Hey, there! Log in / Register
'Self driving' Tesla fails miserably on the streets of South Boston
By adamg on Tue, 02/01/2022 - 9:21pm
Tesla is currently beta-testing something called "Full Self Driving." Taylor Ogan, CEO of Snow Bull Capital, a local investment firm, decided to try it out in Southie, you know, where the roads are mostly laid out in a grid. Fortunately for him, the car, other drivers and pedestrians, he never took his hands more than about an inch off the wheel:
Here's a typical drive on Tesla FSD Beta, and why it's the opposite of useful. It's most certainly not even close to being "safer than a human", by any factor.
22-minute drive, 4.5 miles, and WAY too many interventions. pic.twitter.com/oKmkRZ3Tgh— Taylor Ogan (@TaylorOgan) February 1, 2022
H/t Shamus Moynihan.
Neighborhoods:
Topics:
Ad:
Support Universal Hub
Help keep Universal Hub going. If you like what we're up to and want to help out, please consider a (completely non-deductible) contribution.
Comments
The numbers of self driving
The numbers of self driving Tesla deaths are grossly underreported, because usually the car will exit self driving mode a fraction of a second before impact and the crash gets counted as operator error. They also have a big problem with wheels falling off while driving (Google "whompy wheels") and with people getting locked in their car after an accident and burning to death. I can't wait until the day Elon Musk does the world a favor and launches himself to Mars. Guy fucking sucks.
What are you talking about?
Tesla doesn't have self-driving yet. It's still just beta and very limited users have it. Everything else is assisted driving and is still classified as essentially fancy cruise control. The driver is still in charge completely. They just don't have to do much.
"Whompy wheels" is a crusade from some guy in Australia to sew FUD. Some early Model S cars had a suspension/ball joint issue. They issued a service brief to get those fixed and newer models haven't had the same issue.
People haven't been locked in their cars. The door handles didn't extend for someone outside to help an unconscious driver. And that has happened like 3 times?
Musk is an asshat, but the Tesla is a pretty good car. Self driving (without a full radar array too) is still a pipe dream.
Teslas are not good cars.
Teslas are not good cars. They have extensive quality control issues, especially for vehicles at luxury price points. Just over the past week or so it has come out that in Germany's vehicle road worthiness certification Teslas have the worst failure rate of any EV - 10.7% for relatively new cars. That's about double the average failure rate.
Tesla has "Autopilot", which is a dangerous fraud
and is almost certainly what ZachAndTired was referring to, rather than the new "full self-driving". (Which is probably similarly fraudulent, but I haven't read up on.)
Customers know what "autopilot" means in common language: Not having to pay attention. But that's not what Tesla is actually selling, and this is killing people.
I don't know how the numbers actually shake out -- does lane-keeping and automatic braking result in less injury and death overall, even when you account for the times a Tesla speeds directly into a bollard or the side of a large truck? Not sure, haven't looked for studies.
Bottom line: Tesla is lying to people about the car's capabilities, using a term that people associate with a common understanding of "self-driving car", even when it's very much not able to do that.
(Which is a shame, I'm sure it's a very nice car in other ways, and I would love to see more gas vehicles replaced with electric in general.)
Tesla
Tesla has pushed other automakers to drastically increase their investments and seriousness into electric cars.
We wouldn’t be where we are, even right now, has they not built what they built. The options before, and the designs forecasted by other makers during early Tesla days, were far behind what Tesla was doing. It forced major capital investments by those brands to advance their products for quality.
Tesla has pushed others to actually embrace and move towards electric. They are a significant catalyst in the fight for climate change.
this is a well articulated
this is a well articulated position, it's a great example of the
https://en.wikipedia.org/wiki/Great_man_theory
I'm not a fan of Musk or the theory, had Musk never walked the earth we'd not be sitting around baffled about the transition away from fossil fuels.
