I'm not a subject matter expert; but my surprise isn't "life"(there's some sort of extremophilic bacterium cracking molecules that would make a biologist cry and only a chemist would identify as a possible energy source basically anywhere we've been able to look); but that it's big, energetic life.

These probably aren't the world's peppiest fish; but even so, a fish is a big, demanding, multicellular, operation. Some sort of spore-former bacterium that wakes up and divides a couple of times every decade or two is one thing; but fish populations mean a fair amount of active cellular metabolism swimming around in what you would expect to be a very low-energy zone.

I'm pretty sure that any fish in Lake Vostok were eaten by shoggoths millennia before the age of man. That's the other reason you take greater precautions.

Given the value of having other people using the same software that you are(they encourage commercial support and/or contribute to FOSS support, they sometimes save you a ghastly bug-slog by running into it before you do, the work that they start but don't finish, or that needs maintenance, may be your next job, etc.) isn't 'being underrated' itself a defect, however unfair it feels?

Ubiquity isn't always a good thing, especially if it makes it harder for everyone to distinguish between barely adequate crap and excellent stuff; but (with the specific exception of somebody who has mastered a specific set of skills and tools and would be very pleased for it to become an esoteric specialty just in time to land a few lucrative consulting gigs before retirement), are there really situations where you say to yourself "Yeah, Language X is great and all; but it would be better if there were fewer people using it, less incentive for commercial support or non-bitrotting FOSS support, less useful advice floating around, and fewer openings for people with a knowledge of it."?

It is obviously the case that a pure monoculture is not a recipe for success(barring a yet-to-be-invented language that can in fact be all things to all people, well); but a language that is good, possibly even modestly superior; but lacks some specialty feature of elegance and power, are you ever better off on the underrated one?

In this case, it's actually rather impressive how badly the twits appear to have forgotten.

"Hey, let's select a group of our most influential users and then annoy them with an unexpected and minimally useful nag screen when they try to use our service!" is a plan that sounds like a joke, not a strategy; but apparently twitter is now doing exactly that. Are they really gambling that all those users are just morons who are too stupid to realize that twitter has a given set of features; but would totally love to embrace them over a competitor they already use if only they are nagged enough? That seems...a trifle optimistic.

"Twitter vs. Instagram" is a frankly solid entrant for 'year's most meaningless first world battle' and we haven't even made it out of January. Nice work.

Some applications can get away with the 'trickle charge the capacitor, wake up and work quickly once the threshold voltage is hit' approach(works nicely for solar data logging, as long as you don't need moment-by-moment results); but a nanoamp is likely to fall below the self discharge rate of any capacitor of reasonable capacity; and would sleep for a long time even with an idealized 100% efficient capacitor.

The durability is impressive. It's not like cleanroom fabrication and high-purity metallurgy were exactly top of the line in 1840, so I would have naively guessed that some mixture of corrosion and non-current-generating side reactions among impurities or airborne contaminants would have trashed it in less than a century, possibly a lot less, depending on the exact arrangement of the battery, even if the energy density is totally plausible in physics-experiment-land.

Snakebites are a bit of an edge case: the production of antivenoms essentially involves inducing an immune response (in a convenient, usually large, animal) and then extracting and purifying the neutralizing protein produced. So, it is very much the case that you can prime an immune system to recognize and respond to venom.

The trouble is that snakes tend to (in the case of actually dangerous snakebites, a dry strike is just a couple of puncture wounds) introduce a substantial amount of venom into the wound, and the venoms frequently kill (or cause nasty localized tissue destruction, there are lots and lots of neat variations) substantially faster than the human immune system can synthesize the necessary counteragent, even if the person has prior exposure.

An antivenom has the advantage of being a relatively massive amount of the correct counteragent, ready to be injected into the bloodstream faster than you could synthesize it yourself.

For the less dangerous venoms, and the lower-volume strikes, acquired immunity is more useful.

I've sometimes wondered whether the techniques used to produce vaccines against exogenous drugs could be modified to produce vaccines that suppress endogenous ones. If enforced nicotine withdrawl is unpleasant, I can only imagine that, say, losing the effect of endorphins might really ruin your day...

Aside from the crass pragmatists' "Well, I bet I can develop a drug that compensates for weakness of character and lack of willpower faster than most of the population can develop strength of character and lots of willpower..." Why does this bother you?

Is there evidence that people actually develop more willpower(rather than just smoking more) when these 'replacements' are available? If there isn't, surely reduction in smoking related mortality is a win regardless of willpower, and even if there is; exactly how many people of weak character are on the acceptable losses list?

On the more theoretical side, would you condemn a drug that was actually a general-purpose willpower simulant? That actually gave the person taking it all the changes associated with 'strong will' while it is in their system? Or would you consider that to be a great breakthrough, a drug that produces a highly valuable personality trait?

It is my (layman's) understanding that nicotine is not entirely harmless; but can also have some positive effects, and overall is considered a fairly low risk compound at suitable doses(it'll kill you good and proper in quantity).

Given that, why so much work trying relatively esoteric techniques for nicotine vaccines, or low-success behavioral interventions for smoking cessation, when the only real problem that is actually killing smokers right and left is the fact that they get their nicotine by huffing a grab bag of unpleasant incomplete combustion products?

Is it that there is something particularly compelling about cigarettes, such that even people with access to nicotine by other means still seek them out? Is it just an echo of drug warrior concern that somebody, somewhere, might be employing a psychoactive without suitable risk of death or imprisonment?

I don't get it.

Ah, that was ambiguous. I meant that there are zillions of 32-bit ARM devices in the field. Only a hilariously tiny percentage of those are RT devices; but given the absurdly gigantic number of ARM architecture CPUs shipped, and the fact that 64-bit ARM is still very new and a relatively modestly player even in higher end stuff(much less the 'just a bit more than a microcontroller' market, where it probably never will be, I'm assuming that 32-bit ARM is going to be sticking around for a good while yet. RT, Not So Much.

Unlikely. The effective death of Windows RT only affects some 32-bit ARM devices, of which there are still about a zillion in the field, and more rolling off the production line as we speak(and likely to continue to be for some time to come, unless ARM Ltd. decides to piss off every customer who cares about cheap CPUs and has no need to even touch the boundaries of a 4GB memory space.

The non-RT 'Surface Pro' devices were 64 bit x86s from the start(though there were a few devices that shipped with 64 bit CPUs and 32 bit OSes because Intel didn't have some feature working quite right in 64 bits at the time); but are unaffected, so irrelevant in any case.

This will also have no effect on systems that either have 32-bit Atoms, or 32-bit UEFI(will 64 bit Windows boot from that? it certainly caused a schism among mac models at one point), which are all x86.

You are certainly getting safer as time goes on in ignoring 32 bit OSes, especially x86; but this announcement will have no effect.

That isn't strictly true, unless you ignore the fact that x86s are available(what they'd cost if Intel weren't attempting to buy marketshare might be less exciting) at more or less the same power envelope as the punchier ARM SoCs. They still have nothing on the low end of what ARM can do; but that hardly matters for phones and tablets.

Windows/x86 devices are pretty common in similar sizes and prices to Android or iOS on ARM(and, actually, some Android/x86 devices are virtually indistinguishable from a Windows/x86 device from the same vendor until powered up). There is also still the more-or-less-complete-NT; but somewhat different UI and application layer in WP8, which isn't being axed.

I'm not sure why anyone would mourn the worthless abortion that was Windows RT. All the cruft of full Win8(more, in fact, since the 'WIMBoot' feature never made it over there), including a full desktop because they couldn't be bothered to port Office to their own new UI; but with pointless cryptographic lockdown to the wonderful world of a mostly impoverished app store. All with the mediocrity of a Tegra3, and at relatively modest savings over a real computer! What's not to love?

If they actually wanted to have a go at making NT multi-architecture again, that'd be one thing; but taking pretty much all of Windows 8, then gimping it just because you have a hard-on for Apple's app store success? An idea that stupid deserves death.

It depends on exactly what they mean by 'can't upgrade our computer systems'.

If that just means that, because our budget is smaller, we are being forced to go from an X year refresh plan to an X+1 or X+2 year plan, well, too bad, so sad. That may be strictly true; but it's also one of the (within limits) easy things to adjust as an IT budget moves around. There's a line below which it just isn't worth tossing near-new gear, and a line above which keeping legacy systems in production starts to cost more, sometimes radically more; but modest modifications to the basic systems are easy.

If they are saying that 'We can't cut Legacy Horror X off life support and adopt Shiny System Y; because of budget uncertainties', it is very likely that they are correct: A transition between two systems is not a fun time to hit a budget freeze that leaves you with almost zero room to deal with the unexpected. I don't know how urgently they actually do need to migrate; but if that is what they are talking about they are at least talking about a plausible issue.