Hacker Newsnew | past | comments | ask | show | jobs | submit | toxik's commentslogin

How about smaller buses then? This is quite common in the Nordic region it seems.


They are required by law in no small part because car manufacturers want it to be. Compliance is a moat.


Rearview cameras are effective: https://www.iihs.org/research-areas/bibliography/ref/2130

I agree with having simpler SKUs, but rearview camera is not where to start


The problem isn't that they're effective. It's that they're a regulatory solution to the complaints that the same demographics had 20yr ago (it's too easy to back a big fashionable in the 00s SUV over a kid) and as a result of it now all cars have crap rear visibility because there's no reason to be good when you have the camera.


The primary factor that correlates with reverse cameras reducing backup accidents is age - people over 70 have higher backup accidents rates without cameras/sensors. FTA:

> When averaged between the 2 automakers, effects were significantly larger for drivers 70 and older (38% reduction) than for drivers younger than 70 (1% increase); effects were significant for older but not younger drivers.

A big SUV is probably an exacerbating factor, though.

Also, for any kind of car, rear cameras and sensors decrease impacts while parallel parking. I see far fewer damaged bumpers on newer cars these days.


Visibility has gotten worse in many vehicles as a crash safety thing. Rear visibility is so blocked because the "beltline" of cars has moved up as crash standards get more stringent. A car that has a small rear window and high 'beltline' will do better in a crash.


TBH, I'm hoping we have front-view cameras that maybe kick in at under 20kph or something.

Front visibility is famously poor on SUVs and trucks, and even aside from pedestrians, I suspect there are a lot of small but very expensive bumper taps because you mis-judged the distance to the crap at the back wall of your garage.


Funny story about this.

There was a woman who backed over her own kid in the driveway. For some reason, she was not imprisoned for vehicular manslaughter. So instead of not being in prison, she spent the next half decade lobbying congress to make backup cameras mandatory. And it happened. So now everyone's car costs $3k more.

It would have been cheaper to put her imprison than impose a $3k cost per every car sold in America since 2018.

Lots more people need to be imprisoned for manslaughter, and lots of people need their license taken away for "backing crashes".


https://www.cars.com/articles/lawmakers-to-jump-start-backup...

> A 2012 Harris poll suggests that the public agress with the mandate despite the technology’s costs. NHTSA says adding a backup camera to a car without an existing display screen will cost around $159 to $203 per vehicle, shrinking to between $58 and $88 for vehicles that already use display screens. The Harris poll found that consumers care more about safety features like backup cameras than they do about multimedia systems.

I'm not sure where you're getting your $3k backup cameras from; the camera is a $30 part, and pretty much every new car has a screen in it already.


People don't understand and appreciate additional costs until they actually have to pay them. You can see this play out over and over again with additional tax increases jn response for new and improved public services - or customers asking businesses to do a "Made in America" product line, but then not putting their money where their mouth is and actually paying the upcharge for a MiA product.


I’d say 99% of drivers care about backup cameras more than about how hard it’s to replace spark plugs.


ill never appreciated paying an additional costs


>So now everyone's car costs $3k more.

$3000 for a backup camera, okay.


The screen and the camera is $3k

There's no reason for the screen other than the camera, therefore the camera is $3k


Source? A cell phone can be made much cheaper and has all the required parts.


A similar argument could be made for any safety feature that adds cost to vehicles--literally, all of them. If a death is preventable and adds on a relatively inconsequential amount to the cost of a vehicle, then it is the morally correct choice optimize for safety.


The logic doesn’t scale. You can’t impose arbitrary and subjectives thresholds to gloss over this fact. The obvious conclusion is that safety is one of many moral factors to balance.


You know prison sentences won't save any children, right?

And taking away licenses is acting too late.


You do not think a mother killing her own child is punishment enough? It's very unlikely she intentionally planned to kill her child in this manner to cover it up as an accident.

Besides backup cameras have use beyond just making sure a child is not behind you, such as assisting with parking, or seeing if there is oncoming traffic when there is a larger vehicle parked next to you.


I’m sure that factors in, but let’s not pretend that safety is also major contributor


I often wonder what it's like to think everything is some grand corporate conspiracy.


I often wonder what it's like to put words into other people's mouths.


I think it's just a matter of style. It comes across as pretentious.


I think people often forget about correlations between literacy and class, especially in the past. All of those fancy 19th century high school curricula teaching Latin and Xenophon were for the sons of gentry.


What in Rube Goldberg is this setup?


Heath Robinson, in fact, as I’m British.


My lights are like this, I have a remote placed on top of each traditional switch. I never use the remotes, because most of it is automatic (e.g., dim lights for movies, turn off lights when we leave, turn on lights when we come home). Actually I do use my phone to switch to the "put child to bed" lighting.


The departure times dominate latency and throughput in metros. The gates are not the bottleneck.


When a full train empties out at a specific station you can get massive delays. Euston platforms 8-11 come to mind. Two arrivals of 600+ people (including standing) trains in a minute or so in say 8 and 11 can cause chaos.


It depends. Usually you'd be right, but for some big events, the stations and platforms can be incredibly packed. In those cases the extra delay from gates could really hurt. One example is Comiket, where you have thousands of attendees all coming to the same few stations around the venue. Both times I was there, there was a massive crowd spanning from the platform to the outside. Having to wait the extra few hundred milliseconds on each card tap would have been painful.

Here's an example video to show the gates in action: https://www.youtube.com/watch?v=YffjxN3KsD4



I could use my European smartphone (well, smartwatch) as an IC card in Japan. I don't think it was slower.


If you're alluding to an Apple watch, it has Felica support wherever you buy it.

PS: For anyone in doubt https://atadistance.net/2017/09/12/iphone-x-keynote-global-f...


It also works on iPhone, with built in wallet support for Suica.


It depends on the vendor and whether they are willing to pay for global licensing. For Garmin devices for example, only the APAC version have NFC-F support.


There is, and what's worse, it assumes a global lock will keep things synchronized.


Does it? The GIL only ensured each interpreter instruction is atomic. But any group of instruction is not protected. This makes it very hard to rely on the GIL for synchronization unless you really know what you are doing.


AFAIK a group of instructions is only non-protected if one of the instructions does I/O. Explicit I/O - page faults don't count.


If I understand that correctly, it would mean that running a function like this on two threads f(1) and f(2) would produce a list of 1 and 2 without interleaving.

  def f(x):
      for _ in range(N):
          l.append(x)
I've tried it out and they start interleaving when N is set to 1000000.


It is guaranteed that the models will become salespeople in disguise with time. This is just how the world works. Hopefully competition can stave it off but I doubt it.

It's also why totalitarian regimes love it, they can simply train it to regurgitate a modified version of reality.


Are there any samplers that aren't basically greedy? I.e. actually searches the tree. I realize it's an absolutely insane branching factor and quite expensive to expand nodes at that, but it always seemed odd to me that we don't actually search.


Besides beam search and it's variants? (there are many including the little known but awesomely powerful constrained beam search: https://huggingface.co/blog/constrained-beam-search)

Does MBR (minimal bayes risk) sampling count?

Also there was this paper at ICLR which is relevant to this question: https://arxiv.org/abs/2410.03968

This paper basically claims that non-heuristic methods (like beam search) are harmful compared to the heuristic ones.


Beam Search sampling is sometimes getting used


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

Search: