This is a property of the distribution of light timings across a territory, and is still not something to do with red-light cameras.

If his home town used 3-second yellows without traffic cameras, without excessive enforcement, he’s going to roll through anyway

He’s going to blow a red light? And you think that’s a better situation than him not blowing it?

In my country I believe all traffic lights have approximately three seconds. It’s enough; you need to have a good idea coming up to each light of when you’re going to disregard the light, so that you’ve done your decision-making before it changes. Then, if it changes just before that point, you need to brake positively, not leisurely. Somehow, all learner drivers manage it.

If you view amber lights as “stop as long as able” rather than “continue if able” you wouldn’t be in the embarrassing position of seemingly advocating purposefully blasting through red lights.

Huh, doesn’t render that way on my instance!

How dyou like them apples?

I had read it, obviously, and had no idea what point you are trying to make.

Guess we’ll never know!

Do you have glaucoma? You should easily be able to see the light change when not looking directly at it. And once you’re past your decision point, you don’t even need it in your field of vision.

I love getting contextless links to long articles. /s

The threadiverse is not a Linux stall.

If you slam on the brakes at high speeds on a modern car, you will stop in much less than 2 seconds. In practice you need some time to react as well, but the time to react to an anticipated event is short and should be about a quarter of a second.

The “state of the LEDs” that you keep referring to is the signal that tells other drivers they may proceed. Stop minimising it.

Not relevant to this discussion. You can google to determine that such countries exist, if that’s what worries you. Or just not believe me.

It’s explaining why the bullshit is left. If you don’t want to hear that, and instead just want to moan about how AI is not solving a problem that it literally cannot solve because it doesn’t have any physical existence, then that’s your choice, but it doesn’t seem a productive use of time, like complaining that your dishwasher can’t drive you to work. So yeah, I’m not going to talk about that, because it would be dumb.

AI hasn’t taken away any of the good parts of my life. I do more art and writing than I did 5 years ago. I play the same amount of music. It has impacted my job but it’s not yet clear how significant that will end up being.

Anyone should be able to stop their vehicle within 2 seconds. For an emergency, that is generous. For a traffic light, that is quick - but it is easily achievable for exactly the same reason. And, because you can see the green light, you are able to anticipate it changing, and so react quicker than you would to an unanticipated event like an emergency, so it is easier still.

Amber light timings in my country are, I believe, 3 seconds universally.

I’m saying that the thing that you said here is not relevant to what I said. It’s not a problem to the enforcement of amber lights.

Every driver should have an idea of their own stopping distance, regardless of different so-called definitions, so that they can drive at an appropriate distance from the vehicle in front (fat chance of that, given driving standards). If you are 3 seconds away from a light the moment it turns amber and decide not to stop for it, you went through when you were able to stop, and ought to get a ticket. A police car capturing video would be able to determine that objectively. Hence: objective enforcement is possible.

How short?

Eh. Pretty sure that aspect is trivial. I already have an automated system to determine which jobs need doing - it’s a task tracking app and it just puts them on rotation. The mental capacity it would take to think “does the bin actually need taking out” when prompted by the app is negligible compared to the capacity required to remember to think about the bin, and the physical effort and time required to take it out and replace the bag. So no, I don’t think what we lack is a way to determine “which jobs need to be done”. If an AI tool somehow could tell me that, it would make virtually no difference in my life. The task is the same every single time and so does not need any efficiency planning either.

The fact is that a robot which can successfully take the bag out of the bin, tie it off, carry it through the flat, down the stairs, open the dumpster and throw it in does not exist or, if it does, it’s still a research project not an affordable consumer product. If it did exist, it would save me time and effort every week. Whatever it is you’re talking about would save virtually nothing.

Laundry, to return to that, needs doing at a near constant rate, so there is usually no question of whether it needs to be done. Putting the laundry on is, again, the same task every time (maybe you want it to separate your whites, but the difficult task is the physical separation, not deciding how to do this). Hanging it out would be an interesting software challenge but, once again, if you don’t have a robot capable of hanging out laundry, I don’t need an AI telling me where to put each sock, something which gets me nothing.

The remaining things we can automate with household chores are almost all physical. The prospect of software being able to know exactly whether the bin needs taking out or the laundry needs doing is completely insignificant in comparison.

All articles on lemmy about the Online Safety Act seem to assert secret, conspiratory motives to its passing, without evidence.

I think the OSA is a privacy disaster and refuse to use age verification for that reason, but people: you need to wake up to the fact that most people don’t give a shit.

You:

they have it set up so it will use tools like Me:

It farms arithmetic out to dedicated, non-language based sub agents

Is there an echo in here?

Even if I interpret you charitably, you are using the word “it” to mean “LLMs”, instead of “AI”. That isn’t valid or useful. Nobody cares very much whether specific classes of neural network are good at maths; they care whether ChatGPT is going to give them the right answer. Go waste some of OpenAI’s money right now by asking it a question on mathematics, and see whether “AI can do maths”.

This gets even more ludicrous when you look at what AI is enabling in real mathematics, rather than in arithmetic.

What you’re asking for is not an ai but a robot. A robot that loads and unloads appliances does not need AI (though it may be the easiest way to get it to work now) because the tasks has to do at so restricted, but it does need manipulators capable of holding wet crockery without dropping it.

The reason we don’t have that is because, if the ai art sucks, it is of no consequence, but if the robot dishwasher unloader fucks up, it smashes all your plates. And robotics is harder than LLMs.

We already solved all the easily automatable problems in laundry and dishwashing.

It is quite competent at maths, including cooking up with real proofs of theorems people have been actively searching for for decades.

Ordinary people aren’t put off by AI art, either. My partner’s grandparents had their TV tuned to a stream of Christmas themed ai-generated videos at Christmas. They were not good, no-one had assessed them for quality for sure. No-one but my partner and I thought this was especially worth commenting on.

So it can do something in both domains…

It can do maths. It cannot do arithmetic. And actually nowadays it seems ok at arithmetic. (It farms arithmetic out to dedicated, non-language based sub agents)

Two weeks ago I was testing a draft of a puzzle to see how ai would do on it, with a page of about fifteen maths problems ranging from basic arithmetic up to easy integrals. It got all but one correct. (So that puzzle needs adjusting…)

I think people who say this remember how it couldn’t count letters three years ago and think that’s the end of the story.

Fuckin weeb