I’ve said this a thousand times. The difference between SpaceX and Tesla is Gwen Shotwell. She is an excellent check on Elon Musk.
You need both. It’s the difference between a company with exceptional engineering talent that also runs well, and a company with exceptional engineering talent that doesn’t run well.
I’m really happy with my Model 3 and I continue to wish them success. But these kinds of moves are just pure uncontrolled Elon Musk as-the-engineer. He’s as bad at estimating as I am. I knew the state of machine learning when I decided to just get EAP and not FSD. I’m not celebrating their miss, I’m just saying all this was avoidable.
How on earth are you seriously claiming that Tesla "doesn't run well"?
They've gone from nothing, to a major competitive auto manufacturer, in 10 years. Not only that, they've built out the service and infrastructure to support themselves as a major auto manufacturer, as a side project.
If it weren't for their ridiculous stock price nobody would call them a major competitive auto manufacturer. Their sales are not major and if it weren't for their ability to sell credits they would not be remotely competitive.
Sure, except their stock prices reflect the brand itself. Drive in front of a 5 year old and they know "mommy thats the car that can drive itself!" What car brand has ever has that type of reach? If you have a sportcar, you get a generic ooo look a ferrari/lamborgini. Something most will never reach. But a tesla will be their next car.
The only other brand that has this reach would be Apple. Every single communications player have been gunning for them since iPhone 1. No one has even come close.
I don't know about that. My plain white Model Y gets a LOT of attention from people of all ages, including the under 10 crowd. I've driven other, more flashy cars, and have never gotten as many "Oh look!" reactions in my neighborhood. It constantly surprises me, in a good way.
Im currently sitting at one of the Tesla Superchargers in Phoenix. It’s full of cars. There is a CHAdeMO charger across the lot and I’ve never seen somebody use it.
Have any of Tesla’s competitors (other than the Ford ~~model t~~ mustang) even shipped yet?
“Nikola” appears to have been a blatant scam. Rivian hasn’t shipped. Lucid hasn’t shipped.
Meanwhile Tesla has a month long waitlist (at least), a retail store network, charger network, service network, and is on the 4th generation of actual shipped products.
That’s sounds exactly like spacex to me. The competitors (if you could even call them that yet) are trying their best to copy SpaceX but so far even Jeff Bezos seemingly infinite wealth hasn’t been able to do it.
Also depends on how/what you are measuring. The Zoe is almost $25k cheaper than the base model standard range Model 3 (at least here in Sweden), which explains why it sells so well. I don't think anyone would claim it is a better car than the Tesla, or ahead technically in any way, and you could argue that they aren't in really competing since one is so much smaller and cheaper than the other.
But yea, in Europe a lot of manufacturers sells decent fully electric cars.
Doesn't Phoenix have Waymo full self driving taxis that literally anyone can take upon landing at the airport?
Tesla might be (way) ahead of auto-manufacturers when it comes to churning out individually-owned electric cars and providing a great experience for their customers, but I don't think Tesla is ahead at all when it comes to self-driving.
This is how I feel about the Steve Jobs and Jony Ive dynamic. The two of them together was a great match, but when Jobs died Ive was left to his own devices and kinda just ran with his own worst impulses ("make everything as thin as humanly possible", etc).
While this is certainly true for falcon 9 I wonder if Shotwell is deliberately letting Elon go unrestrained on Starship. Maybe SpaceX winning the Artemis contract changes things but until now Starship was an area where they were free to make whatever claims they want and to fail as much as they like. Elon the hype man could run wild.
People need to stop with this idiotic idea that Shotwell is some firefighter constantly holding Musk back. This is a meme at this point.
Shotwell and Musk are working together very closely and have for 10+ years. Shotwell is often just as aggressive as Musk. On Starship specifically Shotwell has been just as agressive as Musk.
And Elon is not a 'hype man' he is the Chief Engineer no matter how much people want to ignore that and call him a marketing guy or business guy or a hype man.
NASA was clearly impressed with what SpaceX is doing with Starship and that was during the time when 'hype man' was running it (according to you). Accept of course we know that Shotwell is very much involved.
Starship is developed the way it is because it is the most of how ambitious it both for operations and production. And those criticizing have clearly never developed a fully reusable rocket the size of a sky scraper so their criticism is mostly worthless.
I will look for that YouTube video that Shotwell herself describes their roles.
To (badly) paraphrase, Elon is definitely the visionary and chief engineer. He’s the one who says things like “we can go to Mars in 7 years” and everyone gasps and says he’s out of his mind. Then he lays out “no, we can do x,y,z and make it there, correct?” And then people kind of nod. And then she takes his ridiculously ambitious vision and has to figure out how they run a company with that goal. It really is that sort of dynamic. I am saying that with her, they are a terrifyingly effective duo. He needs that in Tesla.
I own Tesla stock. I want them to continue to wildly succeed. It’s for that reason I want Elon to have a true peer there too.
Was debating whether I should get FSD or not on the Model 3. Eventually decided not to despite the egging ons of some people at a Tesla forum I frequent to pull the trigger.
From everything I could glean, the technology does not seem mature yet, and shock!!!, despite the disbelief of some robotaxi/FSD evangelists, I actually like driving. Ok, so I don't like stop and go barely moving traffic, but I don't think FSD can handle that today or anytime soon anyways.
I totally agree with you about Tesla needing a Gwynne-like figure – I’ve said it myself before on this forum.
However, away from the controversy about their approach to and messaging around full self driving, Tesla is otherwise pretty exceptional, if you look at what they’ve achieved. (As detailed by some of the other replies to your post.)
Tesla is one of the most successful companies in American history, for you to say it doesn't run well reveals either unreasonable standards or misplaced priorities.
I’m actually a happy owner of a model 3, and generally a cheerleader for the company. I want companies like Tesla that turn industries upside down to succeed wildly. Let’s start there.
However, Tesla as a company seems to jump from burst to burst of technological wizardry. Build a great car, get stuck in manufacturing hell. Solve the manufacturing problem, you’ve injured a bunch of factory workers. Solve that problem, you’ve now created exceptional battery cooling tech. Solve that problem, but owners can’t get their cars fixed in a timely manner. Solve that problem and then you have suddenly raised solar panels prices 3x. Build amazing solar tiles but then your FSD was always a fad.
It’s a constant thing where they are jumping around from tough problem to tough people to constantly put out fires. Again, they have a world class engineering talent so they solve these problems with exceptional solutions that often leapfrog their competitors. But it feels like they jump around a lot.
It's highly unlikely that cars will be able to replace drivers in general in five years too. I don't think Elon is stupid, and I think he's fully aware of that. He's just getting rich by tricking customers.
I don't get why Tesla is allowed to continue like this. There is two companies I know of that can do basically whatever they want without consequences: Tesla and Apple. Their users believes anything they say.
once people are true believers and that belief has become part of their identity their cognitive biases kick in to ensure no deviation from the belief. We only have to look around at the political arena or the religious arena to see that. No difference in the scientific, financial and business arenas because it's the same human brain involved.
Literally the next line in the email, which the article conveniently omits:
> Please note that Tesla’s development of true autonomous features (SAE Levels 3+) will follow our iterative process (development, validation, early release, etc.) and any such features will not be released to the general public until we have fully validated them and received any required regulatory permits or approvals.
The Beta program is Level 2, this is well known already. It’s intended to collect data amongst volunteers, not as a final end consumer product. The end goal for FSD that most paid for is for Level 5. Not sure why this is confusing. The 10k FSD “add on” is NOT for the Beta they are planning to offer soon, but the final post-beta release, which will be Level 5. The information sent to regulators is for the BETA they are hoping to release in the next month or so. I don’t know why the press in general can’t seem to acknowledge the difference here.
Tesla needs to explain better, but given the intentional mis understanding by journalists lately, I sense they feel silence is better than being taken out of context yet again.
It doesn’t seem at all guaranteed that Tesla can ever achieve Level 5 autonomy, in any timeline, using the hardware deployed into Tesla vehicles today which were sold with “Full Self Driving” (FSD) as the $$$ expensive upgrade option.
I doled out ~$5000 for FSD on a Model 3 Performance bought 2018 which has now had a new computer fitted (at no cost to me) and always considered FSD pretty much a “bet”, which is not IMHO a complete loss since the HW3 upgrade got me more Level 2 features on a 2018 MY vehicle, something which is quite unprecedented in the auto industry.
I suspect Elon may be wrong about LIDAR and for any manufacturer to get to Level 5, it may be required and cannot be replaced entirely by cameras and computer vision. Since it is not present on any Tesla today and cannot be sanely retrofitted, barring what feels like a miracle to me, the “Defects per Million Miles” (e.g. disengagements, and accidents leading to property damage or non-fatal injury) and the “Fatalities per Million Miles” (where one was enough to tank others Autonomous Driving programs) will probably continue too high for public and regulatory acceptance that Tesla has attained Level 5 autonomy with their current vehicles and their current hardware.
Elon and LIDAR reminds me of Steve Jobs and his stance on the the concept of a smaller iPad or bigger iPhone (both which materialized after his death IIRC).
In both cases, both people seem fundamentally and adamantly adversarial for whatever reason.
I think cameras and computer vision are better than LIDAR but the issue will always be the hardware supplied.
The computer for processing the cameras will probably be swapped out a couple of more times before the cameras need to be swapped. The question will be if your 2018 car will get FSD in 2035 by having all of the hardware replaced instead.
Why do you see it as an “intentional misunderstanding” on the part of journalists (and regulators and customers, for that matter) rather than an intentional misstatement, or at least exaggeration, on the part of Tesla?
Literally the next line of this email to the DMV says they will follow an iterative process and get regulatory approval as they release updates with Level 3+ functionality.
This article is essentially disinformation. If the article actually included the relevant source email, it would be obvious to everyone that it is a nothingburger.
Musk explicitly promised L5 since 2016. This is the first honest statement Tesla has made about “FSD” (actually less-than-half self-driving).
When asked about a cross-country autonomous trip;
“[by 2017] all you will need to do is get in and tell your car where to go. If you don’t say anything, the car will look at your calendar and take you there as the assumed destination or just home if nothing is on the calendar.” [1]
He explicitly promised L5 by end of 2020:
“I am still confident we will have the basic functionality for level 5 complete by this year“ [2]
He’s a conman, and the cult of personality is going to end like any other cult.
I recommend following the FSD beta drivers on YouTube. If you’ve at all been following the progress, I think you’d be impressed what is already possible. I’ve see no major corner cases that require hardware changes yet. It’s all software now.
Tesla is just asking for a lawsuit at some point. First "Autopilot" and now "Full Self-Driving"—it's no wonder people keep treating these vehicles as more capable than they are and getting in accidents! You can't tell someone their car has full self driving capabilities and then expect them to listen when you tell them that they also have to be paying attention 100% of the time. The two ideas are antithetical, and people are going to (consciously or unconsciously) believe the one they see pushed the most.
Can we please, please stop with the whole "Autopilot is so misleading thing". Tesla's Autopilot is more sophisticated and feature-rich than any other autopilot (i.e. in an airplane) that you've ever experienced.
The issue isn't how sophisticated or feature rich it is, the problem is the expectations it creates for what you can do. As a sibling has pointed out, the other places where autopilot is currently used are places where, once autopilot is turned on, much less attention is needed to avoid a fatal accident.
For this reason, the connotation of "autopilot" is "the vehicle controls itself with no effort needed on the part of the operator", which is simply untrue for Tesla.
That's a demo of the new "full self driving", not existing Autopilot. I might be on board with calling that functionality "autopilot", but they already used that term up and had to escalate. With that escalation, the marketing is once again out of sync with reality.
i disagree. you don’t get the name autopilot just because you are doing something sophisticated.
autopilot in a plane is a true autopilot as it truly controls the plane. in fact apart from
takeoff and landing most of the flying is done by the autopilot.
is the problem of controlling a plane easier than driving a car? yes. so what?
That is only true for modern autopilot systems in commercial airliners and business jets. But the system has been called "autopilot" since the 70s, when it could barely hold pitch and yaw within a tolerable deviation, with no regard to the engine, altitude, yaw or a dozen of other parameters that still had to be maintained manually. Most plane autopilots are somewhere in between these extremes, and require constant supervision by the pilot in command when flying.
Autopilot on the old planes didn't require full attention of the pilot. They can take the hands off the controls. Fill out forms, run checklists etc. It is like level 3 self driving. Telsa autopilot requires drivers to keep their hands on the controls. I am not aware of any self pilot system that requires 100% of pilots/drivers/captains focus to be on act of controlling of the vehicle/vessel/plane except Tesla.
Most other car manufactures label similar level 2 systems as lane centering and adaptive cruise control. Cruise calls it super cruise in reference to being a better cruise control system. If it was level 3 capable, autopilot would be a descriptive name.
i was thinking about the big commercial airplanes.
it’s probably true that airplanes were also ambitious at first, but the difference is that flying (paradoxically) is way way more forgiving than driving. Also, in flying you have a support system that is built to promote safety as everyone understands that mistakes == death when they happen. When was the last time you thought about the fact that you might die when getting into a car?
Can we get over this trope of "autopilot doesn't mean self driving." Everybody knows what normal people mean by autopilot. And even if they didn't, introducing the new "Full Self-Driving" terminology seems to me to have a pretty unambiguous meaning.
I don't know a single person that thinks planes are entirely flying themselves without pilot intervention?
Not gonna lie you have a very dim view of "normal people". Clearly you do not believe planes are entirely flying themselves, so I'm wondering who these "normal people" are that you are speaking for.
>You can't tell someone their car has full self driving capabilities and then expect them to listen when you tell them that they also have to be paying attention 100% of the time
Not entirely, but they absolutely share a portion of the blame. Their marketing campaign clearly wants people to believe that these vehicles can drive themselves. They're trying to create that belief in people. That makes them accountable when some subset actually do believe what they've been saying.
Deceptive advertising is a crime. If you set out to create a belief that is untrue, you're already criminally liable. If the people who believe what you said end up getting killed...
They have to justify the hefty price tag somehow. If the car drives itself and can actually earn you money by driving other people around, then it might be easier for some people to convince themselves that it isn't that expensive.
I guess it depends on what value you get from a car. A tesla model 3 performance has a faster 0-60 time than a porche (or similar) for ~1/2 the price. That makes it a great deal for a lot of people, regardless of self driving.
It's a bit weird that they're only claiming Level 2 since the requirements for a Level 3 ADAS are rather modest. For instance, manufacturers have certified the feature of a car "self-driving" in a traffic jam by rolling forward at a human-safe speed as sufficient for a Level 3 ADAS, because the driver can then let her attention slip for a few seconds in this limited situation without endangering safety.
It’s doublespeak. To regulators Tesla claims level 2 but for customers they brand it as literally “Full Self Driving Beta.” That way they can avoid any unpleasant regulatory certifications of the actual capability of their system.
“Just weight the steering wheel and it’s basically level 3!” - probably something a Tesla salesperson has said to a prospective buyer
The "levels" beyond 2 are do not add any new driving features, but rather shift the legal responsibility increasingly to the manufacturer.
So Tesla has a choice of no unprotected left turns, HD mapped area only or go full in. And they are choosing the full in route which is very Elon's style. And allow them to roll out beta quality features quickly and iterate.
I agree it's a PR nightmare right now. I hope they will eventually recertify the current cars as L3+, but for the regular user that will only mean they can now officially take eyes off the road and only look when it beeps.
California is already claiming that Tesla cars have too much ADAS to drive without Tesla getting a license to operate them in the state. Tesla is actually claiming to CA regulators that the system is not that good so they can avoid needing a license.
Yeah, I don't understand why Tesla hasn't simply been ordered to stop calling these features Autopilot or FSD, as both are misleading, and explicitly suggest more capability than they have, causing numerous deaths and accidents.
Oh, and anyone who paid extra for these features should be given a refund, as those features were falsely advertised.
Well that is literally where the term comes from and it hasn't really been used in any other context until Tesla's usage, so not really sure where the average person's confusion is coming from.
My point is: I don't think this is true. But if you have a survey which indicates that a broad swath of the population believes Teslas with Autopilot don't need any human intervention or shouldn't have someone in the driver's seat I'd love to see it.
Literally dozens of people in this thread are saying "this is what the average person believes" when none of them believe it themselves. Where are ya'll getting this insight into the average person's thought processes?
I mean, it's worth noting that in most aviation and marine scenarios, collision is relatively unlikely. Leaving a plane on autopilot to go in a straight line for a while is unlikely to be dangerous, but it would be very dangerous in a car.
If that is the general perception, how it got that way doesn't change what people think the word means, or what they'll hear when Tesla names a feature Autopilot.
I'm questioning this oft-cited version of reality. I don't know a single person that thinks autopilot means pilots don't need to be in airplanes, and consequently that autopilot means you don't need drivers in the driver's seat.
Of course people recognize that they can get around Tesla's precautions and have nobody in the driver's seat, just like pilots could all leave the cockpit while the 737 is in flight and go get wasted at the minibar or whatever. But they don't and neither should drivers, and Tesla isn't suggesting that driver's should – in fact they've put technological precautions in place that airplanes with autopilot don't even use to ensure someone is at the stick.
You can also run over pedestrians with [insert car here], but nobody is suggesting you do it. People doing it anyway is not [insert car manufacturer here]'s fault.
It may be misleading, but my argument is that it is not misleading to the point that they deprive the people doing these things of agency, which to me is what matters. Everyone who is disregarding the warnings, weighting their steering wheel, weighting the driver's seat, etc. knows that what they are doing is not how the vehicle is intended to be operated at this time. They are knowingly pushing the limits.
If tomorrow Boeing came out and said "hey, Autopilot on the 787 Dreamliner is sophisticated to the point where the plane can totally fly itself, from takeoff to landing", would it be reasonable for all 787 Dreamliner pilots to just leave the cockpit and get wasted with the flight attendants in the back? I don't think so. So I don't understand why people are holding Tesla to a ridiculous standard that they're not holding any other company to. Teslas can drive themselves from A to B in many situations at this point. It's not false advertising to say that. What would not be OK is if Tesla was saying if you buy this car you don't need to be in the driver's seat right now (not in the future, which is how they're currently selling it).
>I don't understand why Tesla hasn't simply been ordered to stop calling these features Autopilot or FSD
They walked back, or qualified a lot of the initial claims and added mitigations to prevent their 'self-driving' systems from being used as such.
>Oh, and anyone who paid extra for these features should be given a refund, as those features were falsely advertised.
There may be some validity to this. Musk/Tesla certainly made explicit claims that existing cars have the hardware to achieve full autonomous driving via a future software patch.
> They walked back, or qualified a lot of the initial claims
The product names are fraudulent, there's no walking back when they're still called Autopilot (a term which in all other uses essentially does not require operator attention) and Full Self-Driving (which is definitely not what the feature is).
It's fraud, plain and simple. Until the features are renamed, Tesla continues to commit fraud.
An autopilot is a system used to control the trajectory of an aircraft, marine craft or spacecraft without requiring constant manual control by a human operator. Autopilots do not replace human operators. Instead, the autopilot assists the operator's control of the vehicle, allowing the operator to focus on broader aspects of operations (for example, monitoring the trajectory, weather and on-board systems)
Autopilots don't stop crashes, or change lanes. So Tesla's Navigate on Autopilot already does more than a standard Autopilot
The issue with this comparison is that in scenarios pilots or ship captains use autopilot, collision is extremely improbable compared to that of a car.
Most laypeople who encounter autopilot do so on small sailboats near the coastline where a collision is extremely likely if you just turn on the autopilot and go chill in the cabin.
>It's fraud, plain and simple. Until the features are renamed, Tesla continues to commit fraud.
That's for courts to decide, but I don't think so, at least not to the level of hostility that you seem to hold.
This kind of reasoning is typical online where deductive logic based on strong assumptions made by the posters, is then used to reach equally strong conclusion (in this case, that a marketing term implies outright fraud). The reality is that regulatory, judicial and cultural boundaries have much more slack than you give them credit for. And that's a good thing.
It's not just the names that are the problem, it's also the fact that they are trying to do a little more than other driver assist systems despite being very unreliable.
IMO there are some features that shouldn't exist until full L4 so that people aren't tempted to let the car do the driving when it can't
I am reasonably okay with a driver assist system attempting to do more. But I'd agree they need to be better about both knowing what they can't do well yet, and that they need to be much better about detecting driver presence and attention, so that the car can't be used in manners they don't intend.
There is a few companies releasing PR announcing they’ve achieved level 3 and 4 but I don’t feel like I’ve seen any on the street (or maybe they are hard to spot?).
I’ve been in a few Tesla and they drive themselves quite well (I’m not an expert in self driving but I’m a good driver)
I’m genuinely curious, which other manufacturer has a similar level of self driving as Tesla, can deliver an electric vehicle within a week on my doorstep within the same price range?
I think that's a dangerous way to evaluate self-driving systems. With a little practice, any 10-year-old can "drive quite well" under limited conditions. But what makes for a good driver is not the 99% of the time dealing with optimum conditions, but the how they handle the long tail of increasingly challenging/dangerous situations.
I suspect a lot of companies could deliver a Tesla-equivalent system. They're just opposed to killing people in the process. See, for example, this thread from a host of The Autonocast contrasting Tesla's approach with others: https://twitter.com/Tweetermeyer/status/1383778011766398983
Level 4 doesn't actually include the long-tail though, that's Level 5. It's really a terrible benchmark system. Tesla could be level 4 tomorrow if they said (like Waymo) "we're gonna enable FSD only in the greater Phoenix area", where the roads are well-maintained, they could map them all out, etc.
The issue here is that Tesla actually wants a generalized system that will work anywhere a human driver can drive, and does not seem interested in geofencing at all until they've reached that (level 5) point.
With a geofenced area you can hard-code behaviors in any area where FSD struggles. I see no reason to believe they couldn't do this – FSD can already drive quite a few places with zero assistance, and most of the hiccups seem to be around edge-cases, which are easily pre-programmed.
When I say "hardcoding" I don't actually mean hardcoding. I mean forcing the ML system to keep iterating on specific problems that it did poorly at, with human supervision.
e.g. the car did the wrong thing at the intersection of 9th and Elm? Give it 100 similar situations (at 9th and Elm), let it generate lots of possible behaviors, and then have humans tell it which are correct. Then whenever a car is at 9th and Elm, it uses that training as a priority. For a relatively small geofenced area that is a reasonable approach and should work well. Unfortunately it doesn't scale to the whole country/world.
Why doesn't it scale? That's basically how humans do it. We learn general-case rules and then we learn how to handle specific situations specifically. You can observe that just by driving through an unfamiliar city with a local. They'll feed you a stream of tricks for optimizing the drive.
I find this strategy very close to the lean startup model (MVP). Go to market with something you believe is better than what’s out there early, have rapid feedback and iterate.
Feels like a risky bet for Waymo and others to reach level 4 or 5 only in a limited environment, you might end up with so many untested assumptions
Yes, to me this is also another reason Tesla decided to not go with LIDAR (besides cost): if your self-driving car uses a different sensing system than a human, there are no guarantees that this car will be able to drive anywhere a human can drive. Meanwhile, if your system is relying on entirely the same thing humans rely on, the likelihood seems higher you'll end up with a more generalized solution, since all roads / etc on earth are built around vision, not LIDAR.
Most humans do not, indeed, work well in those 1% conditions. If they wanted to avoid damage to their cars, other cars, or human death/injury, they would pull over or choose not to drive when the weather calls for bad driving conditions, or when drunk. I imagine Tesla and the rest will be on-par with humans in these driving conditions within 5 years, but it’s definitely fair to ask that they be 10x as safe as a regular driver which might take 15 or 25 years.
I don’t believe lots of companies can deliver what Tesla offers, or if so, have the ethics of not delivering for the sake of saving lives. If anything, history showed that most car companies ethics is very questionable when competition is fierce.
Unless I misinterpreted the recent Tesla safety report, the Tesla FSD seems statistically safer than human drivers (please correct me if I’m wrong).
The statistic is highly misleading. Yes if I only turn on a feature in the most safe situations (e.g. on the highway) it's going to get better safety statistics than general driving which includes non-safe situations.
I’m not sure a random portlander with no expertise on this tech who makes their money selling self published ebooks on why you should be afraid of Tesla is exactly the right person to include in this conversation.
What others can deliver totally irrelevant. Can I sell my drug as cancer curing (even if in clinical trail it didn’t show much success) just because there’s no other cancer curing drug?
I haven’t faced a situation where comfort drug was sold as a cure. The few experiences I had with the matter, the outcome and justification was clearly stated by the nurse or doctor
Tesla’s FSD is exactly that case. It’s glorified level 2 system (like a comfort drug, aka, increasing comfort while driving), that’s cleverly advertised as a level 5 system (cure, aka removing driving responsibility from you).
It seems to me that you’re arguing that advertising is misleading. In that case I totally agree with you and I would add that all advertising I see is misleading. I’ve personally learned to not trust marketing at face value and try to rely on facts as much as possible
If attributing my argumentation to a speculative agenda is your way of leaving the discussion without losing face it’s kinda pathetic.
The world we live in is full of improvement toward a illusive ultimate solution.
Marketing messages seem to be the problem here but I’ve experienced self driving a few times in heavy traffic (including at night a year ago) and it was really reliable and felt very real to me.
Discussing the philosophy of self driving technology using analogies was your attempt of making your point but it will always fall short of discussing the topic at hand.
FSD is not full self driving by many standards but it is doing a lot and saving lives too (like collision avoidance). To go with your analogy it is not only comfort but provides real benefit to an existing problem
As far as I understand car manufacturing is a regulated industry.
I honestly do not understand why cannot we have a regulation that if both hands are not on a steering wheel for more than 10 seconds, then car should gracefully stop?
There are technologies in place which measure driver's blinking and tell of driver is tired. Why cannot we detect if driver is looking at the road at all. And if driver does not look at the road for more than 3 seconds, then car should gracefully stop?
Maybe I am wrong, but I believe it could make driving much safer, because these stops would be extremely annoying. We already have enforced stop as part of car anti-theft systems and rental car control and it is considered safe enough, so why the heck not?
A car coming to a halt gently in the middle of a highway is extremely likely to precipitate a multi-car crash.
Heck: this is the type of situation which most self-drive systems currently have a very hard time dealing with - a stationary vehicle where they don't expect it.
“Gracefully” could mean put the hazards on, gradually reduce speed, and move to the shoulder. It doesn’t have to mean “come to a halt in the middle of I-95”.
As another layman this seems like a good idea for the next mandatory safety feature now that driver assist is commoditized - lane keep, adaptive cruise and attention monitoring is enough to implement it.
Tesla cheaped out on the touchscreens to save a few bucks. https://www.thedrive.com/tech/27989/teslas-screen-saga-shows... They have been extremely resistant to adding any hardware even after other companies have used capacitive sensors on the steering wheel and/or eye trackers.
The cars by themselves are really good. But the problem for musk is without self driving the electric is already commoditized and so the value proposition will be a race to the bottom on prices. So he is hyping FSD and selling cars.
* citation needed. Tesla level quality control and wonky reliability is what personally turns me off from paying top dollar to be both 'owner' and beta tester.
No defence of Tesla here at all, but if you read any large car forum regularly, you will see that delivery of new cars with significant problems (e.g. paintwork, interior trim, even major mechanical parts) is not unheard-of from many major manufacturers. And major manufacturers have also had engineering issues with whole generations of cars.
Throw back to how Tesla promoted Autopilot as recently as 2019:
>Full Self-Driving Hardware on All Cars
>All Tesla vehicles produced in our factory, including Model 3, have the hardware needed for full self-driving capability at a safety level substantially greater than that of a human driver.
You're definitely right that I got Tesla's described goal wrong.
That said, I don't think self-driving system has come close to "better than human" yet, so my point still stands.
Some may do better in limited circumstances, but nothing even comes close to handling the range of weather and weird roadway conditions that humans can.
And please don’t reply with “well he didn’t literally say, ‘we have the software.’”
It is extremely clear that Tesla is either intentionally lying about their self-driving capability or have lied to themselves to the extent that they actually believe their own bullshit.
