4 ms·
This is perhaps a silly question from a pedestrian/driver, but why exactly is this bad for cyclists? From the article: > It is really surprising if you go on t
by jeffchien 5y ago
This is perhaps a silly question from a pedestrian/driver, but why exactly is this bad for cyclists? From the article:
> It is really surprising if you go on the road and see something foreign, and it’s coming at you at 15 miles per hour.
If it's going down the bike lane on the wrong way then I agree, that's bad and should be corrected.
There's also this:
> Some cyclists remain uncertain about exactly how a REV-1 would move out of a cyclist’s way
Would most of the safety objections go away if it just went faster and/or became easier to pass (i.e. smaller width, stay to the right side)?
- addicted 5y agoHow are you supposed to pass this non human powered vehicle in a bike lane? And if I am biking at 5mph, will it stay behind me?
- jeffchien 5y ago> How are you supposed to pass this non human powered vehicle in a bike lane? If I bike at 15mph (and other cyclists are faster), what is supposed to happen? Do I stay to the right, or if the bike lane is too small, get out of the bike lane and onto a ramp when safe? If the bot can do that (like a rear-facing LIDAR), then would the bot be more bearable? > And if I am biking at 5mph, will it stay behind me? Good point. I hope they have solved or will solve that.
- britch 5y ago>> It is really surprising if you go on the road and see something foreign, and it’s coming at you at 15 miles per hour. > If it's going down the bike lane on the wrong way then I agree, that's bad and should be corrected. 15 mph is close to top speed on a bike on even ground (for me at least). If I'm causally biking to dinner, I'm not going that fast. The robot could be "coming at you" from behind. ---- Beside the risks of the robots malfunctioning, I think this would be bad for bicyclists if it goes well. Right now the bots are pretty rare but let's say this is a huge success and a really cheap way to deliver food in a city. I can imagine there being stretches of bike lanes where the right-lane is a bunch of delivery-bots. That does not leave much room for bikes in the bike lane. "Bike lanes" become "delivery lanes". Whether or not this is a likely outcome is another discussion but making it a rule that robots can use the bike lane could have bigger consequences than this trial run. edit: bikes go faster than 15mph
- alanbernstein 5y ago15mph is not close to top speed on a bike.
- britch 5y agoYes, 15mph is not the theoretical max speed of a bike. For me, on my $300 Trek hybrid, traveling on flat land, that's about as fast as it can go. I'm sure someone in good shape on a nice road bike can go faster.
- hytdstd 5y agoFor a cyclist sitting upright (CdA of .5) but with good tires and pavement (Crr of .0035), it takes about 120W to hold 15mph on flat terrain. I would agree that this is about the limit for most casual riders. Serious riders with more power typically don't have a problem riding around other bike lane users. What I'm saying is, I don't think the robots are burdensome for the average bike lane user, as far as speed is concerned.
- mdasen 5y agoThis isn't a silly question, but I think it has a lot of answers. It's bad for cyclists because the self-driving robots are probably buggy. Having a buggy vehicle hit you from behind and push you into a truck driving in traffic that pulls you under and kills you is a very bad outcome. That might sound extreme, but we're seeing Teslas kill their occupants (and potentially other road users). Do we really think "Tesla's AI will drive into a solid barrier, but I'm sure the pizza robot won't bump a bike into traffic"? Another problems is that it's very hard for a machine to replicate what humans are expecting right now. One of the problems with self-driving cars is that they're often overly cautious in the presence of uncertainty (or certainty for a human, but uncertainty for the AI). If one of these pizza robots suddenly stops short with no reason to be doing that, it can cause accidents. We've seen self-driving cars cause accidents that "aren't their fault" by normal road-fault standards, but are kinda their fault. Stopping in weird ways, making odd judgements about what to do, etc. might lead to accidents. > Would most of the safety objections go away if it just went faster and/or became easier to pass (i.e. smaller width, stay to the right side Realistically, no. You can't really make them smaller in width. You need to be able to hold a 16" pizza and you're going to need at least another few inches on either side at a minimum. 