Flattering; but I suspect that he would have actually done the math, at least to the level of plausible approximation, rather than handwaving that part in favor of some dreadful puns.

That said, if anyone with a knowledge of railguns wants to calculate(or test, be sure to record) exactly how lousy a salmon would be as ammunition, I'd read the hell out of it.

Alpha Centuri had the ability to raise/lower terrain with terra former units... It also had a weather model good enough that this effected rainfall and thus nutrient production. It wasn't usually an efficient use of resources; but building 'moisture walls' and then watching your hapless opponent's population starve sure was sweet...

I cannot independently confirm the truth of this; but I was told, in all apparent seriousness, by someone I know well and who I know to have a long association with the hydroelectric generation business, that the term for what happens to a fish that fails to avoid the turbine intakes is "turbine induced stress". As one might imagine, this 'stress' tends toward the lethal end of the spectrum.

The engineering considerations surrounding such a device seem formidable indeed. Most of the data available are for humans(who, shockingly enough, have most of the medical budget dedicated to measuring delicate electrical signals through their muscle tissue); but if we assume that salmon tissue is approximately similar to human muscle, at least for the purposes of the currents and voltages a railgun implies, we can conclude that (A) the math is obnoxious. (B) fish are shitty conductors (C) fish have other obnoxious properties like 'capacitance' and non-homogenous conductivity.

Given the substantial resistance of our pisciform projectile, and the railgun's need for heroically high peak currents, supply voltage will have to be quite high, introducing additional insulation challenges, risks of air-gap breakdown between the rails, damaging arcs in other areas of the apparatus, and so on. Further issues may arise because of the projectile's non-uniform conductivity and substantial fluid content: with current flow, and resistive heating, highest along the most conductive regions, the projectile may exhibit substantial internal deformation, or even catastrophic loss of structural integrity, during acceleration or at a very early stage of flight. While it may have valuable specialty applications, this so-called 'frangible fish' effect markedly reduces effective range and almost entirely precludes survival of the projectile.

It is conceivable that advances in Aquatic-Preservation Discarding Sabot technology will allow a suitably packaged salmon to successfully traverse the accelerator rails while retaining the buoyancy necessary for continued survival by discarding the conductive jacket before entry into the target body of water. However, such developments are presently theoretical and cannot form the basis of a viable ecological dominance capability in the near term.

Shareholder Value, obviously. Inefficient state-owned propagandists can only maintain their activities through continued forcible appropriation of wealth. Our innovative private sector propagandists can spin the truth and turn a profit at the same time!

It's not just money(though that doesn't hurt).

Journalists practically worship 'access'. This behavior is adaptive, since it's hard to get stories written without information; but it comes with the nontrivial downside that the people the stories are about are in the best position to provide information. The competent ones have learned to take advantage of this by cultivating a relationship with the press: any really juicy story has a comparatively safe penumbra of tidbits, unattributed statements, unofficially sanctioned leaks, and so on. If a journalist is a nice, cooperative, team player, (like the quisling in TFA), they'll be well placed for a steady supply of such things.

By contrast, the uncooperative journalist might, on occasion, get a really nice scoop on where the bodies are buried(sometimes literally); but whenever that isn't available he'll be regurgitating press releases and stale news.

There's also the fact that the campus is likely mostly private land, while downtown areas tend to have public ways near them.

Depending on the local PD, your right to peaceable assembly may or may not be treated as adorably fictitious and/or a good chance to break out the cool 1033 program toys and play soldier; but you don't even have a theoretical one if you can just be rounded up for trespassing before things even start.

Trying to protest on MS's campus would just make it a question for PR of whether the visibility is lower for ignoring you and keeping the cameras away, or having you hauled off for trespassing before you make too much noise.

Are there any laws we shouldn't gut in order to be 'competitive'?

Given that it's both free, and free as in Apache 2.0, it's arguably a question that will sort itself out(though this doesn't make discussion and conjecture on the matter illegitimate or anything). Aside from the trademarks, there's nothing stopping anyone from striking out and having their own docker-in-all-but-name. So long as Docker, Inc. appears to be handling the benevolent dictatorship thing competently there isn't much incentive, though, unless you just adore maintaining a fork for the sake of it. If they upset people enough, without actually losing to another technology, we get an x11/Xorg style move; and if something displaces them, they just sink into the murky depths never to be heard from again...

It doesn't solve any of the annoying questions of making people with disparate objectives and potential personality flaws play nicely with one another; but a liberal software license does lower the stakes: If the controlling entity exerts very substantial power, it really matters how that entity is constituted. If they don't, it just doesn't matter as much because their ability to mal-administer is lower and the potential that everyone will just wander away rather than having to go down with the ship is there.

It looks like I was thinking of something quite different from the 'Authenticator' app and got confused: Recent Android versions include the 'keymaster' HAL component, which uses a software-based set of cryptographic capabilities by default, or can interact with a device-specific hardware module(mostly on Nexus devices, not sure about any third party implementations, usually based on 'trustzone' rather than TPM, just because ARM SoCs do that already).

Looks like it addresses a different use case; but that's what I was thinking about when wondering whether the behavior varied between hardware and software backed platforms. Sorry about that, I get fuzzy sometimes.

If the weapon is sufficiently fiddly and delicate, and the attacker has limited time to subvert it, a variety of means might work (many of them already explored with nukes and/or SALT arms reduction verification stuff in the late cold war); but for simpler, more durable, gear, and hardware subject to prolonged attack, Not Happening.

In particular, nukes are (relatively) easy to secure because they include a fair amount of conventional explosive, improper detonation of which will produce a mess but a fairly worthless yield, which offers a nice failsafe option. With devices that aren't as intrinsically touchy, you don't have the same leverage.

Oh, believe me, I'd love to see some absolutely ruthless discovery through every last scrap of material that passed through Yelp's shakedown/ad sales division over the years. It'd be almost as good, and a lot more legal, than just locking their HQ and setting it on fire.

However, I just Don't Even Want To Think about how awful the internet would be without Section 230. Even by the somewhat unimpressive standards of the takedown-laden world of DMCA safe harbor, life without 230 would be a killing field.

Do 'States' Rights' have any applications that aren't kind of embarrassing?

Does the behavior of the 'authenticator' app differ depending on whether the device has a hardware-backed keystore or not?

Uptake on such hardware is rather patchy(more likely on newer gear; but hardly assured), so I assume that there is a software-based fallback that just obfuscates the keystore as nicely as possible; but if the application were talking to an actual hardware keystore device, Titanium backup(or equivalent) would have absolutely no effect); but updating the application(or even switching to an entirely new one, in some cases) should be doable without losing anything.

Android key handling is an area I really haven't poked at all, so I know little about it; but I'd be interested to know how it is supposed to work.

While the idea of having a well vetted testing system in place that would allow customers to choose software that had been so vetted seems like a good idea, I have to wonder if it's doomed to the fate of SSL, at least outside of a few niche applications that mostly demand high levels of verification anyway.

With SSL, we all wanted the security; but everyone wanted it to be cheap, and wanted to avoid a monopoly over certificate authorityship. So, what did we get? A mass of CAs, many painfully shoddy, who will issue you a fancy looking cert for peanuts. Then we got "EV" certs, which are supposed to actually do what the original certs were supposed to do, only more expensive.