So still 8 months left. Wonder what they mean by "basic functionality"
The current FSD beta test is the best that's available on a car you can buy. However the instructions say "be ready to take over at any time", so it's obviously level 2 currently
Probably whatever definition they want to make up in hindsight so Musk fanboys will be able to twist things around to make it look he was technically right all along.
My point is that the eventual level 5 system will also be called FSD. And they sell FSD but they clearly state it in the Beta and you don't actually get Level 5 autonomy.
When you buy it they are very clear on what features are included and what features are not. And they say what features will be added soon. Promising more updates as features come online.
I have very little experience with the industry, but I think car manufacturers are playing a game of chicken to see who will sell a Level 3-certified car first because, while the law supposedly clearly lays out the legal details, in practice there is no prior work for judges and juries to work off of. It's brand new territory, and no one wants to be the first one in.
Tesla has discovered they can bypass this because "Full Self Driving Beta" is not a regulated term and it sells like hot-cakes to consumers. I am sure other companies are/will be following suit. No one wants to stamp an official "LEVEL 3/4/5 CERTIFIED" sticker on the official sales sheet.
I’m surprised every time I see anyone believe any claims made by Elon Musk at this point. He’s an excellent marketer with a clear business strategy - promise an exciting new technology (“Level 5 by 2018” [1], “wait actually 2021” [2]), take money for “pre-orders”. No need to deliver. Same idea with the 1 million robotaxis and a roadster by 2020, and the Cybertruck by 2021 [1]. I’m just surprised so few people have caught on. Maybe CEO of one of the largest tech companies in the world really is just “bad at estimating”, but if it’s a marketing/fundraising technique, it seems to be working. Not for the customers of course. Bummer for everyone who pre-ordered fairy dust.
LOL. Well Duh! Most ML etc. is criminally and delusionally overhyped. But hey, Silicon Valley baby! Am-i-rite!
It doesn't take a rocket scientist to see that we are 10-30 years away from level 5! Level 2 is barely possible and only half-assed if you are in an ideal environment like a "planned city in AZ". Absolutely NO existing self-driving tech could ever handle:
* Snow
* Heavy rain
* Country Roads
* Dirt Roads
* etc.
The problem with reality distortion fields is they last a short amount of time. Tesla’s is already losing power. You have to deliver regardless of the hype, and frankly, Tsla hasn’t.
Elon’s gamesmanship to maintain access to capital so he can lengthen the runaway is a tactic, but again, you have to deliver and this can only go on for so long.
> You have to deliver regardless of the hype, and frankly, Tsla hasn’t.
Based on what exactly? They have fast growing revenue and growing operation income, good margin, huge new investments and an exiting product product pipeline with 100000s of per-orders.
What exactly outside of getting FSD to Level 5 have they failed to do?
> Elon’s gamesmanship to maintain access to capital so he can lengthen the runaway is a tactic, but again, you have to deliver and this can only go on for so long.
They have 20+ billion of cash and the have lots of free cash flow as well.
Your statement seems totally divorced from reality of the balance sheet.
Most of them aren't even intending to ever sell them anyway. You'll hail a Waymo on an as-needed per-trip basis; you won't own one. Owning one and then leaving it unused the vast majority of the time would be incredibly wasteful.
If they just renamed it to “eventual full self drive” then all the people who refuse to read the fine print or just aren’t getting the nuance of it would be on the same page with the people that do.
I agree the responsibility for clarity is on their marketing to fix, but I’m also saying that if you read the fine print on the site while configuring your purchase it’s pretty clear what FSD is and isn’t.
It is impossible with ML alone to get to Level 3. You have to include a conceptual computing system that can understand what things are, how they usually behave and what they were just doing, so that the system can predict what they might do next and react accordingly, as we humans do instantaneously.
Tesla is being squeezed on the 2 things its brand marketing is built on: EV and self-driving. I think it's going to be hard for them to compete as the major European and Japanese brands get into these segments with all their weight, with Chinese manufacturers just around the corner.
Lex Fridman had Jim Keller on his podcast and they talked about his time at Tesla and his work on auto pilot and autonomous driving in general. Keller was so cavalier about the solvability of self driving. I was wondering how much of Elon’s boundless optimism and reality distortion field had rubbed off on him. His basic point was that with enough high quality data the problem can be solved, but never really addressed the algorithmic complexity of driving. I used to be very bullish about it myself, but after paying more attention to my own driving and how many edge cases and corner cases there are, I am much less convinced.
Sorry, I've been over Tesla for a while now. Huge respect for the engineers, but their hypeman isn't doing them any favours at this stage in their lifecycle.
Massive respect to the company for altering and accelerating the auto markets move to EV's but I wouldn't buy a Tesla just on brand alone.
The Ioniq 5 is currently the top of my list and if I was in the market for a model S or X, I'd instead be looking at Lucid Air. I'm glad realistic alternatives are finally getting there.
Not interested in FSD at all. I won't be using my car as a Robotaxi and I actually enjoy driving. Lane assist is enough for me.
I don't understand why people think Lucid is going to be so great when they aren't even mass-producing yet.
It's easy to forget that building a concept car or small quantities is dramatically easier than mass producing something that's affordable to the masses.
I own a model 3 and I certainly wasn't dumb enough to pay the 10K for the FSD. That being said I'm not a fan of the other electric cars because the charging infrastructure is garbage. Whenever I stop at a non Tesla charging station I have a very high probability of having to try two or three different plugs before I can get remotely decent current. Electrify America is the network VW was forced to fund after the diesel emission scandal and it's pretty bad. ChargePoint is the other one and they're not very good either.
Definitely don't think the build quality on Tesla's is what it should be and I look forward to the factory in Texas coming out to solve that because Fremont is not going to cut it.
That being said these other cars might be better for the price but good luck charging them.
New tooling, new technology and the ability to design it all out from scratch. The Fremont factory was a retrofit of a Toyota factory iirc. It also has been the workhorse for most Tesla cars so it's hard to do any extended updates to it while they didnt have other factories to pick up slack.
Not that I expect the Austin factory to solve all their fit and finish issues, but I am delaying my Model Y purchase till it comes online.
One possibility: they have said on several occasions that it’s very difficult to change how a factory is working when it’s working flat out. And of course, with the huge pressure on Tesla to deliver the numbers, they can’t afford to slow down or have outages for days or weeks while they change the production line.
As a result, meaningful changes or innovations seem to be limited to the design/opening of new factories.
> I don't understand why people think Lucid is going to be so great when they aren't even mass-producing yet.
I assume it’s because their CEO/CTO was the chief engineer for the Tesla Model S. So, not only has he done it before, he presumably has some lessons learned on how to do things better the second time around. Agreed, that’s no guarantee though.
>> I don't understand why people think Lucid is going to be so great when they aren't even mass-producing yet.
Change the name in that statement from Lucid to Tesla and go back in time. People gave Tesla a chance at the beginning. And that's why you're driving a model 3 today.
I'm not in the market for a Lucid or a Tesla, but I think the Air is far more beautiful than the S, and with decent specs, and I would definitely choose the Air.
But that's an early adopter thing, I guess. Luckily there are enough early adopters around so you don't have to worry about the mass-producing issue.
I'm waiting for the VW.ID Buzz. Classic bus in modern & electric, due September 2022. Just seems like a fantasy come true. Although I'm afraid the final version is unlikely to look like the awesome concept car...
Ioniq looks good but doesn't seem very roomy in the back, so kids+dog+luggage won't fit.
Yep I don’t get why they always build these fantastic concept cars but always dumb down the design for the release.
Interestingly in some podcast Franz von Holzhausen (the guy responsible for Teslas design) mentioned that he was tired of always designing concept cars and never releasing them before he went to work at Tesla.
> Yep I don’t get why they always build these fantastic concept cars but always dumb down the design for the release.
Probably because 90% of a concept car is one-off (and that’s for the rare cases where it actually is a whole car, and not a shell), and re-tooling to create it (not to mention addding all the usual safety features) would make the mass production car cost hundreds of thousands.
So why build the concept car in the first place? As a consumer seeing them at this point is just aggravating. It's just setting me up for disappointment. What's the point in setting your potential customers up for disappointment?
> So why build the concept car in the first place?
Because it builds media attention to the potential related production development; there’s usually particular features that are envisioned being for future production.
> As a consumer seeing them at this point is just aggravating
Consumers aren’t really the target audience, their exposure is more of a side-effect.
To hype the shit out of it. If the concept car looks like crap, noone will be interested in the real deal... welcome to modern consumerism. The best part is companies do it because otherwise other companies will.
The ID buzz has got me excited too, but the video on the VW site [0] does the exact same thing we're knocking Tesla for.. "autonomous driving" displayed as the driver leans back, puts his hands behind his head and the steering wheel folds away!
No wonder those in the market are confused as to what these cars are actually capable of.
Yeah, fair points - I've not see the ioniq 5 live yet so will have to wait and see. I thought they'd nailed it as they had a longer wheelbase than normal. We'll see.
VW.ID Buzz looks pretty bad@ss, to be honest. But if we're going down that route, I'd be sending up some prayers for a VW.ID Beetle. I can dream.
I think the buzzkill (apologies for the pun) on the ID.Buzz is the price. The classic VW Bus was quite affordable - basic ones cost the equivalent of today’s $25k way back when. Electrek reported an expected price of 40-60k euros for the ID.Buzz. So yeah, while it’ll look like the old Bus it won’t be as affordable for average people.
Same. We used to have 2 cars, we ditched one early on in the pandemic and haven't missed it. Now we just have a small Toyota Aygo which barely fits a stroller for the kid.
I'm trying to postpone the need to buy a car as long as I can to have more electric options.
We just bought an electric cargo bike with a kids seat and that has been great too so far.
Electric bikes and cargo bikes are seriously under appreciated. They make life so much better in moderate climates. With a climate like the Bay Area, it's such a shame to see so much infrastructure devoted to deadly cars, and so little to making biking accessible to more people. Ebikes are half the equation, protected lanes so cars can't endanger your life are the other part.
Another interesting one is the Ioniq 5 sister model the Kia EV6. I find that both Hyundai and Kia are really doing interesting design, very different looking to everyone else. And the 800V operation is also quite interesting.
Does lane assist work at night when its raining or snowing ? I dont think I have ever needed lane assist during daytime however I find driving at night in snow or rain to be very stressful.
Back before GPS became common, I had an idea for a navigation aid that would also be useful for things like lane assist and would not be affected much by weather. I have no idea if this would be feasible.
The idea is that when paving a road, mix a whole lot of passive RFID tags into the paving material. By a whole lot, I mean that every square meter of road should have a dozen or two tags. The tag IDs do not have to recorded in advance, and there is no need for any systematic ID assignment. Just order a bazillion tags with random IDs and dump 'em into the road.
Then afterwards you send vehicles over the road that have readers than can gather data to make a database of what tags are in what lanes of what roads, and where. The where part just needs to be the distance along the road from some reference point.
That data gathering can be done by paying commercial vehicles to install and use the readers.
The idea was to then make the data available to the general public. You could then make car navigation systems (remember, this was pre-GPS) that could read the tags in the road, consult the database, and figure out that they are on road X and they are Y meters from the reference point of that road.
You could also continue to pay commercial vehicles to gather and report data. By comparing that data with previous data, you might be able to detect damage to the road. For example, if a pothole develops, that should lead to a loss of some tags, which should show up in the data.
I find lane assist and adaptive cruise control incredibly nice if you frequently have two hour driving trips on the interstate. I can focus on watching the environment and cars around me, rather than spending a lot of effort on lane-keeping and making sure I'm not going faster than the car in front of me.
Sure, it doesn't work in heavy snow, but I also wouldn't trust it in heavy snow either. I appreciate that my car doesn't provide false confidence in dangerous conditions.
My other vehicle lacks any of the modern radar features, and the one time I did the same drive with that car was... surprisingly more stressful.
My Kia will follow the car ahead of it down to a complete stop and resume when the car ahead of it moves, I believe. Though I rarely enable it anywhere but the highway.
(Kia understates the capabilities of their ADAS a lot, probably so people don't abuse it, but it will pretty much drive itself as long as your hands rest on the wheel.)
Adaptive cruise control + lane assist can keep the car in your lane at highway speeds without input in specifically clear weather and well-marked roads and as long as the road doesn't curve too much. Lane assist doesn't work when the lane markers are obscured, so it fails specifically during the times you find lane keeping stressful.
Having just driven a brand new VW Passat that had a lane "assist" keeping that tried to jerk the wheel without so much as a beep multiple times, this sounds like the perfect way to describe it.
Tesla M3 lane assist works well at high speeds with heavy rain (dark conditions but not night iirc). It felt like it was better in recognising the lanes than me, maybe because the front camera is mounted higher than my eyes :D
Given that I mainly drive in the city I don't care much about lane detection but I do care about automatic break system and similar.
Some which supposedly Tesla had the best of.
Until people ran tests on it and realized it's much worse then some much less hi-tech solutions. (And it wasn't just one group which ran similar tests and came to the same result).
So I'm always very careful when Tesla (or other Car manufacturers) promise that "their tech works the best"/"very good" until I have seen some reasonable independent tests.
Lane assist during a snow storm would be a game changer. The number of times you end up driving in poor conditions in the northland is a pain. The 2008 Kia Rio had the best window for driving in snow. The A-frame window that looked directly at the center line was brilliant.
There are certain roads where I live that are normally two lanes in each direction, but in the winter there's not enough traffic to need two lanes. When they are heavily snow covered, they become one lane in each direction, straddling the painted lines. This is unofficial and probably illegal, but clearly safer, especially when snow from the plows starts piling up in the outer lanes.
I wonder what self driving cars will make of situations like this?
I wouldn’t trust any driving automation to handle a situation that I couldn’t handle manually, especially if it relies only on sensors equivalent to what I see.
These automations are extremely useful to relieve long and easy drives. Not to handle difficult conditions better than a concentrated driver.
The reason to buy a Tesla, IMO, is range and charging station availability.
The software/FSD capabilities aren't good reasons to get one. Further, I think their current price tag of 10k is absolutely ridiculous.
That said, if you are in a nation with good EV charging stations, plan on just using the car as a commuter, or if the buildout of current charging stations are good enough then I don't see a reason to pick a tesla over anything else on the market.
