Comment by saghm | Hacker News Reader

Comment by saghm | Hacker News Reader

Parent Full thread

> A major theoretical discovery of K="computers can't do XYZ" would be another such thing for all of us who in the modern world chase these fads about computers reinventing finance or art or the modern economy. You'd have people in the year 2100 saying "can you believe only 75 years ago, everyone would just believe that computers were magically going to fix X and then they wouldn't and everyone would suddenly believe they would now fix Y and then they wouldn't and then Z and they wouldn't and and and... straight faced rationalist businesspeople acting like 'Tech-Jesus is coming' on August 1st, oh sorry it's September 1st, sorry -- October 1st? November? Maybe in the coming year? well SURELY this decade."

This might be a higher-level thing that you mean given that I'm not sure it's something "provable" or "disprovable", but I'm constantly amazed how incredulous people are that I don't expect any car to be fully self-driving (like in the sense that there wouldn't even need to be a steering wheel and someone without a license could ride in it alone) any time in the next few decades. To be clear, I'm not claiming that it's something that _definitely_ won't happen, but I don't think it's super likely, whereas the vast majority of people I share this opinion with (even people with similarly high levels of technical expertise) seem to think that it's almost a foregone conclusion that this will be solved within maybe 30 years.

Have you experienced any of the current driverless taxi services?

Personally i see the current limits to rollout as ones of hardware cost, but this is steadily declining.

There are only 3 cities in the entire US with generally available driverless taxis. All three of which are in sunny areas with generally no real inclement weather.

At this point, driverless taxis are more science fiction than reality for a supermajority of Americans, let alone the rest of the world.

Meanwhile up in Canada, I'm still struggling with the fact that google has completely mislabeled two of the streets that touch my corner lot, and invented a through street that has never existed on my neighbors property. All of this is information that I have submitted a correction to, as well as being available through our excellent public GIS system.

The existence of these sorts of ambiguities in the map, and in the real world means that driverless taxis will be doing exactly what the poor pizza guy does periodically and wandering down my dead end road thinking they are on a completely different street. Luckily the pizza guy has the ability to knock on my door and get directions, while a driverless car will just... turn on its flashers and block the road as far as I can tell.

When driverless cars can operate in Salt Lake City, a city that is incredibly tech friendly, low regulation, car centric, but with genuinely bad weather I will be impressed. In the meantime I would settle for more frequent bus service.

from people who work in those companies:

those cars could be in every city and they’d run great (except snow. snow is tough, rain not so much)

but why aren’t driverless cars there?

It’s been decided you want a perfect track record, as close to flawless - to appease regulation bodies.

regulation is what kills and regulation is the biggest fear. The cars and tech are always getting better, the ‘slowness’ is deliberate, the bet is on the future.

> It’s been decided you want a perfect track record, as close to flawless - to appease regulation bodies.

This is why I feel China will be the first to have unlimited self-driving. They would be able to treat self-driving cars the way that we treat trains: liability on the person hit by the train, so long as the train was functioning properly.

Maybe the Chinese self-driving cars will be programmed to kill any pedestrians they accidentally hit. Apparently it's cheaper that way.

https://slate.com/news-and-politics/2015/09/why-drivers-in-c...

According to the DOT over 70% of US roads are in areas that receive snowfall. (https://ops.fhwa.dot.gov/weather/weather_events/snow_ice.htm)

The inability to operate in conditions where people are most likely to not be able to use active transportation seems like a pretty major flaw.

People who work there will of course claim that. Maybe they should demonstrate it.

> All three of which are in sunny areas with generally no real inclement weather.

This isn't an accurate description of SF, but more importantly, the sun doesn't come out at night.

In terms of inclement weather for driving purposes, SF is more similar to Austin than say, Seattle.

My point wasn't that sunshine specifically helps, but that the complete lack of snow, ice and other roadway altering weather conditions is a serious asterisk that needs to be solved before this technology can be moved out of Beta for most of the US.

Self_driving cars seemingly can't handle snowy conditions. 70% of the roads in the US receive at least five inches of snow per year. (https://ops.fhwa.dot.gov/weather/weather_events/snow_ice.htm)

Yes they may not be ready for this decade, but the arguments against remind me of those that said nobody would ever have a computer in their home. Even some stupid connected lightbulb have more computing power than some of the machines at that time...

Don't trust anybody that believe they can predict the future.

Reply on news.ycombinator.com