If every miner stopped today, they network would adjust the difficulty down, down, down, to the point where a single miner on a desktop PC could quickly process the low volume of transactions for the whole network and start farming up BTC.

If every miner stopped today, Bitcoin would die. Difficulty isn't adjusted based on time, but on blocks mined. Specifically, after every 2016 blocks, the difficulty is adjusted up or down by no more than a factor of 4. With nobody mining, the difficulty is never adjusted, and the required computing power to re-start the network is the same as it was when the shutdown happened.

Bitcoin can survive gradual changes in available computing power, but not rapid shifts.

It's not even remotely practical. An airplane's wings and fuselage are made out of a combination of high strength-to-weight materials and clever geometry. Batteries have pretty much zero structural strength, and seriously restrict what sort of geometry you can use.

A human can't maintain the level of vigilance needed to take over in an emergency for more than a few minutes. It's better to lift the "human driver" requirement explicitly than it is to pretend that the passenger sitting in the front-left seat is a safety measure.

None of the N-1 explosions killed anybody. Maybe you're thinking of the Nedelin disaster, where the pad explosion of an R-16 killed about a hundred people. The Soviet Union didn't have just one space program, it had three, and the N-1 and R-16 were from different programs.

And although it would turn up a lot of false positives, it would also catch a lot of cancers very, very early, making survival rates higher and costs lower.

It's not something people like to talk about, but early detection doesn't correspond to improved survival (see, for example, Screening for prostate cancer: systematic review and meta-analysis of randomised controlled trials ).

Forty years ago, prostate cancer was considered a fast-moving, highly lethal cancer. PSA screening was introduced, with a heavy push for annual screening of men over the age of 50. Large number of men tested positive, and a metric shitload of tumors were found and treated. And the death rate didn't go down.

For the past several decades, there's been a heavy push for frequent breast examinations and mammograms for women, the idea being that if breast cancer is detected in the early stages, it's easy to treat and many lives will be saved. The detection rate has certainly gone up, as has the number of tumors treated. The death rate? Hardly budged.

It's something of an article of faith among anti-cancer activists that screening and early treatment save lives. In practice, the vast majority of improvement in cancer survival has come from improved treatments, not improved detection. Most people with early-stage cancer either have something so slow-growing that it can be safely treated at a later date (or not at all), or something so fast-growing that they'll die despite treatment. The percent of cancers where early treatment will improve the outcome is believed to be in the low single digits.

Power 7 and later, yes, Power 6 and earlier, no, and I don't know about System/Z.

Skip the fix. It doesn't help you.

There are three threat situations involved here: one process attacking another process, one process attacking the kernel, and sandboxed code (think: Javascript) attacking parts of the process outside the sandbox.

Intel's fix is a partial fix for the first two situations (cache contamination is not the only way for information to leak, it's just the easiest and most reliable one to exploit), and is important for cloud and shared-hosting providers. It does absolutely nothing to mitigate the third situation, where sandboxed code is trying to read memory outside the sandbox, yet that situation is the one that desktop users are most vulnerable to.

All Intel CPUs that support speculative execution are vulnerable. That means the Pentium Pro and newer, all Celeron and Xeon CPUs, all Core CPUs, and all Atom CPUs except the early "Bonnell" architecture. If you've got an original-flavor Pentium or earlier, you're fine. If you've got a first-generation Atom CPU, you're fine. If you're one of those suckers who bet on Itanium, you're fine. Anything else, you need an update.

Does anyone know how this is going to affect the embedded world?

ARM is largely immune to Meltdown and Spectre. Both attacks require out-of-order execution; most ARM CPUs are strictly in-order devices.

I don't think you quite understand how steel works. Steel gets dramatically weaker as it gets hotter. At a temperature of merely "too hot to hold", it's already lost a measurable amount of strength. By the time you hit the ignition point of wood, your typical structural steel will have lost about half its strength (and will stretch like taffy, making materials testing an exciting proposition).

This would be a good thing to sell policies for, though... the odds of the company not also being obliterated in either the impact itself or subsequent extinction event/societal collapse and having to make payouts would be very small!

Actually, it would be a pretty poor thing to sell policies for. Most asteroids aren't dinosaur-killers. They're 20-meter city-shattering rocks, and if one of them actually hits a city, the resulting payouts would bankrupt most insurance companies.

Forget Ryzen. I'd like to see one of the latest CPUs benchmarked against a Core i7-3960X. 6C/12T, 3.3GHz base clock, 15MB of cache, fully buzzword-compliant. Oh, and it's almost six years old.

Honestly, it's hard to get excited about "bringing the heat" when we're talking about single-digit percentage gains. There hasn't been a breakthrough in either clock speed or IPC in years, and even core counts have remained pretty much the same.

What happens with phantom limbs is twofold:

First, the nervous system uses both positive ("there's something happening") and negative ("there's nothing happening") signals. If you amputate a limb, the brain stops receiving both types of signals, and the absence of negative signals is interpreted as sensations from the limb.

Second, the boundaries between the parts of the brain controlling different parts of the body isn't sharp. If you cut off somebody's hand, signals from other areas such as the "arm" part of the brain will spill over, and there won't be stronger "hand" signals to override them. Since the signals don't come with tags indicating their source, the "hand" part of the brain sees them as coming from the hand.

Congratulations -- you found the edge cases, the few works that continue to bring in substantial profits for a long time.

For literature as a whole, 99% of profits are made within the first decade of initial release. For music, within a year. Magazines make their profit within a month, and newspaper articles, within a day. Movies probably fall into the "one year" bucket, but Hollywood accounting makes it impossible to tell.

The single greatest threat to most creators is copyright terms. Most people aren't the next Shakespeare, or even the next Douglas Adams. Their best bet for keeping their works in circulation isn't a company raking in the millions, but communities of dedicated fans, and copyright terms -- even if they were a simple "author's lifespan" duration -- prevent that.

The newest StreetView data around here is ten months old, the oldest ten years. Most small towns have coverage of exactly one road; some don't even have that. Even in cities, there are gaps in coverage, and not just minor dead-ends: I'm aware of a three-block gap in a major arterial.