What do you mean "even less safe" ? The roads in most of the developed world are safer than they've ever been, and improving rapidly. For example, here in Norway when I got my drivers licencee we had around 400 casualties a year, now 20 years later we've got around 170.

And that is despite the fact that driven-kilometers has almost doubled in those 20 years. Thus fatalities-pro-driven-km has fallen by something like 80% in 20 years.

I expect autonomous cars will continue the trend, and in another 20 years we'll have double-digit-fatalities, despite another increase in kilometers-driven.

True. So perhaps that's more easily doable than I imagine. A solid-fuel booster is similar to a explosive shell, but it's not throttlable and poorly controllable in general, is that sufficient for orbital insertion ?

People have -not- been making liquid-fuel rocket-engines that survive 60000G for 100 years, indeed they like to blow up even under 1G if built imperfectly.

Seems to me there'd be plenty application.

What about artillery-shells that are able to do last-second course-corrections ? I suppose in atmosphere that's more easily done by aerodynamic fins, though.

I'd not call it impossible either, but it does seem to me to be a fairly tricky piece of engineering.

orbital insertion engines that can withstand 60000G and survive to function are in somewhat short supply.

Turkey comes to mind. The military there has on several occasions defended the constitutions againts attack from government, and then handed the keys back over to the ones who, by the constitution, should hold them; namely the people (trough elections)

True, you have to stay secure for the length of time the message has value. This varies. If you're the military, and reporting the position of a patrol in the field, this doesn't need to stay secret for very long. (3 days later the info is pretty useless anyway)

Breaktroughs in algorithms makes this hard. You can nest encryption, which means you're safe unless *all* of the levels are cracked, but it's a hassle.

That's both right and wrong. It's right in the short term: If there's no civilization, what you need is knowledge of how to survive without one - designing CPUs isn't a useful skill in that setting.

But bootstrapping civilization should be -much- faster than discovering everything for the first time, even if 99% of humanity dies out. I thinkt it's perfectly plausible for CPU-design to become useful again in a generation or two, short enough that normal books should survive if protected from the elements.

Then get sure. This argument was old a decade ago, and is getting ridicolous by now. (you hear it against solar too, not just about wind)

Windmills produce (over their average lifetime) 15 to 20 times the energy needed for producing them in the first place.

Photovoltaics produce 5-10 times the energy needed for producing them.

A large fraction of new PCs come with 8. And it's a problem. For example, my wife bought a 8-PC a couple of months ago, and hates it. Now she wants to buy a new computer for her mother, except she can't imagine going with a win8-thingie since she conciders those horrible. Might end up going with Mac, or scrounging for one of the new PCs still being sold with 7.

You know things are bad when nontechnical users are willing to pay a premium to get your 4 year old product, rather than the current one.

Or post things you want googleable under your full name and things that are more controversial under a pseudonym. For typical employers it's not even as it needs to be hugely well-hidden, they seldom do more than look at the top 1 or 2 pages when googling your name.

Is it possible to find out who (say) Indian Homemaker or Conjecture Girl is by researching on the web ? Certainly. Is it likely that the antics of these pseudonyms will be linked to the identities behind them when they apply for a job ? Not really.

No, they'd be _less_ efficient at higher switching-rates. That's because in an ideal world the waveform is square, i.e. goes instantly from on to off and vice-versa, but in the -real- world switching isn't instant, and there's a (short) "half-switched" state where some energy is wasted as heat in the switching-circuit. This wasted energy would go up with higher frequencies, simply because there'd be more switching.

Not that it matters, I suspect we're talking sub-percent of the energy-budget of a screen anyway, atleast as long as you don't want to go up into Mhz.

Yeah. And he forgets about the fact that the photoreceptors in the eye -also- don't react instantly, as demonstrated by the fact that a spinning white/black disc will appear uniformly gray. The highest detectable frequencies for flashing is in the 10s of Hz, certainly not in the Khz.

Yes, when you buy stock, the money goes to the -previous- owner of that part of the company.

But long-term stock-ownership is still influenced mostly by the performance of the actual company, whereas that's entirely irrelevant for HFT. Apple is valuable today, compared to 20 years ago because they as a company grew in every way over those 20 years.

From todays discussion it seems likely that the NSA stores not only all of the above, but the actual call-content too. Just saying.

Not that it detracts from any of what you say.

You need the permission of the author to make copies, but there most certainly are implicit permissions.

Write a blog-post, put it on a website with no mention of any licence, and you have just given me implicit permission to download the post (creating intermediare copies in RAM and possibly in disk-cache).