I say all this as a current owner of a 2018 model 3 with FSD.
Do you not have any S&P index funds or anything like that? Because Tesla is included in a lot of mutual funds these days. Just wanted to point that out since this disclaimer seems unlikely unless you don’t have stocks at all, or only buy individual stocks.
The whole point of an index fund is to make those details irrelevant. One is putting trust in the individuals deciding what goes in the index to select quality securities, not vetting each individual one in the index.
> When will the Elon fanatics stop listening to his bullsh*t?
When someone else will sell us a car.
I literally just ordered a model Y, with the $10k FSD package. Would I have bought something else if I could get it? Maybe. But this is here, and I can get it. Look, the product as it exists today is real and has value. Maybe you're not exited by it, but I am. And I can buy it. And I just did.
I'm old enough to remember when "make something people want" wasn't considered "bullshit" on this site.
Again, I submit myself as an existence proof to the contrary.
The FSD package as it exists gets you autonomy on highways already. That alone was enough for me to buy it. The beta looks pretty acceptable to me too, and I look forward to that feature rolling out.
Exactly what am I not getting? This really seems like a strawman kind of thing. You're inventing a product and a market that don't correspond to real cars or consumers, and then arguing about what those consumers "really" wanted or what the car was "supposed" to do, and not what they actually want or what the hardware actually does.
I don't see how that's helpful to anyone other than an investor short on TSLA. The way to prove that this is a garbage product, again, is to make a better one.
I hate to break this to you, but you've been Musked.
> The FSD package as it exists gets you autonomy on highways already.
This is incorrect. It's a level 2 driver-assist feature no matter what road or surface you're driving on. This means keeping both hands on the wheel and paying full attention to driving 100% of the time.
With level 2 driver assistance, you must always have your hands on the wheel and be in control of the vehicle, paying attention 100% of the time. Autonomous implies that the car is acting without human involvement.
When you think about it, level 5 is the only level that should be described as autonomous. For all we know, level 5 could require strong AI to achieve. My guess is that the gap between levels 4 and 5 is much wider than most of us are assuming.
This is a silly argument because humans are autonomous and cannot drive in all environments safely. Level 5 means matching human performance, so the cars can still get stuck and require a tow, if we reach level 5.
I am using the word autonomous as defined by the SAE levels, which are described as “levels of autonomy”. You are yak shaving and contorting the words to your liking.
Also tesla is level 3, not level 2. It also doesn’t require hands on wheel 100% of the time, although they say it requires this
As I said to a different poster: this is a shameless strawman. Instead of arguing with me about what I actually want or the features of the car I actually purchased, you're inventing a fictitious consumer of a car that isn't actually for sale and trying to tell me why I didn't buy the car I did for the reasons I think I did and/or why it doesn't actually do what I know damn well it does.
All because you apparently can't handle the fact that Tesla's $10k FSD package actually provides value to consumers?
I mean, be real: I almost certainly know more about this vehicle's capabilities than you do, and I assure you that I am more expert on my own purchasing logic.
If you take delivery of this thing and you’re pleased with what it does that’s fine, but it is demonstrably not the thing it’s seller says it is, which in the aggregate consumer protection realm isn’t fine.
This is precisely how the feature is described right above the button that adds it to your configuration in the purchase UI. I really don't think this is unclear at all, and absolutely not the kind of misleading stuff everyone here keeps talking about without citing:
Full Self-Driving Capability
* Navigate on Autopilot
* Auto Lane Change
* Autopark
* Summon
* Full Self-Driving Computer
* Traffic Light and Stop Sign Control
Coming later this year
* Autosteer on city streets
The currently enabled features require active driver supervision and
do not make the vehicle autonomous. The activation and use of these
features are dependent on achieving reliability far in excess of human
drivers as demonstrated by billions of miles of experience, as well as
regulatory approval, which may take longer in some jurisdictions. As
these self-driving features evolve, your car will be continuously
upgraded through over-the-air software updates.
I don't understand. My use of "autonomy" was perfectly in line with existing convention, and this text describes the product and my understanding of it perfectly. Really, can't you just assume good faith and figure that people like me see value in and want to purchase this product knowing full well exactly what it does? Why must everything be a lie to your ears?
Maybe... you just want something different? Why are you so angry about a product you aren't buying.
I'm not angry. I just find the wording misleading. I thought you did so too, given your sentence. Given the interpretation of many people in this thread (including mine), I think there is plenty evidence that it is indeed misleading. It is nice you were not one of them, given that you seem to be the only one who bought the car.
As for assuming good faith, giving all Elon Musk's actions and statements, it is hard to assume it at this point.
raises hand I'm annoyed that these cars exist on the road. They're a danger to all drivers.
People don't need to "learn" how to operate semi-autonomous vehicles that have insufficient safety features. Tesla's autonomy does not train drivers to be attentive. In accident situations, a driver who's been lulled into complacency has no chance to respond appropriately.
Every other company took a more conservative approach to self-driving. Only Tesla barrels forward, head in the sand, ignoring all standards, like being able to exit the vehicle in an accident without needing to find the latch behind the speaker cover. A random passenger would not know this. Tesla breaks things that were not broken and claims they are innovative solutions, meanwhile wasting public investment money.
In consumer transactions the fine print is superseded by the statements, the "express warranties", made by the seller. That is the law in all fifty American states.
Not so much if you live in North America. There are a few plug-in options, but it seems like the good ones are staying in Europe and Japan because in 2021, the big automakers still don't think there's a market for EVs in the Americas.
Tesla does see a market in the Americas, so they actually sell cars here. That gives them a big leg up, even among people who don't like them.
> Not so much if you live in North America. There are a few plug-in options, but it seems like the good ones are staying in Europe and Japan because in 2021, the big automakers still don't think there's a market for EVs in the Americas.
We have, and love, our Chevrolet Bolt. In addition to the Bolt, Kia, Hyundai, Volkswagen, Porsche, Audi, Jaguar, Polestar (Volvo), and more all have electric vehicles, all available in the US.
Yep, I live in the US. There were tons of Leafs were plugged in every day at my office pre pandemic. More of those than Teslas. Also plenty of Bolts there. I see lots of non Tesla EVs on the road every day.
None with the kind of autonomy features we're discussing here. And FWIW: my judgement was that the model Y was still a win as a pure EV choice, though I agree that the other manufacturers are getting much closer there.
I guess I see your point in the context of this thread. But it's hard to see it in the context of my life where EVs are a big deal while all this self driving hype is just a silly side show. My car is pretty easy to drive already...
I think he wasn't aware that he was overselling it. He suffers(?) unrealistic optimism. This is a double-edged sword: the drive and self-belief can spur progress where prior consensus was that something truly feasible was unrealistic, but it can also be an almost Theranos-esque "we're almost here, it's really happening!" delusion.
He’s been in tech business for decades, and has long history of both failed and successful projects - he knows how R&D projects work, and how rarely they deliver, and how never they’re there on time. He has board of directors to control him.
He’s lying on purpose, because he knows he immune from any consequences. And he’s right - those lies propelled him to be one of the richest people in the world.
"Unrealistic optimism" is a feeling one has. We all get it sometimes, including everybody else leading a major company. Making repeated public statements like this is at best a mix of enormous arrogance and negligence. At worst it's a billion-dollar fraud.
