I've always considered "turns against" to be an unlikely scenario.

The first problem is that you've skipped over how it was created and you're focusing on how it took over once it was created.

And if you're going to do that then you can replace "AI" with "aliens" or "mutants" or "witches" or "Satan".

I envision the machine becoming an "infallible advisor" ...

And if that was what it was intended to do then it is operating within spec. So what is the difference between that system and a non-AI system designed to provide the same service?

Obligatory car analogy - an AI designs a more efficient car. A non-AI expert system designs a more efficient car. What is the difference between the AI and the non-AI?

You are still on the grid in one form or another, anywhere you'd care to be.
The electric grid.
The phone Grid.
The postal grid.

I know those "Forever" stamps could not be trusted. And the mailman? Do you think he's innocent? Don't you know that he delivers computers?
http://www.amazon.com/CPU-Processors-Memory-Computer-Add-Ons/b?ie=UTF8&node=229189

Do you think there's any safe place? Do you?
https://en.wikipedia.org/wiki/IP_over_Avian_Carriers
There is no escape.

FWIW, *both* military and factories are already well hooked up to proto-AIs.

I think you're using an overly broad definition of "proto-AI".

It's true, their decisions are tightly focused, but High Frequency Trading is only the most obvious example.

Again, I think your definition is overly broad. HFT just follows the set algorithms (written by humans) as fast as possible within the limits of the connection to trading computers.

It's my opinion that the first true general purpose AI will arise by accident.

Possibly. But that means that the AI must also be able to run on the same processors that run non-AI systems. And on the same operating systems. Which means that an accidentally evolved AI would probably be seen more as a glitch than a threat. Instead of running the apps that it is supposed to, it starts running SelfAwareness.py and eating up all the processor time and RAM. Time to reboot.

http://www.guardian.co.uk/world/2011/apr/06/georgian-woman-cuts-web-access

Now you not only have to take out the machine, its minions, but the country itself. And if more than one government back the machine? Such as NATO, or CSTO? Then what? Now you have to take out entire military alliances.

Then you're not talking about a machine apocalypse but rather business-as-usual. It's not until the machine turns against its creators/owners that there is a problem. Otherwise it is doing exactly what it was spec'ed to do.

From TFA:

The institute was sparked in part by a conversation between Price and Tallinn, during which the latter wondered, "in his pessimistic moments", if heâ(TM)s "more likely to die from an AI accident than from cancer or heart disease".

Someone doesn't know the difference between "pessimistic" and "optimistic".

In short, the answer is "no".

Not only that, you have things like EMP and nukes, not even the best AI is capable of thwarting getting bombed or nuked.

https://en.wikipedia.org/wiki/Colossus:_The_Forbin_Project
That depends upon what the AI is hooked up to. And that is EXACTLY the issue with this kind of speculation. Unless the AI is hooked into military command and control infrastructure OR controls a manufacturing plant then it will be more of a novelty than a threat.

If it is even recognized as an AI and not a glitch which gets it wiped and re-installed.

If it works out as expected everybody in Iowa benefits.

While broadly true, it is also true that the benefits will be very unevenly spread.

I'm more interested in the total square footage of these data centers and the average/median pay of the employees (also total number of employees). It is probably pays better than farming the land. But data centers usually have lots of servers and very few highly paid/skilled LOCAL workers.

I think they already did. From TFA:

Sometimes the hype tends to pan out and concepts such as âoee-commerceâ become a normal way to shop.

60% of the time, it works every time.

Either way, the term âoesoftware definedâ is with us to stay, and there is real meaning and value behind it if you look past the hype.

Except that the term "software defined" is not itself defined except by whatever marketing department wants to make it fit their product.
And the term will eventually be replaced with another marketing term.
Just as SaaS replaced ASP.
Just as ASP replaced thin-client.

What all these âoesoftware-definedâ concepts really boil down to is: Virtualization of the underlying component and accessibility through some documented API to provision, operate and manage the low-level component.

Which means that you'll only have the access and granularity that the API gives you.

