By 2021, according to various Silicon Valley luminaries, bandwagoning politicians and leading cab firms in recent years, self-driving cars would have long been crossing the US, started filing along Britain’s motorways and be all set to provide robotaxis in London.

1 January has not, however, brought a driverless revolution. Indeed in the last weeks of 2020 Uber, one of the biggest players and supposed beneficiaries, decided to park its plans for self-driving taxis, selling off its autonomous division to Aurora in a deal worth about $4bn (£3bn) – roughly half what it was valued at in 2019.

…Prof Nick Reed, a transport consultant who ran UK self-driving trials, says: “The perspectives have changed since 2015, when it was probably peak hype. Reality is setting in about the challenges and complexity.”

…Looking back, Reed says “the technology worked … people had the sense, it does the right thing most of the time, we are 90% of the way there. But it is that last bit which is the toughest. Being able reliably to do the right thing every single time, whether it’s raining, snowing, fog, is a bigger challenge than anticipated.”

Full report

Bubblecar said:

…Looking back, Reed says “the technology worked … people had the sense, it does the right thing most of the time, we are 90% of the way there. But it is that last bit which is the toughest. Being able reliably to do the right thing every single time, whether it’s raining, snowing, fog, is a bigger challenge than anticipated.”

Well who’d have thought it?

The Rev Dodgson said:

Bubblecar said:

…Looking back, Reed says “the technology worked … people had the sense, it does the right thing most of the time, we are 90% of the way there. But it is that last bit which is the toughest. Being able reliably to do the right thing every single time, whether it’s raining, snowing, fog, is a bigger challenge than anticipated.”

Well who’d have thought it?

I like the confusion in the wording above, it sort of falls back on a linear progression idea, 90% it says, a fortunate thing everyone learnt percentages at school, of course it’s not meant to be taken literally, but appeals to the confusion. I’d suggest it’s more accurate maybe to say the technology didn’t get half way to doing what most people do most of the time to avoid accidents, like for example they stay home, drive as little as possible, program that into driverless vehicle AI

of course i’m talking bullshit, which seems to be in great part what makes the human world work

It’s only stalled for some companies. Tesla is still making good progress with their Autopilot system, it’s about midway through the beta testing stages now. About a year or two ago it was showing to be at least twice as safe as the average US driver.

transition said:

of course i’m talking bullshit, which seems to be in great part what makes the human world work

I’m not sure I agree, but it raised a smile anyway.

The Rev Dodgson said:

transition said:

of course i’m talking bullshit, which seems to be in great part what makes the human world work

I’m not sure I agree, but it raised a smile anyway.

what would honest driverless AI discuss with potential travelers, in the vehicle and others on the same road

it might first ask is your travel necessary?, then there could be some exchange about what velocity is safe, and if any velocity is safe, it may even encourage you to stay home and watch TV, from the safety of your sofa

imagine your car AI driver is a bit of a philosopher, it might first quibble about if the vehicle is really driverless, explore who the driver really is, perhaps it’s someone that works for the company that made it

it might further inquire about travelers competence of physics, ask why V^2, that exponentiation in .5M x (V^2) = KE, then become very uncooperative

i

Update on my post about the self-steering Commodore I was driving a couple of weeks ago: With the road just right, and the conditions just right, I got it to self-steer for about 20min (continuous) at 100km/hr down the Monaro Highway a few days ago. It takes a bit of getting used to….

Rule 303 said:

Update on my post about the self-steering Commodore I was driving a couple of weeks ago: With the road just right, and the conditions just right, I got it to self-steer for about 20min (continuous) at 100km/hr down the Monaro Highway a few days ago. It takes a bit of getting used to….

>>It takes a bit of getting used to….<<

Resisting the urge to go to sleep at the wheel?

…Looking back, Reed says “the technology worked … people had the sense, it does the right thing most of the time, we are 90% of the way there. But it is that last bit which is the toughest. Being able reliably to do the right thing every single time, whether it’s raining, snowing, fog, is a bigger challenge than anticipated.”

Is the right thing every single time realistic, its not like we humans drive that well ourselves.
What’s an acceptable error rate for an accident with a driverless car

Cymek said:

…Looking back, Reed says “the technology worked … people had the sense, it does the right thing most of the time, we are 90% of the way there. But it is that last bit which is the toughest. Being able reliably to do the right thing every single time, whether it’s raining, snowing, fog, is a bigger challenge than anticipated.”

Is the right thing every single time realistic, its not like we humans drive that well ourselves.
What’s an acceptable error rate for an accident with a driverless car

the question might be who takes responsibility for errors, to which a court case would explore that, test it, and such a case may determine the vehicle is not driverless at all

so it’s a question of possible liabilities, of which car manufacturers are already well acquainted

transition said:

Cymek said:

…Looking back, Reed says “the technology worked … people had the sense, it does the right thing most of the time, we are 90% of the way there. But it is that last bit which is the toughest. Being able reliably to do the right thing every single time, whether it’s raining, snowing, fog, is a bigger challenge than anticipated.”

Is the right thing every single time realistic, its not like we humans drive that well ourselves.
What’s an acceptable error rate for an accident with a driverless car

the question might be who takes responsibility for errors, to which a court case would explore that, test it, and such a case may determine the vehicle is not driverless at all

so it’s a question of possible liabilities, of which car manufacturers are already well acquainted

Perhaps non liability contracts were you acknowledge something may go wrong, outside testing of vehicles by non vested parties, software patches would be a given based on user feedback for example, sensor data, etc

Cymek said:

transition said:

Cymek said:

…Looking back, Reed says “the technology worked … people had the sense, it does the right thing most of the time, we are 90% of the way there. But it is that last bit which is the toughest. Being able reliably to do the right thing every single time, whether it’s raining, snowing, fog, is a bigger challenge than anticipated.”

Is the right thing every single time realistic, its not like we humans drive that well ourselves.
What’s an acceptable error rate for an accident with a driverless car

the question might be who takes responsibility for errors, to which a court case would explore that, test it, and such a case may determine the vehicle is not driverless at all

so it’s a question of possible liabilities, of which car manufacturers are already well acquainted

Perhaps non liability contracts were you acknowledge something may go wrong, outside testing of vehicles by non vested parties, software patches would be a given based on user feedback for example, sensor data, etc

you still can’t determine the risks to others that aren’t associated with the operation of the vehicle, the risks or liabilities go further to bad press, which effects share value and investment

soon as you build safety systems into anything, and importantly make claims it does this and that (avoids obstacles etc/crashes), the liabilities increase with the added claims it does more, and no less so if sold on the promise of doing those things, and so called driverless (a term I don’t buy) vehicles will be very much sold on claims they do so much

the usual thing presently is when something goes wrong the police look hard for a mistake, driver error, someone did something wrong, someone screwed up, failed to do something or whatever

it’s a long way off before police will be routinely looking for a software bug or hardware failure in a autopilot system of a car