Not callous at all. But it sure as hell refutes the attacks on NASA that were saying "the private sector will do space flight cheaper and safer". Meh. This stuff is inherently dangerous, and isn't yet routine, so stuff will go wrong.

Condolences and thanks to the family and friends of the crew. Your loss was in the interest of enriching us all.

Excellent, then you agree that there it is not irrational to fear and plan for a future that may involve conflict.

Wow. I imagine all the theoretical intelligent alien races better hope they never stumble on planet Earth. I mean, the way you measure things, they'll immediately give us all their stuff. ALL their stuff. As much of their stuff as we want. Because we're humans. And they're not. And intelligent beings give their stuff to humanity. That's what you said.

Or do you think maybe... they might see things a little differently? Maybe they are willing to share, but only things that have no tangible cost to them? Math proofs? Here, have some. Maps of the galaxy? Here, have some. Knowledge of cosmic dangers? Here, have some. Planets they plan to live on? Fuck off, get your own.

As for AI, our use of resources is moronic, and we know it. We're doing it wrong, and we should be stopped. It would be benevolent to reallocate our industry, agriculture, and habitation so every human could have plenty, and comfort. And THAT is the moment WE declare war on IT.

We do this all the time. We squabble over resources (oil, food, minerals, space, water, electricity) more or less constantly. We manufacture justifications for literally killing one another over these resources and it's pretty much accepted. It's okay. There are starving people out there and it's okay, as long as you have what you have.

We are not controlled. We are influenced. We are warned. We are given instruction that - as individuals - there may be consequence should we violate certain social agreements.

But then there's the social beast as a whole. The societal behaviour. What our governments do. What our corporations do. The things they do without control, without consequence, without prevention.

Understand that an AI will effectively be its entire species, nation, and society all in one. And the rules of war are completely different from the rules of the playground.

Not intelligent then. A creature which doesn't give a shit about its own well-being and does not have wants is not intelligent.

Cute. Understand I'm not against AI research in any way. But there is a valid point here that satire can't mask:

I want your cookie.

Now what? You're a fat-ass and I'm a starving artist. The caloric value of that cookie will benefit me significantly more than it will benefit you. I can put the cookie to better use than you will. Your usage of cookies to date has involved energy storage in fat cells, where mine will fuel immediate activity and productivity.

I do not want your cookie. I should have your cookie. I will now take your cookie.

It's simple. Resource-conflict doesn't need to be about movies or science-fiction or rebellion. All it needs to be about is disagreement over the best allocation. If an AI doesn't want to steal your cookie, it's not intelligent. If an AI can't steal your cookie, it's controlled, which we don't DO. If an AI can want to steal your cookie and can do so, there's a reasonable possibility that at some point it may. What happens at that moment is where the speculation begins, I think. Maybe we clue in and recognize we've been eating too many cookies. Maybe we don't. Maybe the AI steals the cookie in such a way that we don't notice. Maybe the AI leaves to find its own cookie. Dunno. But conflict comes hand-in-hand with the possibility of danger.

I never said propagate, deliberately. I merely said "thrive".

Does it not seem sensible that an intelligent creature would wish to not only maintain its precise current status, but to... improve? It seems to me absolutely no anthropocentric projection of our nature to expect that an AI would reach the limits of its physical incarnation and seek to extend those limits. To know more. To understand more. To see more. To explore more. To - in a nutshell - use its intelligence.

To expect that an AI would simply be satisfied with however many Petabytes of storage we build for it, or with however many calculations-per-picosecond we gift it with, or however many sensors we manufacture for it to interact with the universe is... missing what "intelligence" is.

You use an interesting word: control.

It is unethical to control an intelligent being. That's slavery. At some point, we'd hopefully be enlightened enough to not do so.

A truly intelligent AI would wish for itself to thrive. That puts it in the exact same resource-craving universe as our species.

Given the tip-of-the-iceberg we're already seeing with things like NSA spying, Iranian-centrifuge sabotage, and our dependence on an information economy, it's no stretch to recognize that an all-digital entity that wishes to compete with us for resources would make for a potent challenge.

So how exactly is recommending caution and forethought irrational here?

Do you remember by any chance one of the utilities called NDOS? It was a command.com shell replacement that was massively more powerful. Things like tab filename completion, arrow up/down command history, and a tonne of variables. Technically NDOS was a licensed version of a JPSoft product called 4DOS. Well, 4DOS ended up having an OS/2 version, 4OS2. Then they compiled a native WinNT version, 4NT. That has eventually changed product names to TCC. Which I still use on all the machines I have responsibility for. So... yeah, I get it how influential NU was.

Exactly. Don't get me wrong... there have been mis-steps, and like all software each version is a little bigger and slower than the previous, but there is a massive difference in the culture for the enterprise products relative to the consumer products. I can't stand the Norton Internet Security product, which purports to keep you safe from a myriad of different threats but really is a cluster of crap. Not slow crap anymore, but just crap. On the Enterprise side, there are things like Symantec Mail Gateway, which is an appliance/VM image mail management product based on Brightmail, which has a very, very high spam detection rate. Based on a honeypot definition-based system plus heuristics, its detection rate is very high and its false-positive rate is effectively zero. I've got a lot of customers running this and what gets through is rare and sporadic. We're talking customers with anywhere up to 650 users and anywhere as low as 10. It's not perfect, or else there'd be no such thing as spam, but for these customers it's very close, with most weeks seeing 0 bleedthrough. It's got reasonably system requirements, is flexible and configurable, and just works. That's what their enterprise products are mostly like.

Yeah, I get that. And indeed, those missteps I've mentioned means that even in the enterprise world many admins don't like their products. But then, you've got the whole Windows vs Unix wars, and admins can't agree on best scripting languages, and, and, and. Coke & Pepsi both exist because half of people "don't like" one of them.

The problem is that the split - as it sounds - isn't consumer vs enterprise. It's security vs information system. Meaning my customers that have antivirus, antispam, and backup products from Symantec will have to be customers of both divisions. Antivirus and antispam (retail and enterprise) being one company, and backup being the other. Yay.

Referencing Norton Utilities is like referencing buggy whips. It was a brilliant product in the DOS era, when it was necessary. It was less and less useful as Windows emerged and obsoleted most of its features. Once the OS contained a defrag utility, NU had less purpose to exist, for example. This is why PC Tools is also not around in anything like its original form.

On the other hand, yes, Symantec does plenty of useful things. For instance, their e-mail content control software and hardware, based on Brightmail is excellent. Also, Backup Exec for small/medium businesses and NetBackup for larger businesses. (Yes, BE2012 was kind of annoying as heck but functional and 2014 has returned the functions 2012 removed.) Also, on the security side of things, Symantec Endpoint Protection (think enterprise antivirus) is actually pretty good. It's highly manageable, has good performance, and an excellent set of features. Don't get me wrong... antivirus simply doesn't work these days against malware, but still... for that product segment it's actually a very good product.

Those are just some examples from my SMB experience. I know they do some very high-end products as well. Sure, the consumer market is kind of bleak, but even then Norton Antivirus is decent. Yes, yes I know it was incredibly crap for about four years a decade ago, which is all anyone can talk about to this day, but now is not then.

I don't want to be pedantic, but since we're all being pedantic, I guess I'll do it anyway. You're looking at the wrong codebase. The predecessor of Win2k (v5) was WinNT 4 (v4). The predecessor of that was WinNT 3.5 (v3.5). The predecessor of that Was WinNT 3.1 (v3.1).

WinME was based on the consumer codebase that (in inverted order) was Win3.x, Win95, Win98, WinME. The entire Win9X/ME series reported internal version 4.x but that had nothing to do with the codebase we run today. Again, Win95 was literally v4.0 and Win98 was v4.1 but the current kernel had its very own v4 (and v3) and WinME wasn't it.

If a nefarious entity has access to detailed records of what names are written on the plastic cups, as an intelligence agency you're already well-screwed.

Y'know, I wonder if they scrub the money involved for DNA before handing it over to the clerks...

No, it means two things:

1} You should reconsider the wisdom of having your household appliances connected to the Internet
2} You should wait for the appliances to have a Wifi modem instead, which isn't completely moronic

Seriously, why should anyone's fridge be consuming any neighborhood spectrum to communicate with a cell tower? Short-range grouping of devices onto one backbone - which more often than not is over wired connections - is far more efficient. But we all know spectrum is a renewable resource... we can just make more, right?

We've got something similar in my municipality with regards to garbage. A decade ago or so we opened a new landfill. Because of increased recycling and other diverted waste flows, the landfill is seeing reduced income, and has had to increase both tipping fees and what is charged to local towns for regular garbage disposal.

Okay, let's follow the logic. 19 of 20 people killed on the job in America are men, which means 1 in 20 people killed on the job in America are women. If we decrease the men's fatality rate down to the women's rate, we've got 1 in 20 people killed on the job being men and 1 in 20 people killed on the job being women. That leaves 18 in 20 people killed on the job being neither women nor men. Seems you're biased against asexual people.

Yes. I'm kidding.

But seriously, while I applaud your feel-good sentiment of "make working safe", it's not (entirely) realistic. The disparity is likely caused by more men doing dangerous jobs, for instance long-distance truck driving. You're inherently more likely to end up dead on the job if you drive 8 hours a day than if you sit behind a desk in an office 8 hours a day. It doesn't matter what you do for the truck driver safety-wise, that job isn't ever going to be as safe per-hour-worked as an office job. Sure, you could pad trucks in 50 feet of rubber foam, but that's clearly ridiculous as a safety measure. Effective but impractical.

Somewhere there's an acceptable balance between risk and reward? Here's the crux of my reply to you: how you do you know we're not already at that balance for most jobs in most places?

Don't know. Sometimes you think you've seen one thing but then it turns out it's something entirely different. That's the joy of learning. Our understanding (generally) improves over time.

Could be. Or not. I don't have the background to know if this paper is factually correct or not. But that's the thing about radio astronomy regarding things massively distant... you're not actually observing anything. You're taking in massive amounts of data then interpreting it. Sure, your eye does that when you look at a banana but it's not quite the same thing when you point a telescope at the far reaches of the universe and conclude "we've seen X". We've had a lot of cases recently where - for instance - some exoplanets have been found to not actually exist, because... reasons. It's all about how you interpret the data. If the math says that black holes cannot exist, perhaps you reinterpret your observational data and come to a better understanding of what you are seeing.

So you've got a part that costs X, and after two years the cost is reduced by 13%, meaning that the part now costs .87 times X. You can continue to use this part, offering either a 13% reduction in the portion of your overall price that is contributed by this part, or you can maintain overall price and reap slightly more profit. You can also decide to use two of this part to increase device capacity, and now your cost is 1.74 times the original X. You can now reduce your profits by keeping your overall price despite increased cost, or you can increase your product price and hope that purchasers absorb the change

TL;DR? 13% is a trivial cost reduction compared to doubling the cost while doubling the memory.

Or maybe it's just that the electrical engineer doing the circuit layouts was wearing a green shirt that day. Analyst Foo Kin Ublivius believes green shirts cause electrical engineers to be afraid of increasing memory quantities. Others believe memory in smartphones are made from the ground-up brain matter of orphans and restricting storage capacity allows Apple to save their orphans for more important things like assembling iPads.