"Risk compensation is a theory which suggests that people typically adjust their behavior in response to perceived levels of risk, becoming more careful where they sense greater risk and less careful if they feel more protected."
https://usa.streetsblog.org/2017/09/13/wide-residential-stre...
I was first introduced to "wide streets in neighborhoods are more dangerous than narrow" on HN years ago. (I don't think it was the linked article, but that was the first one that came up just now after a search :P )
Since having read that, I've actually noticed how true this is, at least to me anecdotally. When I'm driving in a neighborhood with crowded streets, I can't bring myself to go over 15MPH, much less over the speed limit (typically 25 in neighborhoods in the US).
Wide streets give a sense of security. So I feel like people are less likely to pay attention going around bends, parked cars, etc, than if they didn't have that sense of security.
Forget the average, how about the bottom 10-20% of all drivers? I don't trust the bottom 10% driving with "Autopilot" at all, zero. They are going to use it to go on autopilot while driving, exactly as the marketing implies. I mean there has to even be people who think the car itself is conscious. Car has advanced AI, must be conscious.
To think otherwise is just highly underestimating how clueless some people are.
They then use a "common formula" to show that that leads to more fatal accidents, but didn't actually study on actual crash data.
There is absolutely no doubt I'm a safer driver with Comma than without it. I'm still in control, but Comma not only allows me to expend less effort driving (which allows me to stay alert over longer periods of time), but also be much less emotional when driving. I'm pretty convinced that a large percentage of accidents are caused by frustrated or bored drivers doing crazy things that you just don't feel the urge to do with the assistance of self-driving.
Edit: After one minute I got a downvote.
Sorry, I actually meant to refer to Birken's comment above. Advertising might not have been the best word - astroturfing? If you read their comment but replace "comma" with "tesla" it still reads as spammy.
Yours was fine (though discussing downvotes will always get you downvotes, my comment included).
> Otherwise, is every person here who mentions Tesla advertising too?
Only the ones who needlessly sing the praises of the Tesla autopilot in barely-related threads.
I guess that we have to look at the results here to judge whether too many people are not paying attention. Hopefully the investigation will reveal whether the autopilot incidents of collision with emergency vehicles is significantly more frequent or less frequent than from vehicles being driven in the traditional way.
https://www.latimes.com/business/story/2021-06-29/nhtsa-adas...
Tesla released data in the past, but that’s quite suspect as they have an obvious agenda and aren’t known for open and honest communication.
https://www.latimes.com/business/autos/la-fi-hy-tesla-safety...
I also think there should be dedicated lanes for self driving cars..
A very good friend of mine was a sensor engineer at google working on virtual sensors that interacted with hand gestures in the air... and is now a pre-eminent sensor engineer for a large japanese company everyone has heard of...
We drove from the bay to nprthern california in his tesla and it was terrifying how much trust he put into that car. I got car sick and ended up throwing up out the window...
Knowing what I know of machines working in tech since 1995 -- I would trust SHIT for self-driving just yet.
Autopilot bothers me for a number of reasons. Fundamentally it's a poor driver, spending time in people's blind spots unnecessarily, braking and accelerating in rather abrupt ways, and just generally acting like a teenage driver who just got their license. It simply doesn't practice defensive driving.
I also spend a lot of time driving on undivided rural highways. These are highly dangerous roads, with closing speeds in excess of 200 km/hr at times. In those situations autopilot drives far too much according to the strict rules of the road and can't adjust to the situation. It doesn't use the lane space to leave additional room and it doesn't give a wide berth to cyclists.
It also bothers me deeply that one of the ways to override autopilot is to make steering input and that's also the indicator Tesla uses to determine the user participating. I haven't used autopilot much for the reasons above, but the few times I did active it the steering input required to tell the car I'm there is also enough steering input to move the car several feet in the lane due to very direct steering. It feels like I'm just fighting the car. That is deeply unnerving since the force required to override autopilot feels like enough to jump nearly half a lane and cause a collision.
I'm guessing GM Supercruise since it's the only one I'm aware of that uses eye tracking in production (though Tesla claims to have enabled that in the US just recently. Personally I'm not sure their camera placement can really do proper eye tracking). Supercruise's disengagement rate is low though, generally much lower than Tesla's.
I do like supercruise from what I've seen of it (haven't had a chance to actually use it since GM seems determined to waste their advantage by not rushing it into every car they make).
Ive been driving for 25 years, cars, trucks, trailers, standard and auto transmissions, and I have never once thought to myself "I'd be such a better diver if I didn't have to pay attention to my speed, lane keeping or following distance" Why? Because those mundane things are already on autopilot in my brain.
Posts like yours are so absurd to met that I cant help but think shill.
It's like people who do math by hand and then get a calculator.
I would be terrified to share the road with someone of this mindset. Your vehicle is a lethal weapon when you are driving it around (assisted or otherwise). At no point can someone claim that a tesla vehicle circa today is able to completely assume the duty of driving. You are still 100% responsible for everything that happens in and around that car. You had better have a plan for what happens if autopilot decides to shit the bed while a semi jackknifes in front of you.
The exceptions are what will kill you - and others - every single time. It's not the boring daily drives where 50 car pileups and random battery explosions occur. Maybe your risk tolerance is higher. Mine is not. Please consider this next time you are on the road.
When you automate away the mundane, exceptions are much easier to catch.
Self-driving cars will take that stuff over 99% of the time, but the 1% where it screws up is the dangerous part. There are plenty of examples where a self-driving car seemingly randomly seems to go completely haywire, without any obvious reason.
Instead of spending brain power on driving properly, you now have to spend brain power on looking at what you should be doing _and_ checking if the car actually agrees and is doing it.
Staying 100% focused without actually _doing_ anything is incredibly difficult. Many countries intentionally add curves to their highways to keep drivers alert: having a pencil-straight road for hundreds of miles really messes with the human brain.
That doesn’t help at all if you’re reading a book or playing on your phone, both of which are things I observe Tesla drivers doing pretty often when I’m in the Bay Area.
It's the same thing here. The computer is assisting me so that I can take care of the novel situations, the exceptions if you will. I can pay closer attention to the road and see that jackknifed trailer sooner because I wasn't looking at my speedometer to check my speed.
And I don't have a Tesla, I use a different autopilot system.
He's saying that he no longer has to worry about those things the same way cruise control lets you not worry about the speedometer needle and dedicate more of your attention budget outside the car.
Of course you can be an idiot and spend it on your cell phone but that's not really a failure mode specific to any given vehicle technology.
Obviously that's a single anecdote, and I don't know if it would have gone through with it because I immediately corrected, but that was my experience.
The question is whether a system that absolutely requires that you pay attention going through intersections (which you should obviously do) is safer in aggregate than not having those features enabled at all in those situations.
E.g. are weird lane changes that people don't catch happening more frequently than people zooming through red lights because they weren't paying attention. Only the data can show that, and Tesla should share it.
Personally, I tend to turn off things like lane keeping because I end up having to babysit them more than I would like. It doesn't always read the lanes correctly, though I have not tried Tesla's technology yet.
I don't know if this "AI" has any sort of quality control, but how difficult is it to test if it detects a solid white line on the side of the road in at least 6 out of 10 tries
it also tends to suddenly disengage and pass control to the driver at most dangerous parts of the trip e.g. when passing other car in narrow lane, etc.
This "driver assistant" is a series of disasters in the making.
But, I feel that this depends on the type of driver and their personality. I, for on, have never felt comfortable with cruise controls, even adaptive ones, let alone partial self-driving. I had always found that I am more comfortable when I was driving rather than trying to make sure that the adaptive cruise control is able to make a complete stop in case of emergencies. Perhaps I'm just a little untrusting and paranoid :).
But no, I don't trust it to drive itself. If I'm tired I won't drive, regardless of autopilot.
I'm curious about this part. Do you manually input a limit, or trust it to read street signs?
And how often do you look at your speedometer anyhow? I think on the highway I glance at it maybe once every few minutes and otherwise match speed with the other vehicles, and in the city I look more often but mostly just drive at what feels a safe pace (which seems to match the limits, more or less.)
It's true, I don't look at the speedometer all that often when driving manually, but it's just one less thing to worry about.
I have autopilot on my car, and it definitely makes me a better and safer driver. It maintains my distance from the car in front and my speed while keeping me in my lane, so my brain no longer has to worry about those mundane things. Instead I can spend all my brainpower focused on looking for potential emergencies, instead of splitting time between lane keeping/following and looking for emergencies.
I no longer have to look at my speedometer or the lane markers, I can take a much broader view of the traffic and conditions around me.
Before you say it's impossible to be safe driving with an assistive product, I suggest trying one out.