22" is probably as small as they can go. Have you ever tried to ride side-by-side with another bike? It isn't realistic on a street with traffic. The objections would probably go away if we were willing to dedicate a 10-12' lane in each direction to vehicles around 15MPH (with a curb that cars couldn't cross). Then there would be room to avoid problems. The issue right now is that most often bikes don't have any lane, the lane they have is often obstructed by double-parked cars, delivery vans, cones, etc., and even good lanes are narrow (think 3-4 feet). There isn't room for a self-driving vehicle that is likely buggy and likely to make odd decisions trying to get around a delivery van blocking its path - decisions that might endanger bikers. One thing I'd ask drivers to remember is how angry they get when they're inconvenienced on the road. If a delivery truck stops in a car travel lane, drivers lean on their horns. If you live in a city, remember the last time someone was looking for a parallel parking space while you were trying to get somewhere. Imagine self-driving cars just kinda randomly stopping, sometimes too quickly when there's no reason to stop causing you to run into the back of them - and then you're at fault for it! Heck, just imagine them holding up traffic because they got confused. Heck, just remember the last time someone was in the wrong lane, but wanted to make a turn. Maybe you're an awesome and chill person, but you've definitely seen drivers who aren't. Biking is wonderful, but it's not well-supported in a lot of places and our roads have become more hostile in many ways. Drivers are choosing larger, heavier vehicles with higher hoods which are many times more likely to kill cyclists and pedestrians in the event of a crash. More and more people are watching all sorts of entertainment in their vehicles and taking their eyes off the road to fiddle with touch screens. Pedestrian deaths are rising in the US...And now it's like, "hey, since we've been doing so much for you lately...um...I can't recall any of those things...maybe we could throw some lightly-tested AI-driven pizza bots at you...we certainly hope no one dies...but if anyone does, we won't hold anyone responsible for those deaths...have fun!" As an aside, I think one of the big things is that no one will really be responsible for these pizza robots. If you kill someone with your car, at the very least your insurance is likely to get hit hard. You'll likely feel bad about it. There's the possibility that you'll go to jail for several years. With these pizza robots, if they kill a cyclist, what happens? A corporation loses an amount of money that they don't even notice? Except that they have the lawyers to make sure they don't even lose that?Would the company be willing to stipulate that they take responsibility for any accident involving the pizza robots (regardless of any determination of fault) and that any death requires 10% of shares to be assigned to the victim's family? Would they be willing to risk their business on this? On a more realistic level, who is criminally responsible for any negligence that the pizza robots might do? If you're negligent driving, you're responsible. You can be put in prison. Do we put the CEO of the pizza company in prison? The software engineers? It seems like there's little incentive for the pizza robots to act well in the bike lanes. Why not try to monopolize them and push out cyclists? Why not park in them? If you're developing this delivery-robot tech, being a bad actor in the bike lanes potentially means that you'll get to privatize that part of the road for your delivery-bot fleet. That might sound absurd, but it happened with cars. Streets used to have people in them, but as it became clear that the new cars didn't mix with people well, the people got pushed out. Heck, street cars didn't mix well with cars (because they stopped frequently to let passengers on/off) and cars hated getting stuck behind them and having to go around them so we mostly removed those. As a biker, these pizza-robots feel like someone saying "yea, you only have a scrap of the road, but we want you to share that scrap." I don't know where you live and things can vary a lot based on geography in the US and it's certainly not a silly question. In my city, I can see these pizza robots swerving unexpectedly to avoid road imperfections, stopping randomly, getting confused about turns, potentially trying to make an L-turn (going straight on the green and then turning/orienting left on the other side and waiting for the green the other way) and blocking the bike lane going forward, just plain not identifying a bicycle and hitting someone, having issues with rain or other weather, or just stopping in the bike lane while waiting for the pizza-recipient to retrieve the pizza since there's no place to pull over. Realistically, I think the question might be something to think about in reverse: why would sharing a road with self-driving pizza-cars be annoying for cars. Well, it would get to its destination and then park in the traffic lane blocking traffic for two minutes waiting for the person to collect the pizza; it would have trouble understanding certain situations and cause you to rear-end them - and the company would use its lawyers to make sure you were deemed at fault to protect the reputation of their new pizza-cars; etc.