Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

because other drivers and pedestrians on public streets have a very strong interest in knowing how safely these cars perform and how they glitch out given that they could cause accidents at any point?


How's that different from any human driver on Saturday night?


Human drivers know what to look out for in other human drivers, we don't know what to look out for in autonomous vehicles. If those vehicles perform particularly badly under some conditions that are not obvious the public would benefit from having transparent insight into how the cars function, so they can be appropriately alert.

A research experiment that exposes the public to additional risk should be performed with the maximum amount of transparency possible. It is frightening that this even needs to be spelled out.


Human drivers have their own low-hanging fruit that can be pointed at for accident reduction. Tired, impaired, distracted driving are horribly commonplace.

Its a low bar, for automated cars to be better than human drivers. Sure they'll have their 'blind spots'. That's no condemnation of the whole industry. Because what we have now is not very good at all (fallible humans, all different). And when automated drivers have an issue we discover, they can all be fixed. Try that with humans.


It’s a low bar for automated cars to be better than tired, impaired or distracted drivers. It’s a very high bar to be better than all drivers.

Even tired, impaired, or distracted drivers still behave in semi-predictable fashions - they tend to overreact.


That's why these vehicles have both a backup driver and workers monitoring remotely from the office...

No one is saying entirely driverless vehicles are ready yet. Even some of the earlier hawks who claimed next year have backtracked.

Nothing wrong with that, all software deadlines are usually 1.5-2x longer than initial optimistic estimates.


As the article lays out waymo has accurate data on where the cars glitch out. Take that data, put it onto a public map, make that available for resdients to browse so they have accurate, data driven insight into what they need to look out for when they encounter autonomous vechilces. There is absolutely no reason not to do this, the amount of effort is trivial, and it would enhance the safety of people living in those areas.

If waymo wants to have the privilege of secrecy they can run experiments somewhere not open to the public. That should be the standard we apply to these companies.


But … we have Saturday Night human drivers, and we have no such rule for them. No complex system of reporting maps of bars and festivals and when to avoid them. Doesn't seem fair, somehow.


You're spreading fud. Waymo have a lot of vehicles on the road and most of the time the human drivers run into the back of them while they're stopped. They haven't had any big incidents.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: