In my experience (Tesla), attention monitoring works well even when I'm wearing sunglasses. The camera can still see my eyes even through dark polarised lenses.

It may depend on the sunglasses, however - other people report problems with sunglasses that have mirrored lenses etc.

Sometimes I wonder if Tesla also has a much better software stack than most other manufacturers. IIRC, Tesla has had interior cameras in their cars for years now and I haven't heard about major issues stemming from it.

The camera was not actually used initially. With the old Autopilot software, attention monitoring was (and still is?) exclusively done with steering wheel torque sensors. Our camera got enabled when we purchased the FSD upgrade. I do agree the software is good, however: it's both more effective and much less naggy/annoying than the old "hands on wheel" method.

It still falls back on the torque sensor (requiring hands on wheel) if it thinks you're not paying enough attention, or if it can't see your eyes for whatever reason.

And I guess Tesla must have enabled the camera for all new vehicles now, at least in Europe, given that it's required.

Wow, I did not know that! Thanks for the clarification, that is actually decently well thought out.

It wouldn't be surprising if they did have a better tech stack. also judging from the state of things, it wouldn't be that hard either.

It's a matter of time before someone invents sunglasses with eyes painted on the lenses.

I can answer this, since I have a new car with this camera and polarized sunglasses.

MOST of the time it's good about telling when I'm looking and when I'm not, out of maybe... 5 alerts over the previous 8 months all, but one occurred when I was in fact looking away for one reason or another. Likewise when it's correct my lane-keeping it's been right about me drifting.

Given how inattentive I see other drivers being, on their phones for example, and taking into account that I'm (based on my record) a good driver who is attentive... I appreciate these additions. I doubt that they make us less safe, we just dislike anyone or anything telling us how to drive, because "we already know what we're doing." The subjective experience of being distracted however isn't usually so clear-cut, it FEELS like you're paying attention.

Note: This is a new model Lexus, so I expect this represents that brand as well as Toyota, but beyond that I don't know.

Just because it's available in a Lexus does not mean it's available in a Corolla

In Europe it does. ADDW is required in all new vehicles, including Corollas.

There is no way that training people not to worry is making us all safer. I don't even like how new cars have this thing where they will automatically hold the brake once stopped even if you let go. There is no way it's a net good to train people that running cars just stay where you put them like an inert object does.

The result for me has been pretty predictable, I'm getting a kind of corrective feedback so I drive in a way that prevents that feedback. The practical result is that imo I'm a better driver, I'm more aware of my lapses in attention, my tendency to overcorrect to the right, and so on.

I'm yet to experience a downside, this isn't like using "Autopilot" or some other situation where a machine is taking over for me. I don't see why our skills or caution would be lessened by exposure to realtime feedback.

Cars have had that feature ("hill hold assist") for 20+ years now. Even longer for high-end models. There can't be all that many vehicles left on the road that don't have it.

Besides, you could make such an argument for any safety feature: "I don't like how new cars have ABS brakes. We shouldn't be training people that you can just slam on the brakes on a slippery surface and expect the wheels to not lock up..."

... "I don't like how new cars have seatbelts. We shouldn't train people that they can just crash into things and not have their face go through the windshield"...

The GP seems to be describing "auto hold", which is a different feature from hill start assist, and is newer and less prevalent.

OK. But it's essentially the same concept. Just that auto hold will go a step further and apply the park brake to hold the car in position indefinitely rather than just for a few seconds.

There have been many deaths over the years where drivers forgot to engage the park brake, then stepped out and got run over by their own vehicle. Including, sadly, Star Trek actor Anton Yelchin. If his car had had Auto Hold then he would very likely still be alive today.

The only scenario where this could have a negative effect on safety is where you, the driver, were used to driving a car with the feature but got surprised by its absence when driving an old car. But as I facetiously illustrated in the post above, this is not a valid argument against new safety features. Any accidents that this scenario might cause are far outweighed by all the ones that will be prevented in the long run.

Nowadays cars engage the electronic parking brake (and P on the gear selector) when you open the door. Which has caused another crash, there's a video of a lady stuck at a rail crossing who was screaming she couldn't get the car going, she had opened the door and in the panic didn't notice the car did all that.

Anton Yelchin's death was more annoying, it was a design defect that pushing the gear lever all the way to the top didn't put it into P (like cars of the 80s, of even 50s), the gear lever acted more like a 4-way button, and drivers needed to push it twice (or 3x?) to go to park. Unfortunately his car was probably not new enough to have auto-park when driver door is open. His family got a settlement from the manufacturer..

You have not illustrated anything about the validity of any arguments. You have only expressed a feeling and said some things that sound like reason to you.

Which to be clear is all I did myself, but I did say it that way honestly. "there is no way that..." doesn't pretend to supply any proof, just the reasoning.

Neither of your attempted counter-examples actually amounts to the same sort of reasoning.

Except in a way you are also right and that rightness actually supports my point. I would equate you seatbelt nonsense with bicycle helmets, where it is quite up for grabs if they have been a net positive, because both the riders and the drivers are falsely trained to think the cyclists are safer and don't need to worry as much. There are not many real studies, and the few that exist, you can make them say whichever side your emotional belief wants.

But I wasn't quite going there. Operating a vehicle is operating a vehicle. Until such time as the driver, or rather mere user, bears absolutely no responsibility, like riding the bus or a train , or maybe an elevator, then "there is no way" yes a simple assertion, that it's good to train everyone that they are riding in elevators where they only need to maybe press one button to say what they want but beyond that it's out of their control and so out of their responsibility what happens after that, because we do not actually live in that world yet.

If you had put any thought into either of those, you might have come up with some argument that actually applies. Neither of these equates or relates at all.