Jan and Klodi's Party Bus - part II off topic discussion

Effects · 19-03-2018 11:40am

Ah ok, I guess as it was attacking cyclists I automatically thought it was After Hours!

Wishbone Ash · 19-03-2018 4:34pm

Foley's tea rooms in Castlebellingham burnt to the ground today.

https://www.facebook.com/foleyscastlebellingham/

magicbastarder · 19-03-2018 6:31pm

i know this has nothing specifically to do with cycling, but we've just had the first fatality involving an autonomous car:
https://www.theguardian.com/technology/2018/mar/19/uber-self-driving-car-kills-woman-arizona-tempe?CMP=Share_AndroidApp_Tweet

19-03-2018 6:37pm

Wishbone Ash wrote: »

Foley's tea rooms in Castlebellingham burnt to the ground today.

https://www.facebook.com/foleyscastlebellingham/

Ah no, was only saying to herself today we should take a spin out that way soon as we've not been that way since around this time last year.

Very sad it was a lovely thatched building and lovely people running it.

doozerie · 19-03-2018 7:08pm

magicbastarder wrote: »

i know this has nothing specifically to do with cycling, but we've just had the first fatality involving an autonomous car:
https://www.theguardian.com/technology/2018/mar/19/uber-self-driving-car-kills-woman-arizona-tempe?CMP=Share_AndroidApp_Tweet

It was initially reported that the victim was a cyclist, and I assumed this was a tragic demonstration of the challenges that still remain for autonomous cars when it comes to recognising cyclists.

But it has since been confirmed that she was a pedestrian, which arguably is even more concerning when looking at self-driving cars generally - various of the reports I've read on self-driving cars seemed to suggest that recognising pedestrians and avoiding colliding with them was almost a ticked box already, while this tragic event suggests otherwise.

Eamonnator · 19-03-2018 7:53pm

Wishbone Ash wrote: »

Foley's tea rooms in Castlebellingham burnt to the ground today.

https://www.facebook.com/foleyscastlebellingham/

https://www.dundalkdemocrat.ie/news/home/303332/fire-destroys-thatched-roof-of-popular-castlebellingham-tearooms.html

Terrible pity. Many's the feed, deserved and undeserved, I've had there.

manafana · 19-03-2018 8:08pm

magicbastarder wrote: »

i know this has nothing specifically to do with cycling, but we've just had the first fatality involving an autonomous car:
https://www.theguardian.com/technology/2018/mar/19/uber-self-driving-car-kills-woman-arizona-tempe?CMP=Share_AndroidApp_Tweet

this can happen, nearly all iv read suggest road fatalities would greatly reduce with self drive, however we find death as usual is acceptable but not when it comes from automation as we can't just accept it as human error

guanciale · 19-03-2018 10:01pm

Autonomous cars

I was at a presentation last year about autonomous vehicles given by a specialist Lloyds insurer who provides tech based insurance solutions.
They have 6 semi-auto to full auto vans driving in London and surrounds. They are ollecting data to understand how these vehicles interact with surroundings and vice versa. The use of the data is to allow them to ascertain how they might price fleet insurance for auto cars.

The vans they have on the road collect so much data that the trial had to be mothballed due to the firm (an already data savvy firm) being overwhelmed by the data (cost of storage/noise etc). They outlined the sheer number if interactions in a busy setting and suggested that modelling is not yet advanced enough to cooe with this.

rollingscone · 19-03-2018 10:16pm

guanciale wrote: »

Autonomous cars

I was at a presentation last year about autonomous vehicles given by a specialist Lloyds insurer who provides tech based insurance solutions.
They have 6 semi-auto to full auto vans driving in London and surrounds. They are ollecting data to understand how these vehicles interact with surroundings and vice versa. The use of the data is to allow them to ascertain how they might price fleet insurance for auto cars.

The vans they have on the road collect so much data that the trial had to be mothballed due to the firm (an already data savvy firm) being overwhelmed by the data (cost of storage/noise etc). They outlined the sheer number if interactions in a busy setting and suggested that modelling is not yet advanced enough to cooe with this.

Techbros no likey risk management framework! Techbros go sulk now!

Really interesting though.

eeeee · 19-03-2018 11:19pm

Really, really think we're a good bit away from safe self driving cars. Don't think the technology is near where it needs to be yet. There's also significant moral quandries to be sorted yet, think there's a website you can log on to answer questions to go into the database for the autonomous cars to help them make these decisions .

I fear the autonomous car more than the human driven one!

20-03-2018 12:07am

During trimming today I managed to completely mess up my beard so I shaved it off, and now my face is exposed to the elements and cold. My little girl got a bit freaked out, she's 4 now so last time I had a bare face she was 2 so likely doesn't recall me without one

Hey might make me more areo though right? right?

