Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

The US will finally allow adaptive beam headlights on new cars

arstechnica.com

Post reply on HN
Print view View original

201–210 of 290 posts

Re: The US will finally allow adaptive beam headlights on new cars

#201

Post by spookie »

fallingknife wrote: post #181
bmitc wrote: post #136

Earlier quoted context omitted.

> it's clear that human behavior and existing road design is still vastly worse Is that true? Those human statistics take every human car and driver into account, not just Teslas. I have never seen data that shows self-driving cars anywhere close to human driving capability. If we replaced every car on the road with "self-driving" Teslas, it would be mayhem.

If we replaced every car, it would be an easy problem. The issue is having to replace one at a time while everyone else on the road is still human.

Or... Just make public transportation work? Is this just that difficult to get on the anglosphere or something?

I still cannot grasp what people are trying to do with this, really. Why are they trying to do public transports... With cars? You guys do realize most of accidents come from urban/chaotic areas, right? You don't have to use a car to solve this.

Re: The US will finally allow adaptive beam headlights on new cars

#203

Post by jlund-molfese »

dontlaugh wrote: post #164
jlund-molfese wrote: post #139

Earlier quoted context omitted.

If everyone lives right by the train station that’s fine. What about people in wheelchairs who live in the countryside and prefer to be dropped off at their doorstep? Or someone who can’t drive themself but could use an autonomous car or taxi to go to the hospital in an emergency situation (without waiting a while for the next open train)?

Most people already live near a train, tram or bus station. Or somewhere it would be worth building one anyway. Taxis already exist, just like trains they work fine. Personal cars will likely always exist, but only a small minority absolutely needs one. The bulk of human transportation doesn’t need cars.

Trains are great. They’re safer, more efficient, and often more convenient than cars. But it’s not either-or.

My comment was emphasizing that smart road technologies would be useful to many people also. There are many places where it’s not possible or desirable to build train infrastructure—these are areas that can be served by autonomous vehicles.

Re: The US will finally allow adaptive beam headlights on new cars

#204

Post by fastball »

Retric wrote: post #155
fastball wrote: post #137

Earlier quoted context omitted.

Do you need everything to work reliably 99.9999% of the time? Is that the reliability rating of human drivers?

Depnds on how you count. If you consider every ~5 seconds you could have an accident, and a minor or worse accident happens once every 2,000 to 20,000 hours of driving that’s 99.99993% to 99.999993% reliability. On the other hand if you start talking per trio then we aren’t even close to that.

Why in the world would you count in 5 second intervals? If the measure is accidents per hour you don't then divide it by 5s.

Reliability with 1 accident per 2k hours would be 99.95%.

Reliability for 1 accident per 20k hours is 99.995%.

Re: The US will finally allow adaptive beam headlights on new cars

#205

Post by fastball »

spookie wrote: post #197
fastball wrote: post #137

Earlier quoted context omitted.

Do you need everything to work reliably 99.9999% of the time? Is that the reliability rating of human drivers?

A single person might die, go to the hospital, or even go to jail depending on the accident. But if it's the car's software fault... Who gets blamed? Bear in mind we are trying to make this economically sustainable.

Presumably the company making the car/software/hardware – whichever part is at fault. If one company is making all three (Tesla) it's pretty easy.

Re: The US will finally allow adaptive beam headlights on new cars

#206

Post by jlund-molfese »

mschuster91 wrote: post #182
jlund-molfese wrote: post #139

Earlier quoted context omitted.

If everyone lives right by the train station that’s fine. What about people in wheelchairs who live in the countryside and prefer to be dropped off at their doorstep? Or someone who can’t drive themself but could use an autonomous car or taxi to go to the hospital in an emergency situation (without waiting a while for the next open train)?

> What about people in wheelchairs who live in the countryside and prefer to be dropped off at their doorstep? The important thing is to build a system that works for wide masses - even replacing half of individual car transit by rail and bus works wonders for space consumption [1]. The more complex cases - rural people, mostly - can be dealt with later on, e.g. by providing something like a "phone bus" service. Here…

> Saying "no" by default because a system is not entirely perfect at the beginning is a destructive discussion tactic that only serves those in power.

Who are these people in power, and what are their goals? I don’t see what this has to do with my comment—I’m not arguing against trains, merely in support of autonomous vehicles, in areas where they might be useful, like far away from cities.

Re: The US will finally allow adaptive beam headlights on new cars

#207

Post by danaris »

qwerty456127 wrote: post #145
prng2021 wrote: post #4

We allow companies like Tesla to jeopardize people's lives while the company beta tests "full self driving" technology, but are scared of allowing proven safe technology like adaptive headlights.

This is because the only place for progress in America seemingly is within yet-nonexistent unregulated areas. As soon as something gets regulated in the US it mostly is doomed to lag decades behind the rest of the developed world (because American policymakers apparently are not fond of updating existing laws/rules and the bureaucracy is not flexible enough, often prioritizing the letter of the law). Of course there…

A big part of the problem is our ongoing and increasing legislative gridlock. When one of the two major political parties has as their stated goal preventing the other from getting anything done, and undoing anything they do (leaving aside their open efforts to replace democracy with autocratic fascism), it's very, very difficult to make any meaningful progress.

Re: The US will finally allow adaptive beam headlights on new cars

#208

Post by danaris »

bb88 wrote: post #32
prng2021 wrote: post #31

Earlier quoted context omitted.

Yes drunk driving is a terrible problem but beta testing software on the open road is not the only option available. It's nothing but irresponsible. Here's one of many: use sensors and AI in the car to detect a drunk person behind the wheel. This is HN so I'm sure readers could spend 10 seconds and think of numerous ways this could be accomplished and many otherideas about what could be done when a drunk driver is de…

You may find it interesting the NHTSA is pushing for breathalizers in cars starting in 2026. Personally I'm more in favor of smart road like technologies where a highway has a mechanical or physical properties to assist self-driving vehicles stay in line with the road. Mostly because a 100% software solution will likely not work, particularly in the midwest and great lakes states during the wintertime. It need not be…

Any solution to self-driving cars (that purports to be a universal solution) that requires specific modifications to the roads is doomed to failure.

Even leaving aside all the gravel and dirt roads, and the many places where, at any given time, there's construction moving a lane 5 to 15 feet to one side temporarily, do you know how many miles of road there are in the US alone? I sure don't, but I know it's a staggering amount, a hell of a lot of it is rural (and some seasonal), and getting all of it retrofitted with even the most simple and low-tech self-driving guidance assist tech like your magnetic nails would cost vast amounts of money and either take a decade or three or cost about ten times as much.

Re: The US will finally allow adaptive beam headlights on new cars

#209

Post by potamic »

sethammons wrote: post #152

The bright-in-your-eyes headlights (oncoming or following you) are a problem, especially oncoming on curves. I learned early to follow the outside lane marker with my eyes, effectively taking my eyes off the road to avoid getting temporarily blinded. My wife suggested a fun tech solution: night vision and making all lights less bright. Kind of a fun idea: a slight glow or dim lights so you can still see vehicles unai…

You don't even need night vision. Your eyes adjust to low light and can actually see better when the peak brightness is low. It's ironical that all the solutions we come up with to aid night driving involve increasing the brightness, whereas we should be doing exactly the opposite.

I once drove on a stretch of highway that used reflective paint for the lane markers. My feeble beam could light them up better than any HID could. It was very relaxing to drive in the dark and never once did you have to strain to make sure you can see the road. I think this is really the future of night driving.

Re: The US will finally allow adaptive beam headlights on new cars

#210

Post by ericmay »

lotsofpulp wrote: post #190
ericmay wrote: post #183

Earlier quoted context omitted.

People drink and drive in America because they have no choice. You can’t walk him from a bar, and even if it’s available nobody is paying $40 for an Uber when they can drive their car. If we want to lower drunk driving incidence rates we need people walking to the bar or taking public transit. It requires us to fix our outdated and asinine transit culture. All of the technical solutions in the world can’t solve this…

>People drink and drive in America because they have no choice. You can’t walk him from a bar, and even if it’s available nobody is paying $40 for an Uber when they can drive their car. List of choices: 1. Not drink alcohol 2. Drink less alcohol, and switch to water at least 1 hour before driving. 3. Pay for a driver 4. Live where you can walk to an establishment that serves alcohol 5. Rent a hotel room nearby And so…

  1. Don't drink alcohol
Well yea but we're talking about people who do drink alcohol

  2. Drink less alcohol, and switch to water at least 1 hour before driving.
I think this is a good solution, though of course different factors can and do easily disrupt this.

  3./5. Pay for a driver/hotel
Hmm, I've only had a few and I feel pretty good. And an Uber is $50 just one way and I still have to pick my car up in the morning (if it doesn't get towed). Vast majority of people (hence why we have DUI and drunk driving problems) will choose to risk it.

  4. Live where you can walk to an establishment that serves alcohol
This is tautological w.r.t the OP. If you could walk to an establishment that serves alcohol you'd live in an area where you could walk, which would mean you didn't really need a car. But that's not how we build in America which is what precipitated my comment.

Related to #6 from the poster below - you can get a DUI for sleeping in your car while intoxicated, you can get woken up because you aren't allowed to sleep where you are parked, you can get towed, etc.

Post reply on HN