Yes and as with Tesla "FSD", supposedly the driver is to devote full attention to monitoring the car whenever activated. But this is silly because as these tools get better, the driver will inevitably monitor less. But it won't any time soon (ever?) avoid random errors every now and then. But now these will be catastrophic failures.
Driver training should be updated to include what to do when self-driving modes are used (ie what do you actually do to supervise the vehicle, which checks to do, etc).
I don't really see how the needed monitoring would be more interesting than fully driving the car, maybe with a few convenience and safety assistances. Planes can use autopilots because there are not litteraly surounded by traffics 2m away. If a car needs monitoring, and they will probably need it for a very long time, you should as well drive it.
> If a car needs monitoring, and they will probably need it for a very long time, you should as well drive it.
This is the crux of the intractable problem of AI/self-driving marketing vaporware, and why until this nonsense is shut down, the situation will probably have to get scandalously worse before it gets better (better meaning regulating fake "self driving" off the road as a public safety travesty).
> But this is silly because as these tools get better, the driver will inevitably monitor less.
And for good reason. The main benefit to FSD, as I see it, is the ability to do something more productive with my time while getting from A to B. Most wealthy people end up paying for drivers not because they want an exceptional driver, but because once you reach a certain threshold of income it is more profitable to work (phone calls, laptop, w/e) or sleep, and pay to be driven everywhere than to drive yourself. FSD lowers that income threshold tremendously and could be a significant boon to overall human productivity. If it ever happens.
>is the ability to do something more productive with my time
Seems to miss my point, I may have not stated clearly. Yes I understand the sales pitch of self driving cars stated above...the "marketing" / "hype" of AI/self driving cars/ Tesla FSD evokes the idea of completely offloading the responsibility to the car and doing whatever you want with your time during commutes, as if it was public transit.
But my point was anything less than this is a complete waste of time and useless as a product because until you are completely hands and attention off, you have to be completely hands and attention on/ready...no matter how good or near actual FSD it gets.
Because 99.9, or 99.9999, or 99.999999% effective is effectively 0 percent effective given the catastrophic outcome of any error if you aren't devoting full attention to monitoring the car.
But aside from all of this theorizing, the actual product is NOWHERE near high 9's reliable. As it gets better, attention will drop but errors will still be pervasive. And with this as backdrop, the fact the marketing has ever even hinted at being general-AI type FSD (the name itself for example) is a scandal. The videos of people riding in the backseat, or any other instance of not monitoring the car fully because they think it can drive itself. I predict there's going to a be an endless stream of accidents from these cars then FSD will be shut down until fully proven, which won't be technically possible for a long long time if ever in our lifetimes.
> But my point was anything less than this is a complete waste of time and useless as a product because until you are completely hands and attention off, you have to be completely hands and attention on/ready...no matter how good or near actual FSD it gets.
Agreed. If I got a Telsa tomorrow I wouldn't do so with "FSD".
> Because 99.9, or 99.9999, or 99.999999% effective is effectively 0 percent effective given the catastrophic outcome of any error if you aren't devoting full attention to monitoring the car.
I'm not sure what effectiveness you're measuring. There will always be fatal accidents with heavy vehicles moving at high speeds. It's more of a regulatory issue where the accident rate of auto-pilot systems is sufficiently lower than human drivers (because I believe it is already lower in many situations, just not sufficiently lower or ubiquitously lower) and we decide its good enough to no longer require monitoring by the driver.
Edit: I guess its even more complicated than that. Using hypothetical numbers, if the accident-free rate of a human is 99%, and auto pilot is better at 99.9%, but combined (auto-pilot with human monitoring and rare intervention) is 99.999%, can we justify allowing no monitoring in that situation? Can we expect people to follow the law or will it turn in to more of a speed-limit type grey area.
Not all errors or accidents are created equal. As someone else mentioned in this thread the nature of human error is qualitatively different than the nature of seemingly random errors we see with self-driving. Code bugs can produce any scale of error. Humans do no in any given situation have open-ended likelihood of making any errors.
Search youtube videos of Teslas if you haven't seen them yet. The cars will randomly swerve into parked cars, inexplicably steer out of turns, aim directly at medians on the highway. Human errors on average do not equate to this if we want to compare human error "rates" to machine. What is the rate of a random catastrophic error? Machines are way way too high for the foreseeable future, and because of the nature of software bugs they arguably always be higher than humans. The scope of what AI has to be to achieve "general" driving on open-map roads with near infinite variables nearly equates to "general" AI itself.
You can't allow "halfway" monitoring the car. It's all or nothing. It's always monitored or it's not monitored because errors can happen at any time. But anyways, humans aren't robots. Their attention basically not focused if it's not focused completely... multi-tasking is false, and humans already fall asleep fairly easily driving manually. If they have to sit there "focused" but not driving this will lead to even more instances of sleeping or just not monitoring generally.
The numbers are a thought experiment what I mean by effectiveness or mistake or error is just the driver (man or machine) doing something they shouldn't have. Messing up. Humans have very low rates of these mistakes/bad actions/errors. They have very very low rates of catastrophic mistakes. I predict Tesla's won't be 99.anything for a long time. But think of how error-free they have to be to account for the huge domain of roads, signs, weather, terrains, other cars such that when they make their mistakes it doesn't result in disaster... or just "accidents" galore.
