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

271–280 of 290 posts

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

#271

Post by midoridensha »

bergenty wrote: post #234
midoridensha wrote: post #40

Earlier quoted context omitted.

American drivers are simply too stupid to be able to comprehend curved mirrors, so the US government bans them on US-spec cars, and instead you get flat mirrors with a warning message.

That’s a ridiculous hot take. What’s so hard to understand about curved mirrors? The mirrors already have a warning on them that objects are closer than they appear.

The EU mirrors aren't curved consistently; they curve more towards the outer end (IIRC), to eliminate the blind spot and show more. Apparently this is too much for the brain of an American driver to comprehend. If you disagree with this, you'll have to take it up with the federal government, because this is their opinion.

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

#272

Post by fastball »

hansvm wrote: post #232
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?

Even assuming one "event" per mile that needs to be handled satisfactorily (pothole, pedestrian, inconsistent embedded reflectors, chipped or graffitied sign, ....), 99.9999% is about the same reliability as the rate that human drivers crash in the US, and fatal crashes are around 100x less common. If, as in this comment thread, self-driving cars are being proposed as a solution to the deaths caused by human ineptitu…

Well in terms of fatalities, FSD actually has a 100% reliability rating, so presumably it's ok for it to be deployed, right?

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

#273

Post by hansvm »

fastball wrote: post #272
hansvm wrote: post #232

Earlier quoted context omitted.

Even assuming one "event" per mile that needs to be handled satisfactorily (pothole, pedestrian, inconsistent embedded reflectors, chipped or graffitied sign, ....), 99.9999% is about the same reliability as the rate that human drivers crash in the US, and fatal crashes are around 100x less common. If, as in this comment thread, self-driving cars are being proposed as a solution to the deaths caused by human ineptitu…

Well in terms of fatalities, FSD actually has a 100% reliability rating, so presumably it's ok for it to be deployed, right?

Not really:

- All FSD miles put together are still only 1/3 of the average miles needed to generate a fatal crash using a human driver.

- The vast majority of those miles are in ideal conditions.

- A cursory glance indicates there have been 15 FSD-induced fatalities.

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

#274

Post by alkonaut »

dominotw wrote: post #114
alkonaut wrote: post #107

Earlier quoted context omitted.

Luckily most new cars from the last 5-10 years should have blind spot warnings (even as standard on premium cars) so should be a problem that's at least slowly being solved.

They tell you to not rely on it in their manual. So you still have to check manually.

A US manual or hazard sticker would tell you not to forget your head outside the car when slamming the door shut too. I don't think manuals directed towards US customers are in any way related to safety, just legal fear. I mean it's the place where a coffee cup can have the text "caution, hot!".

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

#275

Post by fastball »

hansvm wrote: post #273
fastball wrote: post #272

Earlier quoted context omitted.

Well in terms of fatalities, FSD actually has a 100% reliability rating, so presumably it's ok for it to be deployed, right?

Not really: - All FSD miles put together are still only 1/3 of the average miles needed to generate a fatal crash using a human driver. - The vast majority of those miles are in ideal conditions. - A cursory glance indicates there have been 15 FSD-induced fatalities.

- Fair point, guess we will see in the next 60m miles, which by the sounds of it will be in the next few months.

- Tesla's FSD in particular (being an entirely vision-based system) likely operates in less-than-ideal conditions much more often than competitors.

- Where are you finding the "15 FSD-induced fatalities" stat?

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

#276

Post by shellfishgene »

snovv_crash wrote: post #170
qwerty456127 wrote: post #145

Earlier quoted context omitted.

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…

Metamizol (Novalgin) is probably one of the biggest examples of this I know. It's a non-NSAID, non-addictive painkiller that was discovered over 100 years ago and banned in the USA in the 60s. It has a similar danger profile to Diclofenac (Voltaren), which is now prescribed like candy. And it is still widely used in Germany and Switzerland, so there is lots of data on 'modern' populations. But since nobody can patent…

While it's widely prescribed in Germany, it is actually pretty restricted, too, but there seems to be a lot of off-label use. It's only supposed to be prescribed for strong pain, or for fever only if other medications don't work. There is a risk of agranulocytosis, which is low, but the reason for the restriction in various countries.

