Beware Driving Automation

[markjay]

They have to do something to justify their £65k salary!

My guess is that the London Underground automation wouldn't have been possible in the first instance, without a deal with Aslef that guaranteed the drivers' jobs.

 

[ChipChop]

Despite the claims of the Tesla PR machine (miles driven per one accident??) the number of road accidents in the US is forecast to increase from now until 2029 by +12%. That forecast is based on road accidents per million US residents. See statista.com.

 

[Doodle]

The problem with all these automated systems is that they're based on rules, and when something happens that isn't covered by that ruleset they don't know what to do (even those with AI - underneath all that supposed 'intelligence' there's still a set of taught instructions it worked from). Expecting the man on the Clapham Omnibus to then jump in and uncluster the situation, when they don't understand why the system has suddenly bailed and with a good probability they weren't paying attention in the first place, is going to have inevitable results.

The day that one of these automated driving systems is willing to sacrifice its owner as the best outcome in a no-win situation, then we'll be making progress.

 

[MrGreedy]

The day that one of these automated driving systems is willing to sacrifice its owner as the best outcome in a no-win situation, then we'll be making progress.

What, like 'my owner is a megalomaniac business psycho who I've just heard on a phone call is immenently going to go ahead with a competitor grudge match power play putting it ahead of the safety of their workforce and customers/consumers, resulting in hundreds of thousands of deaths: I'd better drive off a bridge whilst taking them to work for the greater good in this no-win situation'.
This sort of thing?

 

[Doodle]

What, like 'my owner is a megalomaniac business psycho who I've just heard on a phone call is immenently going to go ahead with a competitor grudge match power play putting it ahead of the safety of their workforce and customers/consumers, resulting in hundreds of thousands of deaths: I'd better drive off a bridge whilst taking them to work for the greater good in this no-win situation'.
This sort of thing?

Might be a bit much for the initial release but I'm sure we can get a firm deliverable for the first update

Probably going to hurt the sales of those halo models all the CEOs are buying though.

 

[clk320x]

Despite the claims of the Tesla PR machine (miles driven per one accident??) the number of road accidents in the US is forecast to increase from now until 2029 by +12%. That forecast is based on road accidents per million US residents. See statista.com.

What’s that got to do with automation? The stats Tesla quote are comparing their own cars being driven with AP or by the driver. Not sure which part you don’t agree with…

 

[ChipChop]

What’s that got to do with automation? The stats Tesla quote are comparing their own cars being driven with AP or by the driver. Not sure which part you don’t agree with…

If Tesla's "statistically AP is safer than unaided driving" was a statement of fact you would expect there would be a decrease not increase in the forecast USA accident rate over the next 5 years.

 

[markjay]

If Tesla's "statistically AP is safer than unaided driving" was a statement of fact you would expect there would be a decrease not increase in the forecast USA accident rate over the next 5 years.

Depending on what market share Tesla predicts for itself over the next 5 years.

At current, less than 1% of cars on America's roads are Tesla. Their impact on overall crash statistics is insignificant.

 

[clk320x]

If Tesla's "statistically AP is safer than unaided driving" was a statement of fact you would expect there would be a decrease not increase in the forecast USA accident rate over the next 5 years.

How? As above this depends solely on the % of vehicles that will be Teslas… Oh dear.

 

[ChipChop]

Depending on what market share Tesla predicts for itself over the next 5 years.

At current, less than 1% of cars on America's roads are Tesla. Their impact on overall crash statistics is insignificant.

But as you recognise there are a number of Tesla cars on Americas roads. And they are contributing the square root of nothing to improving the forecast US accident rate over the next 5 years. "Statistically AP is safer than unaided driving" - a statement of fact? Or Tesla marketing spiel for the ears of their beleaguered investors?

 

[clk320x]

But as you recognise there are a number of Tesla cars on Americas roads. And they are contributing the square root of nothing to improving the forecast US accident rate over the next 5 years. "Statistically AP is safer than unaided driving" - a statement of fact? Or Tesla marketing spiel for the ears of their beleaguered investors?

The fact is that if you own a Tesla and use the AP you are statistically safer than if you didn’t - not sure why you feel that a brand that accounts for 1% of the percentage of total cars would improve the overall rate of accidents in the USA? The fact being that per mile covered AP has had a lower incident rate than when the car was driven manually - and in both cases this was still lower than the accident rate by mile covered as a whole in the US accounting for all cars, this is due to the various collision avoidance features in Tesla’s with or without AP being active.

 

[markjay]

But as you recognise there are a number of Tesla cars on Americas roads. And they are contributing the square root of nothing to improving the forecast US accident rate over the next 5 years. "Statistically AP is safer than unaided driving" - a statement of fact? Or Tesla marketing spiel for the ears of their beleaguered investors?

I don't think that "Statistically AP is safer than unaided driving" is only a Tesla thing - many marques have varous levels of AP tech in their cars (ICE or EV), and based on my personal experience this statement seems both reasonable and plausible. Why is this surprising?

As far as I am aware, accidents while using AP are mostly the result of having an idiot at the steering wheel.... and not the computer's fault. The cases of "unintended acceleration" are very rare, an even so not yet proven.

 

[Bobby Dazzler]

Out of interest, when using cruise control, do MBClubbers keep their foot on the accelerator pedal (without applying pressure) or do you move your foot to a completely different position?

 

[TrueSpirit]

Out of interest, when using cruise control, do MBClubbers keep their foot on the accelerator pedal (without applying pressure) or do you move your foot to a completely different position?

Usually used to keep my foot on the floor but then depending on what is unfolding in front of me on sometimes crazy motorways the right foot may hover towards either the accelerator or the brake!

 

[ChipChop]

I don't think that "Statistically AP is safer than unaided driving" is only a Tesla thing - many marques have varous levels of AP tech in their cars (ICE or EV), and based on my personal experience this statement seems both reasonable and plausible. Why is this surprising?

As far as I am aware, accidents while using AP are mostly the result of having an idiot at the steering wheel.... and not the computer's fault. The cases of "unintended acceleration" are very rare, an even so not yet proven.

Being progressive in outlook is all well and good but when the march of car safety tech, exemplified by Tesla's AP, is forecast to not bring
improvements in (US) road safety carrying on regardless is a puzzling position to take. Cognitive dissonance?

Reminds me of the touted air quality benefits of the expanded ULEZ which the Jacobs report firmly undermines. Yet progressives
carry on loving ULEZ zones.

As for always blaming the driver that rings rather hollow when you would think the primary consideration when designing driver assist features is to keep drivers engaged in the task of driving.

 

[markjay]

Being progressive in outlook is all well and good but when the march of car safety tech, exemplified by Tesla's AP, is forecast to not bring
improvements in (US) road safety carrying on regardless is a puzzling position to take. Cognitive dissonance?

I think that there's a misunderstanding here.

Tesla's AP tech won't make a dent in US road safety stats, not while just under 1% of cars are Teslas, for sure.

But AP tech in general, will make a difference, as more and more new cars (of all makes) are now equipped with this technology.

 

[ChipChop]

But AP tech in general, will make a difference, as more and more new cars (of all makes) are now equipped with this technology.

That is not what is forecast. +12% in the US is not an improvement.

 

[markjay]

Perhaps in the same way that road safety in Europe wasn't transformed when MB invented the ABS, but it was transformed when everyone started fitting it, and finally becoming mandatory on all new cars.

 

[markjay]

That is not what is forecast. +12% in the US is not an improvement.

"+12% in the US is not an improvement" - isn't it?

 

[ChipChop]

"+12% in the US is not an improvement" - isn't it?

12% more road accidents forecast by 2029.