You're thinking to hard. All you need to do is build a list of all the servers used for the speed test, and then prioritize traffic to those servers. No packet inspection required. No filenames required. Just simplicity.

Do it for five or six of the most popular speed tests, and you're golden.

Seriously, I do not want Chrome's PDF renderer. It is ridiculously slow -- I can force a download of a PDF, get it pulled down, launch my own PDF reader and have it open in less than 1/4 the time it takes Chrome to download and render the PDF itself. It is also sorely lacking on features.

If this cannot be disabled, I for one will be removing Chrome from my machine, and I say that as somebody who has used it as my primary browser since it first came out. I am getting more and more fed up with the continuous feature creep and bloat in Chrome.

All of these speed tests are ludicrously easily-gamed, and are thus of next to no value in the real world. They don't tell you what speed you're getting on real-world websites, they tell you what speed you could theoretically get when your internet provider lifts caps on bandwidth, prioritizes your traffic over those of other users on the same cell tower / network for the duration of the test, etc.

And you're naive if you think some or all of the above doesn't already happen.

"if they have that little disregard for others' lives, why should we have any for theirs?"

I'd like to refine this suggestion. Do it in a full simulator capable of recreating accident impact forces. Keep the car / controlled accident scenario, but let's not pansy about: set it to a collision that's double the actual speed they were driving while caught texting. (In other words, head-on collision with another vehicle doing the same speed.) And they have to send a certain number / length of texts with no typos in a certain time to get the car to stop without crashing, but they also have to stay in their lane and not have any violations of any kind, or there's an immediate crash.

I don't think you'd have a single repeat offender. Admittedly, half of them wouldn't be alive to repeat-offend, but if they have that little disregard for others' lives, why should we have any for theirs?

...which is the equivalent of 2,832 people being murdered in the US every year.

Actual US homicide rate, courtesy of the CDC: 16,259, of which 11,078 were using firearms.

So you have a 5.74x greater chance of being murdered in the USA than in New Zealand, assuming your figure was correct. (I didn't bother to check it.) And even if you ignore the firearm deaths completely in the US (but still include them for NZ), you still have 1.83x greater chance of being murdered in the USA.

So much for the whole "guns make you safer" thing. You're less safe in the US in terms of non-gun crime, and you're much, much less safe in terms of gun crime.

Total production of the Model S is expected to reach 24,000 units worldwide, by the end of the year. Three fires in 24,000 units for the Tesla is the same as 4,688 Toyota Corolla fires, if the total production here is accurate (and that's being slightly fair to Tesla, given that we haven't reached the end of the year yet.

Now, the Corolla has been on the market for 47 years. Let's be overly fair to Tesla again, and pretend the production has always been constant. That's still 100 Corolla fires a year, for 47 years, worldwide.

I think if 100 Corollas spontaneously burst into flames each year (and realistically, more like 2-300 given that production in the 60s, 70s, and even 80s will be far lower than in recent years)... we'd probably have heard about it by now. Don't you?

Was just about to say the same thing, glad to see I'm not alone. Together we can eat our way out of this problem!

...for a bunch of "inventions" that are almost certainly blindingly obvious to anybody over the age of four, and under the age of 55. No, the patent system isn't broken at all, nuh uh. Why do you ask?

Indeed. In two decades of driving, I have had precisely zero driving-related injuries of any kind. (And in four decades of being driven or driving myself, I have been in precisely two accidents, neither involving injury to myself, and only one involving injuries to anybody -- all of which were extremely minor compared to those this single biker has received.)

It will show you every last *reported* injured biker. That's a very big and important distinction. Equally important is how many of those injuries were on public roads. Whether or not some kid on an off-road course injured himself is of little importance.

If you look at your 5-inch phone and 32-inch monitor from the same distance, you have a problem.

Because nobody brought it up. I did a find-in-page for "liability" after loading all comments, and it didn't appear once before my post.

As for your other comment, please point me to the many such driverless systems which have gone beyond PR whoring and are now on the street in the hands of the general public. Clearly, if the lawyers approved it and the tech is already available, there will be masses of them.

Oh, wait. There aren't. Because, as I said, it's a PR stunt and nothing more. It's all about making people believe your brand is on the cutting edge so you can sell them massively outdated tech at a premium, not about actually putting cutting edge tech into a car. If you got out and look at your brand new, shiny car, it's almost identical to what you were driving a decade or two ago, in terms of technology. It's just been repackaged once a year to make it glossier.

Point well made, but you misread. He said the drive *should* have taken five hours, but *actually* took three. In other words, his apparently-completely-incompetent-driver Dad was *averaging* two thirds over the legal limit, ie. 42 in a 25, 50 in a 30, 75 in a 45, 92 in a 55, and 117 in a 70. (I think that hits all the speed limits I see around these parts.)

And that's average, which given that he wasn't able to even remotely judge speed, means he'd likely have been going significantly faster / slower enroute than the average, not managing a consistent margin over the speed limit.

Which I still don't believe for one second. This story is either badly, badly exaggerated, or more likely just completely made up.

I'm amazed nobody has brought this up yet, and it's the reason you won't be seeing this in your car any time soon, if ever.

Who is liable when your self-parking car fails to self-park due to any of a million different reasons from a faulty sensor to an unaccounted-for scenario to malicious interference by a third party, and it crashes into my car -- or for that matter, ME?

With a regular car, the liability stops at its driver. (And then maybe, if the driver believes it wasn't their fault, they sue their mechanic or the manufacturer -- but mostly that doesn't happen, because it *was* the driver's fault, and court cases are expensive.)

But now the car is driving itself, and that means it is the manufacturer who's liable when it causes death, injury, or damage. If Ford puts this in a production car, they'd better be damned sure it is perfect, 100% reliable, and tamper-proof, and that if ongoing maintenance is required, that there is either a 100% reliable, tamper-proof system which alerts the owner and/or refuses to start the car if the self-driving system needs maintenance, or that the owner is comprehensively briefed on the maintenance schedule. Or more likely, both.

Otherwise, Ford is going to find itself on the receiving end of a whole lot of lawsuits it doesn't want. Which is why this "look at me" attention-whoring whizbang tech will stay in the lab, intended solely to get headlines and build reputation, but it won't be going in your car any time soon.