on a dash platform standard and stick to it for at least a decade, making it possible to continue to properly update and support existing cars for a reasonable amount of time. Tesla isn’t doing anything special, it’s just a plain old off the shelf ARM SoC running a touch-optimized Ubuntu variant. Surely Toyota, GM, etc could figure this out?
Barring that, they could save money and exclude built in dash functionality all together in favor of CarPlay and Android Auto, but that could be problematic for customers who don’t use smartphones.
I don't own a Tesla. Can you provide evidence on why Elon's claims are fraudulent? I was thinking about a Tesla as my next car and would like to learn more.
Do you know if the Tesla over the air updates are altering critical systems without diligent testing? I assume only Tesla engineers would have insight into this.
OTA updates have had literally deadly consequences. There was an issue in which "autopilot" poorly recognized left-exits, including a particular one in California. It got some publicity and was fixed by an OTA update. Good! Except that a couple of months later another OTA update reverted whatever part of the "autopilot" model had learned not to drive straight into a concrete wall separating a left exit, and it killed someone. So yes, ML combined with OTA can certainly be dangerous, but Tesla fans will soon show up in droves to explain why that man's death is his fault, or the fault of anyone but Tesla.
As far as fraudulent claims, Musk has repeatedly made statements related to self-driving that some people consider to be well beyond the pale of optimism, including promising that every Tesla sold after a certain would include all hardware needed for full self-driving, and then doing the same thing a few years later, apparently rendering the first statement factually false in retrospect.
Musk has made many, many, many statements about timelines that have not worked out, and some that probably still won't work out, and some people label that fraudulent. I think it's just naïve, but as the CEO of a public-traded company, his naïveté is pretty stunning and has resulted in fines.
Mentioning SpaceX and autonomous Teslas in the same breath makes me wonder if the Starlink satellite internet program is related to his plans for self-driving cars.
There's synergy between everything he works on, it seems to be a natural extension to his flow of thinking - expanding further into the next biggest market. It's the ESSO (gas company) model: they needed a truck to transport fuel, so they designed a truck to manufacture, and sell it to others too, and so on.
The difference between NIH and vertical integration is in the result. When accounting for cost of effort, does it end up inferior to some third-party solution, or does it allow for a more efficient business process?
Most of the time, you're right: one is better off going with a solution put together by some other group who made it their core expertise. But if you keep bumping into limits, and have the resources to throw at the problem, you can develop a compelling solution in-house.
It's a bit like the tradeoff (for enterprises) between buying a turn-key solution and going with an opensource project. The latter is often more expensive (once you factor in the cost involved in figuring everything out) BUT will end up much more taylored to your needs—especially if your usage involves controbuting patches to the project.
Apparently it took over a year for the coffee to kick in! I just always shrugged and thought it was a cutesy name for a company that did something boring like dig tunnels. I cannot believe that pun escaped me for so long. :)
Nobody is boring cities on Mars with diggers designed to bore small sewer tunnels. Boring Co still has a long way to go to demonstrate basic competence at digging small tunnels and this is an area where obstacles increase with scale.
It makes sense to pursue technologies that can pivot from expanding humanity's potential to simply guaranteeing survival. But only the former is what people are going to want to hear about, while the latter is where we seem to be heading.
Why not just "cars"? Teslas with or without FSD capability already use Internet connections for OTA updates, maps, audio. Other manufacturers will likely follow sooner or later. That's a good outlook for any global Internet provider... I'd offer maps, audio services to other car makers in his position, wouldn't you?
Starlink will require a phased array antenna that is pizzabox-sized and also not cheap (in fact, making them cheap enough to make the business case worthwhile is a major challenge). Using 3G, LTE or 5G seems like the vastly cheaper and easier option.
> Using 3G, LTE or 5G seems like the vastly cheaper and easier option.
Starlink isn't meant to compete with that. It's main target markets are remote areas than are not connected by fiber, planes, and boats which are all currently serviced by low bandwidth and high latency geo-synchronous satellites.
I don't believe the physical antenna itself is difficult or expensive. The expensive part is all the digital phase control circuits that are needed to actively steer the antenna beam. A electronically scanned array antenna is actually a whole lot of little antennas fed by an equal number of phase control electronics. The phase control electronics need to be small and inexpensive (since there are a lot of them) and also able to accurately adjust the phase of a very high frequency signal (24++GHz). That is difficult.
Not sure how their antennas are constructed - if they are fabbed on a PCB, often at microwave frequencies you need exotic substrate material (e.g., Rogers) to cut down on losses. Those substrates are expensive compared to FR4 (fiberglass common in PCBs).
Metal additive manufacturing has never solved a cost problem AFAIK. The problem it solves is the manufacture of parts that have physical characteristics that are difficult to machine, or in the case of Inconel, materials that are simply difficult to machine period. Neither of those characteristics are going to be a problem for the Starlink antennas.
A small series (when producing other tooling would be too expensive; a single, bespoke part would be the extreme thereof) or ones that would be impossible to machine otherwise are candidates for additive manufacturing.
I don't really see how the Starlink constellation is vaporware. They've launched 2 test satellites into orbit and are less than a week out from launching the first batch of 60 satellites into orbit.
I think when you've reached the point of a 60-satellite production run, integrated on the pad and waiting for launch, you've at least passed vaporware.
Tesla and Musk do a lot of things. Some are clever, eg: Tesla gets marketing that other car companies can only dream of, for free. Other things are insane, eg Musk's "secured" bid of $420 for a company that today traded unter 200, less than a year later.
>Musk's "secured" bid of $420 for a company that today traded unter 200, less than a year later.
This keeps being repeated ad nauseam but what's the relationship between the 'going private' tweet and today's stock price?
It's been proven that the Saudi asked Musk multiple times to take Tesla private between early 2017 and summer 2018 but other shareholders resisted the attempt (with far less capital at hand than the Musk and the Saudi).
Remember that non-accredited shareholders (like me) would also have been forced to sell at that time, if the company was to be delisted (which wasn't Musk's intention but still does not undermine the "funding secured" tweet one bit).
In fact, today's stock price makes a strong case for the company to be taken private and even a stronger one for the attempt to do it when it was above $350.
> It's been proven that the Saudi asked Musk multiple times to take Tesla private between early 2017 and summer 2018 but other shareholders resisted the attempt (with far less capital at hand than the Musk and the Saudi).
Actually, the record seems to be that Musk was interested in taking Tesla private, and met with the Saudis once or twice, where they allegedly indicated some verbal support for the idea while not agreeing (or even discussing) on things such as at what price. Musk then took the verbal support as a full commitment to buy Tesla at a premium over current prices and gave his infamous "funding secured" tweets.
> Saudi asked Musk multiple times to take Tesla private
Do you have a source for that? All I know is that Saudis had one meeting with him where they discussed simple investment, not going private. Further, this meeting could just be them answering a request from Elon musk, and they wouldn't even be interested.
Personally I think Tesla business would get crushed if someone made really well designed plug-in hybrids. Most trips people take are under 30 miles. No need for a 70kWh, 350kg huge battery. Plug-in hybrids could be operated as full electric during short trips and the internal combustion engine could charge the battery when necessary.
No doubt someone will immediately reply with a question: why would you want to change mechanical power, into electric, store it in battery, then change it into mechanic power again, when you can use ICE directly, or go full electric. I'll tell you why. Hybrids allow you all the same advantages as full electric car for short trips, and they are significantly more fuel efficient than normal ICE cars for long trips. All this in a package that can achieve 400+ miles of range and 5 minute recharge in a package that weights less than half of a comparable battery,it doesn't need to waste energy to be kept warm in winter and cold in summer nor it slowly deteriorates whether you use it or not.
We have to have an honest conversation about disadvantages of battery powered electric cars. Everything that makes those cars cool - great torque of electric motors, mechanical simplicity, user interface can be replicated in a plug-in hybrid(sacrificing some of the simplicity as a hybrid needs either a fuel cell or an ICE to recharge). Then imagine buying a 10 year old Tesla model 3 economy, in a climate with -20deg C winters? Lets say you're left with 60% battery capacity, then you don't want to kill it completely so you don't let it go to complete 0 and you don't charge it to full capacity leaving you with perhaps 40%? Is this enough? To go to work and back, maybe, but what if you want to, or have to travel further once in a while? You can't using that car.
"Just throw in an ICE" is your grand idea? That sounds like the argument folks were making to add a keyboard to iPhones.
Perhaps you weren't aware that Teslas are holding 80%+ of their charges over 300k miles and are expected to double that performance with Maxwell technology?
"Just throw in an ICE" is your grand idea? That sounds like the argument folks were making to add a keyboard to iPhones.
It is not "my grand idea". It is a good idea for the reasons I mentioned in my previous post period. Few actual cars that were and are good plug-in hybrids such as Chevy Volt and Prius Prime are(and in case of Volt were) getting way less publicity than they deserve.
I recently heard Chevy Volt was the first hybrid that didn't just start an internal combustion engine when you pressed on the gas pedal too hard, but it let you drive in 100% electric mode until the internal battery discharged.
Personally I think those cars fail to catch on due to the failure to market them properly.
It already happened, the 2nd generation Volt was an excellent car and it just died. Which is a shame because it's the "35k Model 3" most people should have gotten
We arguably have that in the Prius Prime, or had that in the recently discontinued Chevy Volt.
I think the marketing and design is tough.
So far, there really isn’t an electric car market in North America. There is a Tesla market—they sell well, but basically every other electric vehicle is sold at a substantial loss and in tiny volumes to satisfy various government mandates. In my opinion, people buy a Tesla because they are buying a Tesla, and then people can see you driving a Tesla. Same thing we saw with hybrids. There is a Prius market, because you are obviously driving a hybrid car with brand recognition, and otherwise nobody is buying the Honda Insight or hybrid Camry.
And I think it’s tough to market, quickly, how a plug-in hybrid works. It is not the simplest drivetrain, and trying to explain to people when you have to plug it in and how to get the best electric mileage and yes the engine will run occasionally and yeah that’ll be weird when it kicks in...a pure electric car is just conceptually easier to understand.
here in CA the cost of electricity is such that driving on electricity is roughly equivalent to 45mpg in cost. So, a hybrid like the prius can beat that: 60mpg
I’m in CA and have been driving electric since 2013. Electricity costs 13 cents per kWh, about 3 cents per mile. My last gas car, a Honda Fit, cost about 10 cents per mile (33 mpg @ $3.30). So it’s more like 100 mpg equivalent.
The major flaw (among many) with Musk's self driving taxi Tesla vision is that it treats cities, the entities that control the roads the taxis would drive on, as passive entities that wouldn't make any peep of objection.
In reality cities have plans on the books to shift away from car oriented design whether it be electric or ICE. Additionally, cities now understand that ride hailing competes with and erodes the quality of public transit, with some accordingly capping ride hailing licenses.
Cities are not going to want the surge of congestion that self driving Tesla taxis would bring.
Autonomous cars are part of the shift from cars, because they don't need street parking and can "loiter" offsite.
Urban plans for alternate modes of transit (scooters, bikes, etc) will massively benefit from street parking elimination. It's where the lanes can be safely placed.
Autonomous driving will revolutionize public transit. Uber isn't wrong about that, it's just that they are too evil to be trusted with it from a monopoly standpoint. Autonomous can solve last mile problems and centralize transit stations so that you don't really need bussing, and can concentrate on fast high capacity trains with "coarse grained" stations.
It is still a model of having one 15'x6' vehicle carrying 1 person to 1 destination, which leads to congestion no matter how well the cars are timing off of each other. Either these things get really really tiny (self driving moped?) or we cram more people in one of them (self driving bus?) because at the end of the day its a physics problem of there being not enough space on the road for everyone in a city to be transported one at a time, 20' by 8' apart.
The bio diesel relies on a net investment in calories to achieve its green fuel. Most of the investment comes from petroleum products used as fertilizer. (All be it some time ago, I remember it being 4:1 calories invested:harvested for corn).
Battery tech is fine today if companies like Tesla and other EV manufacturers weren't so against open standards. You should be able to pull out your battery and swap it with a full one at a gas station, just like you do with a propane tank (and the gogoro electric scooter).
You don't see a car manufacturer making a proprietary gas inlet that only takes a proprietary pump, but that's exactly what EVs are currently doing because they can squeeze more money this way at the cost of how good the tech could be for people.
Carrying a battery to a car / truck / tractor way out in the middle of nowhere ( US has those places ) would be extremely difficult. Regardless of their availability.
Take a look at the gogoro battery system, you swap your empties for a fresh battery and you go off on your way while your empty ones stay in the rack charging, waiting for the next scooter user.
Petroleum certainly has momentum, however there's no reason to suggest EV and ICE won't co-exist for a couple of decades. I doubt bio-fuel will ever be as cheap as oil, sadly, there are simply too many sources with very easy to extract oil (Venezuela, Saudi Arabia, etc.)
I think there’s a world that a lot of tech people don’t consider. Let’s say Tesla rolls out their self driving taxi service. And let’s say it’s decent. Not great. It kills people infrequently and Tesla manages to argue its way out of liability...
Wouldn’t that be a pretty good business? It has some severe ethical costs, but... a lot of people are ok with that. What should be doesn’t really effect the market’s reaction to things or even consumer preferences.
It would be an amazing business, if in fact it was tesla owners pressing a button on cars they own and letting it make some money for them as a robotaxi. The fact that even that declaration that Musk sees that happening in 2020 didn't move the needle for Tesla stock shows how little the market believes it's actually coming imo.
I agree that was the motivation for the timing of the event. I do think he is being serious though. His track record is reasonably good in that he does intend to accomplish what he says he wants to do with zero to moderate delays. By my mental model, I would expect him to deliver this by 2021.
The way Tesla tends to defend Autopilot-caused crashes makes me think that if someone did take over their servers and then sent an update to a specific car to kill someone, Tesla would hide this from the public.
The U.S. needs better data breach disclosure laws before self-driving cars become mainstream.
Automakers will say and probably rightfully so, it was never about the software it was about the nuts and bolts of making large amounts of vehicles and the act of actually selling them to dealerships. It probably turns out automakers do it the way they do it because it is the best way to make money and not because they are stupid.
Nuts and bolts are becoming commodotized by the day, as they do in any other industry frankly. EVs are a huge accelerator of this process because they are a fraction of the complexity of an IC drivetrain.
The supply chains in automotive are also becoming more dominant and most of the stuff in your car is sourced from Aptive or Bosch or Continental or whoever.
So it's not just a commodity yet, but it will be - especially towards the lower third echelon of the market - in a decade or two (a decade is not long in auto industry terms; it's merely two product cycles).
Bosch was founded in 1897. Here Japanese cars made it big in the eighties. Korean cars in the last ten years. A local factory has produced Saab (now defunct), Porsche and now Mercedes cars, spanning about five decades? Maybe some consolidation but also I think there's a high amount of constant flux.
The ICE drivetrain is arguably the least complex part of a modern car. Converting to electric reduces some complexity, adds some, but fundamentally it's still a car. Nuts, bolts, ECUs, brakes, etc -- all of these have been commoditized for decades. What the manufacturer is bringing to the table is the structure. From the way it looks to the way it crumples in a crash, that is arguably the single most critical element of a car and I don't see any commoditization going on there. And those are two of the biggest decision makers for buyers.
How is a dealership network more intrinsic to success than physical bookstores? Auto dealers are demonstrably not the best way to make money for a manufacturer. This is why dealers use legal means to prevent companies from selling cars without dealers.
If you need laws to protect your business plan, your business plan is probably not the best way to make money for anyone but you.
Interesting... The guy is landing rockets on Earth. He said, it planned it did it.
So on one hand - do you listen to a bunch of wallstreet know it all short sellers pumping out negative news and counting parking lots and shipping containers.
Or do you listen to the Elon case - that it actually may be crazy to buy a car that that is not a Tesla...
One thing is certain, A car like a boat has been a money loosing investment until today, and consumers have been marketed to oblivion and convinced that, that is O.K.
IF THAT CHANGES, even slightly, well, then... Its GG (good game). because that changes everything, and even the die hard truck lovers will be tempted to make a few bucks on the side.
Landing a rocket on earth is technically a lot easier that making a self-driving car. The former is in a complex but reasonably well understood area of physics plus some advance manufacturing. The latter is cutting edge AI sitting in the middle of a legislative/political/ethical quagmire.
Ashlee Vance's biography actually explains this really well, how one paper SpaceX looks far more difficult to pull off than Tesla but how in reality it's the opposite.
And no money to do it. Did you see the results of their last attempted funding raise?
Once again acting like something likely far fetched is right around the corner, because Elon said so. When the money for nothing dries up, we will see what these businesses look like.
I'm a huge skeptic when it comes to Tesla, but JLG raises a good point here. Apple had all sorts of problems in 2007, so many that established phone manufacturers largely brushed them off. A focus on software rewrote the mobile phone industry, and every day I get into my car and watch the onboard computer struggle to handle bluetooth properly.
I need to look beyond the obvious inanity of Musk and remember this shift: software will eat this part of the world too.