And somewhere, someone will have to deal with the real servers and switches and such. And to him, you'll just be another account in a bunch of accounts. Sure, they'll deduct the cost of your downtime from your next bill. Well, the cost of what you pay them per business hour per business day.

Don't like it? Just try to get your data from them so that you can move it to a different provider.

And each provider will be under the same pressures to reduce costs as much as possible in order to maximize their profits.

People are corrupt, greedy, and stupid... this naturally leads to an erosion of individual rights.

Not just that. Freedom is scary. There are always lots of "reasonable" arguments to give up some freedom (even just for a little while) in order to defeat or defend against the "bad guys".

You don't want your freedom getting in the way of fighting the bad guys, do you?

The bad guys will abuse your freedom so that they can attack us good guys.

As always, Fascism begins when the efficiency of the Government becomes more important than the Rights of the People.

... damaged American interests.

That is not the same as declaring war on the USofA.

By telling the Chinese and the world that the US spies on them and leaking important details, he has empowered our enemies.

How? Look up ECHELON. The story here is how much the USofA spies on its own citizens.

Furthermore, he fled to China to escape US justice, and then did a weak post-hoc justification of fleeing to an enemy country ...

So China is an "enemy country"?

Where do you think your mobile phone is manufactured? If they're an "enemy" then we certainly do a lot to help their economy and employment.

I stand by what I said. He's knowingly committed suicide by doing this.

Taking a stand is not the same as committing suicide.

Our Founding Fathers signed the Declaration of Independence knowing that their signatures would be used to convict them if they lost the war. But it was not a suicide pact. It was them standing up for their beliefs.

Anything else is tyranny.

Or, to phrase it another way: if you have the hacking skills to retaliate then you have the skills to be invulnerable to the attack in the first place.

The enemy cracker has a limited number of targets:
1. your router.
2. your firewall.
3. whatever service you provide through your firewall (you do have a DMZ, right?).
4. flooding your bandwidth with traffic from thousands of zombies.

Anyone have any other types of attack that I forgot? And if you cannot secure those (except for #4) then you probably won't be able to "hack back".

I don't know but I've been told
4th platoon got Rick Rolled
Am I right or wrong? (you're right!)
Are we going strong? (we're strong!)

Thailand's Rangers Task Force 45, in response to Army policy, has put its troops to the task of promoting and protecting the monarchy in cyber space, ...

Retweeting Private Ryan.

And yet that made sense to someone in their military.

If you read what I wrote carefully, you'll see that I was talking about the inability of science to do something, not religion's ability to do things.

That part is okay. It's accurate.

Which is why the belief that the universe started with a big bang, for example, is faith-based.

Bullshit. That's wrong.

Science is falsifiable.
Faith is NOT falsifiable.

The CURRENT model says that the "Big Bang" was the origin of this universe. But that model is based upon specific, identified observations and experiments and PREDICTIONS.

Now that model may be incorrect. But whatever new model replaces it will still have to account for all the specific, identified observations and experiments and PREDICTIONS that formed the basis for the "Big Bang" model.

First off, demand that every software vendor provide a list of files that their product installs, where those files are installed by default and different checksums/hashes/etc for them.

It should be possible to boot a machine with a live CD (or PXE) and inventory every single file on that machine and identify the origin of each of them.

At least you'd know whether a machine was cracked or not.

Right now, with existing anti-virus, all you can say is that a machine does not have anything that matches the signatures that you have right now.

No movement to outsource the management of the machines to outside cloud services?

Of course there will be.

And there will be the opposite where things that were moved "to the cloud" are being brought "in house".

It's the beautiful cycle of IT.
Outsource to save money.
Insource to provide reliability/accountability.
Repeat.

That may or may not happen where you are, but there's a lot of it going on, and it invalidates much of your list.

It depends upon which part of the cycle the company is on.

Remember that CIO's do not get credit for "maintaining the status quo". They have to identify and "fix" a "problem".

Accounting servers are expensive and techs to maintain them cost too much. Move it all to the vendor's "cloud".

Can't write paychecks because someone is DDOS'ing that vendor or the ISP flooded or a backhoe cut the fiber? Better bring it in house.