To be clear, I mean "unrealistic optimism" in a way that is pathological, such as seen in bipolar hypomania. Where it is so severe that the person's confidence and grandiosity is beginning to detach them from reality.
That's way too strong of a statement. Musk is personally very popular and that spills over to Tesla and SpaceX. You can't make the claim that had it not been for Musk's bullshit self-driving claims that Tesla wouldn't have been able to raise funds. In fact, I think those claims made little to no difference.
Tesla’s profit narrative changed from making cars to self driving. They even hosted investors day, where they hyped it. Raises where largely made on a story of autonomy being just around the corner. And general hype around Musk being Ironman, that he carefully works on maintaining, exactly through events like this.
Being very popular does not mean that you are not subject to laws.
It is unquestionable at this point that a very large number of rules have been broken.
Just on the securities fraud: Musk has committed securities fraud multiple times and has been let off the hook (if anyone else did the 420 thing, they would have been barred for life for being a public company CEO...it was appalling). The same rules apply to everyone, whether you are a self-aggrandizing billionaire or not.
Multiple consumer law violations. Multiple instances of securities fraud. None of this is legal.
Also, it is possible to raise money without lying. Musk didn't raise money on the basis of these lies. He raised money initially with no claims about FSD. And the money he has subsequently got came from China.
> Musk didn't raise money on the basis of these lies. He raised money initially with no claims about FSD.
April 2019 - Tesla Autonomy Investor Day - Musk makes various claims of your Tesla turning into a profit center via "robotaxis", million robotaxis on road in 2020. "In two years Tesla will be making cars with no steering wheels or pedals."
https://ir.teslamotors.com/events/event-details/tesla-autono...
Okay, and what was the market cap then? The valuation wasn't contingent on FSD happening (I do agree with your general point, my understanding of fraud is that making those claims knowing they were impossible is fraud, so the question is whether he knew...it also isn't just a moral thing for human to do regardless of legality, he shouldn't have said it if he didn't know). I mean, no-one took those claims seriously. But when the market cap is $1tn then it is somewhat obvious that this isn't a "car business" anymore...it wasn't possible to justify that even if they made all the cars in the world.
The market cap at the time was under $50 billion and Tesla was bleeding money, a structured bankruptcy seemed like a real possibility at the time (before the FSD fraud raise).
Yep, I agree (I was short the stock). But the question is whether lying about FSD was a big part of the raise? I don't think it was, back then everyone assumed he was full of shit because he talked literal nonsense almost constantly (the number of topics on which he is a self-appointed expert is quite extraordinary). When the market cap is $1tn, that nonsense was behind the price.
> When will the Elon fanatics stop listening to his bullsh*t?
When Elon stops making such statements? And that will only happen when he faces the consequences from regulators for doling out unrealistic promises and claims.
Level 3 is actually a level where the responsibility is still with the driver.
It means that can mostly drive itself but the driver still has to be alert and ready to take over at a moments notice.
Which is the worst off all worlds, and I believe all manufacturers have said they will skip level 3 in their offerings.
We're using traffic sign and zebra crossing detection as a Turing test. Machine translations in the Play store read as if they were done by 5 year olds. When you buy a garage door, the recommendation algorithm of the most advanced company in the world suggests you garage doors for the next 2 years.
"Autonomous driving is right around the corner" my ass.
I’m no Elon fan, but Tesla FSD is still a generation ahead of anything else available on the market. GM (SuperCruise) and Ford (BlueCruise) make good attempts, but it’s not near FSD quality. Plenty of YouTube videos (@EVDave) out there that compare them against even Tesla basic AutoPilot. Super/BlueCruise are fine for pre-mapped straight line highways, but for the other 90% of the roads out there, Tesla AutoPilot can traverse most of them.
"The Waymo Driver operates at Level 4 autonomy, meaning, Waymo says, that “no human driver is needed in our defined operational conditions.” This, Waymo continues, represents “fully autonomous driving technology,” with the Waymo Driver being “fully independent from a human driver.” "
The problem is less the limits on itself, but more that they might be reached "surprisingly".
E.g. sudden dust storm on the highway in Germany where dust and/or sand storms on itself are a rarity (also this is something humans are bad at coping with themself, as given example caused a accident with multiple dead people).
Less unusual but still realistic example: Sudden extreme heavy rain (I had that once, the car I was driving in was old needed new windows and I wasn't able to see anything at all. The only reason I didn't cause a crash is because I knew the road I was driving on very very well). Or sudden fog.
Humans just tend to be better at handling such potential fatal unexpected situations, e.g. because they have a good memory of the road and contextual thinking in ways which are hard to explicitly train a robot in.
So they kinda need to find ways to still handle things well in such situations and "just stopping" isn't an option as this might endanger people around you (and yourself).
I think they're being very cautious, and for good reason. One preventable pedestrian death was enough to kill Uber's entire self-driving program. As a new technology, self-driving cars are under significantly higher scrutiny than the status quo of human drivers.
If I had to imagine, they need HD maps everywhere and don’t consider their cars capable of level 4 in non-typical circumstances like heavy snow or thunderstorms. As Waymo is targeting the ride-sharing network, they probably need to work in any circumstance in case things happen during an existing ride (so that people don’t need to be dropped off on the side of a busy highway in bad weather).
Tesla FSD isn't Level 4/5 because they're skipping level 4, at least they're trying to. Honestly the level system is pretty stupid. The gap between 4 and 5 can be tiny or it can be huge, depending entirely on what you decide to constrain your vehicle to.
Yeah. If you check youtube you can find lots of Waymo footage of driverless vehicles operating on broad, very well marked suburban roads in moderate traffic. The FSD beta seems to handle that just fine as it is, and users are also posting plenty of footage[1] (including failures!) in difficult situations like five-way stops, unmarked rural intersections that are just pavement, gridlock conditions, etc...
I went looking hard for Waymo-in-extremis kind of stuff, and it just doesn't seem to be there. I think a lot of the argument here is being confounded by the vapor problem.
Full disclosure: I pulled the trigger a few weeks ago and am awaiting delivery of a Model Y with FSD. Two levels in the hand are worth four in the bush.
> Honestly the level system is pretty stupid
Indeed, the SAE Autonomy classification scheme is the OSI Network Layer paradigm of our age. People were sure TCP/IP was going to fail for not being OSI-compliant.
[1] Edit: this is a great example from urban SF: https://www.youtube.com/watch?v=2_zTZjCCoxo&t=0s The car manages narrow lanes, double parked cars, a ton of pedestrians, hilltop crossings with poor visibility, scooters in the lane, unprotected left turns, etc... It looks great in this particular video. And I can't find anything with a Waymo car dealing with similar environments, just big flat suburbs.
You need both. It’s the difference between a company with exceptional engineering talent that also runs well, and a company with exceptional engineering talent that doesn’t run well.
I’m really happy with my Model 3 and I continue to wish them success. But these kinds of moves are just pure uncontrolled Elon Musk as-the-engineer. He’s as bad at estimating as I am. I knew the state of machine learning when I decided to just get EAP and not FSD. I’m not celebrating their miss, I’m just saying all this was avoidable.