Tesla footage of braking before crash ahead [video]

In this particular road and crossing there is a sign with flashing lights indicating "PREPARE TO STOP WHEN FLASHING". It was flashing while the crossing light was still green to indicate that it would turn yellow pretty soon, but the driver still didn't slow down.

it's absolutely trivial compared to self-driving cars which have to perceive their environment, and I don't need to give details. I'm not an automotive engineer.

it's trivial. To show you this is trivial without pretending to be an automotive engineer I'll use the example of a screaming auctioneer who sits on top of your steering wheel. A conversation with him might go: "HEY I NOTICE WE'RE STUCK IN A JAM IS IT OKAY OF I SIT ON YOUR STEERING WHEEL AND OBSTRUCT YOUR VIEW AND THEN I'LL START PRESSING DOWN ON THE GAS TOGETHER WITH ALL THE OTHER CARS IN THIS TRAFFIC JAM SO WE ALL START MOVING AT ONCE OKAY OKAY OKAY JUST TELL ME IF YOU'VE HAD ENOUGH READY OH LOOK WE HAVE CONTACT BETWEEN THE BUMPER TRANSMISSION LED AND RECEIVER LED AND THE CAR AHEAD OF US'S REAR BUMPER TRANSMISSION LED AND RECEIVER LED OH THIS IS AWESOME THE CAR IN FRONT OF US ISNT'T THE ONLY ONE WITH THIS EQUIPMENT THE CAR IN FRONT OF IT ALSO HAS THIS AND LOOK THE CAR IN FRONT OF OF THAT ONE ALSO HAS IT SO WE'RE ALL HOOKED UP AND AND WHEN THE FRONT CAR STARTS MOVING WE'LL START GOING AHEAD TOO ALL RIGHT JUST PRESS THE BRAKE IF YOU WANT OUT OF THIS BECAUSE WHEN THE FRONT CAR STARTS MOVING I'LL PRESS THE GAS AND MOVE TOO! HEY IT'S MOVING!!! I'M PRESSING THE GAS!!! PRESS THE BRAKE IF YOU WANT THIS TO STOP!!!!! WE'RE MOVING!!! WE'RE MOVING!!!!"

it's an absurd example but shows that it's trivial. The only thing necessary to get a stalled phantom traffic jam moving is for all cars to slowly start moving together. they can maintain distance between each other.

it's trivial if it requires special equipment such as bumpers that communicate with the car in front or behind them, proximity sensors/detectors, a way to transmit this information intercar and a way for the system to collectively brake or apply gas (like cruise control).

if you're an engineer I'm sorry if you can't see how absolutely trivial it is COMPARED WITH SELF-DRIVING CARS. It could easily have been done in the 80's with zero environmental perception, zero awareness of road conditions, nothing, and not hooked up to the steering wheel in any way. You don't need the system to handle any steering, period, in order to be able to break up phantom traffic jams.

if you've never invented anything in your life then you'll have to take my word for it that this is trivial, sorry. We're 30 years behind schedule on it and have suffered 30 years of phantom traffic jams that could have disappeared whenever they made their way to high-end cars that could have been equipped with a limited caravan system to start moving together with the drivers' permission. A phantom traffic jam can disappear at every point. Even if just 5% of cars were equipped with the standard.

you don't get what a huge deal it is for a few cars to be able to start moving together, coordinated, instead of one after the other. You don't get how 500 ms - 2500 ms of human driver latency means that traffic jams form which otherwise wouldn't and that once it is formed it is impossible to brake up and snakes its way back - but could be broken at ANY point by any group of cars that briefly acted as a caravan. you'll have to just trust me that the technology for this is easily in the realm of what could be done in the 80s. self-driving cars are lightyears ahead of that.

New Teslas have Nvidia PX2 which is very fast. Can easily run deep neural networks realtime

https://blogs.nvidia.com/blog/2016/10/20/tesla-motors-self-d...

Not sure what HW their previous cars had

I don't think Tesla use lidars, and I don't think the small Quanergy lidars are shipping yet. PX2 isn't widely available but you can pick up a TX1 which is pretty powerful, if you need something small/low power but still fast enough for your own robot. Runs Ubuntu and CUDA so you can run the same deep learning frameworks (and possibly models, if you optimise them) as on your desktop

>You really haven't.

Well, I guess I haven't. Perhaps that's because there are really a myriad of contexts for the word "value" here. For instance, you've chosen to focus mostly on "economic value".

And, sure, I have an argument with that too. For instance, you say this:

>It's saying that we shouldn't waste a human being's potential on some menial task like manually picking corn in a field or driving a car.

Sounds great, but nothing in recent history bears out a desire for this as a driving force or even a significant consideration. In fact, the reality is so far off that this argument is intellectually dishonest. That is, we're not handing over millions of people to a life filled with purpose, now that they are released from the "menial tasks" that were once destroying their potential. Instead, we're really just leaving them rudderless (and penniless) in a society where our value as humans is significantly defined by our work/economic-output.

Further, the economic benefit of this automation accrues to a relative few at the expense of those "newly freed" individuals. It is dramatically redistributing power and wealth and threatening self-determination as well as the democratic model. In short, it tacitly promotes a very real form of oppression.

So, piling onto that with a further denigration of the value of those being disenfranchised is unhelpful to maintaining a healthy society--to say the least.

>The logical conclusion to that argument is saying that we should unwind civilization and technological advancements until we're all hunter-gatherers again.

You have to know that's classic reductio ad absurdum. I've never advocated that we don't leverage tech. But, if you really want to carry an argument to a logical conclusion, consider the very real eventuality that all jobs will be displaced by technology, including "knowledge-workers". That's the difference between the current reality and the relatively low-tech technological displacements of the past to which you keep alluding: here, even skilled labor is at-risk.

This is happening on a continuum. So, I'm suggesting that it is becoming increasingly more important to define our relationship to technology in healthy ways.

Instead, what I'm noticing overall is that there is a certain hostility towards humans (devaluation) when championing tech these days. Whether it's in the labor force, with autonomous vehicles, or otherwise. Until more recently, we've known about and relied upon various tech to keep us safer and perform other critical functions. But, they played those roles as enhancements to humans. Increasingly, however, we speak overtly about replacing humans and we juxtapose the value (any value) of a human with that of technology. It's a dangerous framing.

But, it seems we just disagree here. You think there's nothing wrong with that. I do.