The reality is that Tesla
The reality is that Tesla probably actually slowed the mass adoption of electric vehicles by stock piling regulatory credits and selling them to other automakers. This allows traditional automakers to comply with laws that require they carry a certain number of these credits each year without actually manufacturing any electric vehicles. This is also the source of a large portion of Tesla's revenue. Elon Musk's objective is not to fight climate change, it's to amass as much wealth as possible by exploiting loopholes in government incentive programs.
No
The major automakers operate in “cost centers”. Everything is a cost, with the goal to drive that down to increase profit. Because of this, there was resistance to building quality electric vehicles. R&D was not heavily funded.
Then Tesla comes in, looks at it as product development (as opposed to cost center). They operate as a tech company making a car product.
Because they accelerated innovation, other manufacturers needed to increase R&D to match.
Tesla greatly advanced the electric vehicle market.
This should be shown
This should be shown nationwide on TV and online to let the public know how serious these defects this car has.This car can potentially kill a pedestrian.
Tesla recently had a recall
Tesla recently had a recall of their 54,000 vehicles,reports that the car failed to stop at stop signs..
Any car can have "wompy wheels"
Most common cause is some jackass going nutso with an air hammer on the lugnuts.
Tesla has an outstanding
Tesla has an outstanding lawsuit against them for it and China forced a recall due to it. This isn't a fluke thing.
https://www.hotcars.com/exclusive-whompy-wheels-hunter-whos-the-man-afte...
Chinese recall is bunk
https://electrek.co/2020/10/23/tesla-argues-no-defect-suspension-china-f...
And you can sue anybody for anything. This is one of those "if you want to vilify someone, all you have to do is find enough anecdotes" situations.
Oh well if Tesla says it's
Oh well if Tesla says it's unnecessary then it must be! I'm sure they'd have no reason to lie about it.
Also, what a shock that Electrek is coming to Tesla's defense.
https://www.thedrive.com/tech/21838/the-truth-behi...
Who are you mad at?
Is it Electrek's fault the NHTSA investigated and hasn't issued a recall? Or is it Tesla's?
I'm just saying you might
I'm just saying you might want to find a better source than a blog whose sole purpose is to pump up the price of Tesla stock.
This never would have
This never would have happened in old Southie.
Musk worship is an American
Musk worship is an American disgrace, as a practical matter he sits on enormous amounts of capital that could be used to make incremental progress on really important stuff.
Really is astounding how
Really is astounding how people worship Musk, whose contributions to the world are widely a side effect of his wealth seeking, while somebody like MacKenzie Scott is giving away millions of dollars to targeted charities, making palpable changes for hundreds of people every day, and is largely ignored.
When Tesla can do "self parking"
in South Boston after a snow storm and not get vandalized afterwards, the feature _might_ be ready. :-)
Product idea
This suggests the idea of an autonomous snow-shovelling robot that clears a parking space, and then plants itself in it as a space saver. Every household in Boston would want one.
As usual, the fact this
As usual, the fact this technology is being driven by the egos of men with too much money means it's all pushing in the stupidest directions. You will NEVER get an AI to understand how to drive in Boston. An AI doesn't recognize "a truck", it recognizes a series of pixels that have a 84% correlation to what it's been told is a series of algorithms. It's just not going to work for places with such a dynamic, nonsensical, rules-optional, illogically laid out, and aggressive driving culture. I'd love to see a Tesla try and get itself out of a parallel parking space surrounded on both sides by 5 foot snowpiles. Truly.
What it WOULD work for, and has done well in tests, and is DESPERATELY NEEDED considering the shortages in bodies right now, is depot-to-depot long haul highway driving. Eliminate pedestrians, most turns, cross-path traffic. Introduce long stretches of straight, consistently build road and federally mandated consistent signage, and that 84% correlation can creep up to the 90s where it's actually useful.
But nobody gives a fuck about trucking and Elon Musk wants to sell sexy tech cars to dudebros and incels who harbor dreams of being Tony Stark while using the entire technology to write off billions in taxes and play around with carbon credit trading for other companies.
Tesla in Southie
Parking in Southie after a snowstorm is hard charging your electric vehicle from your triple decker is impossible.
To be fair,
any "self driving" vehicle that could successfully navigate Boston consistently without incident would be a marvel of technology . . .
Any human doing that
would be a marvel of biology.
Should be a test of basic skills
Slowing down and paying attention are not beyond the grasp of the typical human.
Driving without incident, yes
but it's the navigation part that causes problems here.
Tesla also failed to self
Tesla also failed to self-navigate Downtown in the Financial District/Faneuil Hall area. I think they tried last year.
The problem isn't Boston. It
The problem isn't Boston. It's declaring that this technology is safe enough to use in public, when it's clearly not even close.
There's nothing at all unusual or difficult about the driving conditions in the video. If the software can't handle that situation, it can't drive anywhere.
And I don't think it's safe on highways either. There was that news story of the self-driving car engineers who got seriously injured in a crash because the car kept accelerating instead of slowing down to allow another car to merge at an on-ramp. Maybe according to the software engineer's interpretation of traffic law that's what you're supposed to do when you have the right of way, but that's not how human drivers negotiate to avoid a horrible accident.
Not sure self-driving will ever work in the city
My wife and I debate the future of truly self-driving cars. She's convinced they are just a few years away. I'm convinced that there are too many unknowns, as seen in that video, of people and peds doing unexpected things that need to be reacted to. Even if they get the technology perfect, a self-driving car will run over some little kid (even if it was totally the kids' fault and no human could have done any better.) The outrage will cause them to be banned or severely curtailed.
I can see self-driving cars (and trucks) working and being safe on Interstates, but I doubt that the day will come where cars will routinely self-drive in the city.
Not happening for decades
The final 10% is too tricky to resolve. Something like autopilot on a highway system where you are mostly going at a set speed without a lot of decisions sure but not driving around Boston level.
Not happening *safely* for decades, you mean.
My guess is that we'll see it permitted in some states and not others, and that it will cause a lot of deaths and anger and outrage from being rolled out too early.
anger and outrage
I'm with you on "lots of death" but I doubt we'll see any significant anger or outrage. Look how little outrage there is when a pedestrian or bicyclist is killed by a vehicle driver in Boston.
In the scenario you describe (which I find likely), the "full self-driving [sic]" technology will likely be rolled out in some southern and western states. It will probably perform moderately well on highways and outside of cities, and most of the deaths are likely to be pedestrians in urban areas (i.e. poor people), in which case the concerns of wealthy Tesla drivers (i.e. the ability to continue using their expensive toys) will be prioritized over the lives of poor pedestrians.
Only because it's new
and new things are scary and worth reporting on. (-:
Agree/Disagree
Sadly, I disagree. Even right now people shrug when others are killed due to preventable driver error.
There's no way self driving cars will be safe or will be preferable to human drivers. The auto manufactures are already suggesting pedestrians and cyclists carry little beacons to make them more "visible" to the self driving electronics.
But I have no doubt in a few years these cars will be on the road, killing people, and society will shrug.
There will be people
saying that self-driving cars are safer, no matter what the statistics show.
How long
did it take to find a parking spot?
Boston really really needs …
… more cars in circulation.
What I don't understand is
What I don't understand is why the authorities have allowed this flawed technology to be used on public roads. How was this allowed to happen?
What law was broken? How would they stop it if it was?
First, how do you think you'd enforce this? Cops in this city enforce nearly zero traffic violations to begin with. And now you want them to determine if the person is using a program some of the time to steer and handle speed control?
Second, the owner is behind the wheel and still in control of the vehicle. He just using an algorithm to make changes to speed and direction instead of his hands and feet. And he takes over with his hands and feet whenever he wants to make changes to speed and direction. This is barely one step beyond fancy cruise control.
There are already cars all around you on the road that control steering and speed to avoid accidents, stay in lanes, cruise, etc. This is just doing all of those things at once all of the time.