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

#277

Post by rad_gruchalski »

vladvasiliu wrote: post #95
nraynaud wrote: post #88

Earlier quoted context omitted.

when I was a kid in France I remember one some car parks at the beginning of the winter there were people tuning headlights with a machine. I don't know the specifics because I was too young (was it the police? a road safety non-profit?) It has disappeared a long time ago. edit: https://valdeuropeinfos.fr/faites-verifier-vos-phares-gratui... actually it still exists, and it's the police (a technical department dedica…

That may still exist, but I think no one bothers to do that around these parts. Basically, every other car has at least one headlamp pointing up. And I say this as a motorcycle rider, so my eyes are somewhat higher than in a regular car (I can see over the top of cars if they're not vans/SUVs).

On the subject of motorcycle riders. In Europe it’s so common for them to ride on main beam during the day. They always blind everybody around. I guess they do it to be better visible? Guess what, if you are riding on the high beam, be sure I am not looking in your direction.

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

#278

Post by vladvasiliu »

rad_gruchalski wrote: post #277
vladvasiliu wrote: post #95

Earlier quoted context omitted.

That may still exist, but I think no one bothers to do that around these parts. Basically, every other car has at least one headlamp pointing up. And I say this as a motorcycle rider, so my eyes are somewhat higher than in a regular car (I can see over the top of cars if they're not vans/SUVs).

On the subject of motorcycle riders. In Europe it’s so common for them to ride on main beam during the day. They always blind everybody around. I guess they do it to be better visible? Guess what, if you are riding on the high beam, be sure I am not looking in your direction.

I agree with this. Maybe they think they're more visible, but they don't understand that when they're blinding you, it's much harder to judge the distance...

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

#279

Post by maxerickson »

Retric wrote: post #256
sgtnoodle wrote: post #228

Earlier quoted context omitted.

Any fixed timespan chosen to analyze risk is arbitrary, and you can convert between timespans using probability math. I'd say it's typical to choose the average "mission" length, because that's the most relevant to the purpose. For a self driving car, an hour would make sense. For the first stage of a rocket, 5 minutes. You don't ride in a car appreciating every 5 seconds that you haven't crashed yet, do you? Maybe i…

The average American only drives ~35 miles per day. 1 hour may represent the average time driven per day but it’s much longer than the average trip length. Which doesn’t mean it’s a bad point of comparison but distance seems like the more common metric. Failure intervals generally use multiple different metrics depending on context. You see statistics for fatalities per year or per passenger mile as both being releva…

Decision events probably correlate better with distance than time.

Like your self driving car shouldn't be meaningfully more reliable if it slows down by 5 mph (because if it is the operating parameters are poorly chosen!).

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

#280

Post by webmobdev »

TomSwirly wrote: post #173
snovv_crash wrote: post #170

Earlier quoted context omitted.

Metamizol (Novalgin) is probably one of the biggest examples of this I know. It's a non-NSAID, non-addictive painkiller that was discovered over 100 years ago and banned in the USA in the 60s. It has a similar danger profile to Diclofenac (Voltaren), which is now prescribed like candy. And it is still widely used in Germany and Switzerland, so there is lots of data on 'modern' populations. But since nobody can patent…

To be fair, the Netherlands has very tight controls on metamizol as well, and for no clear reason except small-c conservatism, and also a tremendous reluctance for doctors here to give out painkillers. No, I have never asked for painkillers myself, so it isn't personal! However, if God–forbid I came down with pain issues that were treatable by painkillers, you better believe I would be on the dark markets PDQ. (My RS…

> My RSI is no fun, but I deal, and it isn't amenable to painkillers except in mind-numbing doses.

Just a warning - trying to ignore the pain (by bearing it or hiding it with painkillers) is what is likely making your RSI worse. Please try and address the actual cause of the pain, than just taking pain killers.

Post reply on HN