RobertFoster · 20-03-2018 12:24am

Deleted User wrote: »

Hey might make me more areo though right? right?

https://www.youtube.com/watch?v=FkmCLOjk2h8

<1 second over 40km

20-03-2018 12:32am

RobertFoster wrote: »

https://www.youtube.com/watch?v=FkmCLOjk2h8

<1 second over 40km

So the smart man grows some fur to keep his face warm in a long descent it seems, unless you ride for Israel Cycling Academy , they are anti beard :mad:

Lusk_Doyle · 20-03-2018 9:15am

As are a Dutch or Belgian lower level outfit. Can't recall the name.

buffalo · 20-03-2018 9:20am

magicbastarder wrote: »

i know this has nothing specifically to do with cycling, but we've just had the first fatality involving an autonomous car:
https://www.theguardian.com/technology/2018/mar/19/uber-self-driving-car-kills-woman-arizona-tempe?CMP=Share_AndroidApp_Tweet

The follow-up may not be as widely publicised, but here's the preliminary comments from the cops: https://arstechnica.com/cars/2018/03/police-chief-uber-self-driving-car-likely-not-at-fault-in-fatal-crash/

The chief of the Tempe Police has told the San Francisco Chronicle that Uber is likely not responsible for the Sunday evening crash that killed 49-year-old pedestrian Elaine Herzberg.

“I suspect preliminarily it appears that the Uber would likely not be at fault in this accident," said chief Sylvia Moir.

...she "abruptly walked from a center median into a lane of traffic."

magicbastarder · 20-03-2018 9:23am

thing is, that's precisely the sort of incident that self-driving cars were meant to save us from. granted, i have no clue how far away the car was when she stepped out, and the technology is far from mature. the car was reported to have been doing 40mph at the time?

ford2600 · 20-03-2018 9:59am

magicbastarder wrote: »

thing is, that's precisely the sort of incident that self-driving cars were meant to save us from. granted, i have no clue how far away the car was when she stepped out, and the technology is far from mature. the car was reported to have been doing 40mph at the time?

If you write an algorithm to slow/emergency stop on every potential pedestrian stepping out in front of car no body will buy the car.

As it is we kill about 40 pedestrians a year; pedestrian behaviour has evolved primarily driven by fear of traffic. That is even seen within housing estates where rather than traffic slowing we need to build ramps everywhere. Once we change to automated cars; which are always alert/within speed limit/not tired/not distracted it's likely pedestrian behaviour will evolve also.

Would you use an automated car in Johannesburg? Maybe they could fit a module to deal with hijacking.

Alek · 20-03-2018 10:17am

manafana · 20-03-2018 11:41am

nee wrote: »

Really, really think we're a good bit away from safe self driving cars. Don't think the technology is near where it needs to be yet. There's also significant moral quandries to be sorted yet, think there's a website you can log on to answer questions to go into the database for the autonomous cars to help them make these decisions .

I fear the autonomous car more than the human driven one!

in general id feel the predictability of an autonomous car easier to deal with, they will use indicators won't quickly change their mind about what lane they want be in, however i would have doubts in certain areas stills its an improvement, but yeah we need move away from letting big metal boxes dominate our built up urban areas.

buffalo · 20-03-2018 1:08pm

magicbastarder wrote: »

thing is, that's precisely the sort of incident that self-driving cars were meant to save us from. granted, i have no clue how far away the car was when she stepped out, and the technology is far from mature. the car was reported to have been doing 40mph at the time?

It's a contest of correcting the software vs correcting the process. Arguably the pedestrian should not have been jay-walking*, but the car should've been able to anticipate the possibility of a collision and slow to an appropriate speed where it could either stop or prepare for evasive action.

The question is whether we (as a society) say, "well, the software rigidly follows the rules of the road and only expects pedestrians to cross at a crossing, therefore if you don't use a crossing, you've only yourself to blame", or we decide "the software must allow for pedestrians randomly walking out from the footpath suddenly". The latter may include children who don't know the 'rules' or somebody stumbling. The former leads to the possibility of swathes of roads being declared danger zones for people on foot (and perhaps bike).

There are levels of nuance in between of course - in 30kph zones, the likelihood of pedestrians suddenly crossing the road increases, and the car's behaviour should compensate.

*itself a loaded term and leading to victim blaming

DirkVoodoo · 20-03-2018 1:20pm

buffalo wrote: »

It's a contest of correcting the software vs correcting the process. Arguably the pedestrian should not have been jay-walking*, but the car should've been able to anticipate the possibility of a collision and slow to an appropriate speed where it could either stop or prepare for evasive action.

The question is whether we (as a society) say, "well, the software rigidly follows the rules of the road and only expects pedestrians to cross at a crossing, therefore if you don't use a crossing, you've only yourself to blame", or we decide "the software must allow for pedestrians randomly walking out from the footpath suddenly". The latter may include children who don't know the 'rules' or somebody stumbling. The former leads to the possibility of swathes of roads being declared danger zones for people on foot (and perhaps bike).

There are levels of nuance in between of course - in 30kph zones, the likelihood of pedestrians suddenly crossing the road increases, and the car's behaviour should compensate.

*itself a loaded term and leading to victim blaming

The roads are full of unexpected situations though, we tell drivers/road users to always "expect the unexpected", if the software is incapable of this then it is not fit for purpose.

This seems almost like the disc brakes in the peloton issue: it doesn't look like we can have autonomous vehicles mixing with human drivers, pedestrians, cyclists. It's all or nothing.

Hopefully the future will involve autonomous super highways with EVs travelling safely at 200kmhr between cities, transporting goods and people, while the cities themselves become cycling and walking utopias, free of any vehicles, except for the delivery drones that occasionally fall out of the sky and squash some nuns/orphans.

I can dream, right?

doozerie · 20-03-2018 2:02pm

buffalo wrote:

It's a contest of correcting the software vs correcting the process. Arguably the pedestrian should not have been jay-walking*, but the car should've been able to anticipate the possibility of a collision and slow to an appropriate speed where it could either stop or prepare for evasive action.

The question is whether we (as a society) say, "well, the software rigidly follows the rules of the road and only expects pedestrians to cross at a crossing, therefore if you don't use a crossing, you've only yourself to blame", or we decide "the software must allow for pedestrians randomly walking out from the footpath suddenly". The latter may include children who don't know the 'rules' or somebody stumbling. The former leads to the possibility of swathes of roads being declared danger zones for people on foot (and perhaps bike).

There are levels of nuance in between of course - in 30kph zones, the likelihood of pedestrians suddenly crossing the road increases, and the car's behaviour should compensate.

*itself a loaded term and leading to victim blaming

There is a huge challenge in trying to accommodate people and machines/software in the same space.

At one extreme, every single autonomous car will be limited to a speed (in populated areas) that allows it to stop in time no matter how close it is to a pedestrian that steps in front of it. I’m not sure what speed that would be, 20kph maybe? And knowing that we are “safe” as pedestrians we’ll all just cross roads where and when we like, whether empty or filled with traffic, knowing that no car will hit us.

I’m not sure that would be such a bad scenario to be honest, no doubt the car lobby would be utterly distraught at the prospect and would fight it tooth and nail, but fcuk them, they’ve had it largely their own way for long enough that it would be good to see the tables turned for a change.

I’m not sure what it would mean for cyclists though. The attitude already exists that we shouldn’t be on the roads at all, the many appalling off-road cycle lanes were arguably a blatant attempt to work towards that (as opposed to being for the benefit of cyclists), etc., so it’s not inconceivable that autonomous cars would be used as another argument against having human-powered bikes mixing on the roads with “the future of transport”.

doozerie · 20-03-2018 2:03pm

DirkVoodoo wrote: »

Hopefully the future will involve autonomous super highways with EVs travelling safely at 200kmhr between cities, transporting goods and people, while the cities themselves become cycling and walking utopias, free of any vehicles, except for the delivery drones that occasionally fall out of the sky and squash some nuns/orphans.

I can dream, right?

You can dream of nuns, yes, but not of orphans, that's wrong.

dinneenp · 20-03-2018 4:08pm

Documenting a bike build with the Fujifilm X-E3 camera. This is from a photography website so the video focuses a lot of the camera.

photos

video

Wishbone Ash · 20-03-2018 11:06pm

17 year old cyclist in fatal accident in Bettystown this evening. RIP

https://www.independent.ie/irish-news/cyclist-17-killed-in-collision-with-minibus-36726525.html

niceonetom · 21-03-2018 9:44am

Re: autonomous cars. It seems to me that the cultural shift will be at least as hard as creating the technology. Whatever solution to the trolley-problem is eventually implemented, there's likely to be a huge disparity between the way the public react to humans inadvertently killing humans and devices killing humans.

We've become astonishingly inured to human mistakes. Something like a million people are killed on the world's roads every year, but we mostly treat this as a background fact of life. We're habituated to something truly horrific. Something monstrous.

I have no doubt that that number can be radically reduced by replacing distracted chimps with implacable code... as long as we don't stall the project by holding the machines to an impossible standard in the meantime. Every day we wring our hands about an single outlier incident, hundreds of people become victims of the status-quo. It feels like we have a moral responsibility to roll out autonomous system as soon as we can.

The very fact that autonomous mistakes are newsworthy but human mistakes are not is illustrative of something.

I'm sorry for making this point in the wake of another road death. R.I.P. to that person and all the others who died on the roads.

seamus · 21-03-2018 9:55am

niceonetom wrote: »

The very fact that autonomous mistakes are newsworthy but human mistakes are not is illustrative of something.

One day all road deaths will get as much attention as this one because of their rarity.

There is no doubt that machines can/will/do have two things that human drivers never will:

1. Far greater capability to recognise and react to potential danger
2. Consistency. A human driver may give extra space when they pass a parked bus this time, but may not next time. A AV will do it every single time.

One of the more notable side-effects of autonomous vehicles is that there will be less frustration with non-motorised road users. If an AV is stuck behind a bicycle for 30 seconds, the occupants won't be screaming "go go go, get out of the way". Part of the stress of being stuck behind a slower road user is having to try and figure out when and how you can overtake safely. AVs remove this stress.
Eventually most of the time the occupants won't even be paying attention to where they are or what the vehicle is doing. Like going on a bus, it'll speed up and slow down and just do its thing.