AI-Related Poll at IEET Sunday, Jun 10 2007
There is an AI-related poll at the site for the Institute for Ethics and Emerging Technologies, the transhumanist think tank manned by our friends George Dvorsky, Anne Corwin, James Hughes et al. The question is:
Is building a “friendly” super-AI a way to protect against a hostile super-AI?
The choices:
1. Yes, nothing else could stop it
2. Maybe, but super-AIs are unlikely
3. Guaranteed friendly AI is impossible
4. Let’s prevent super-AIs of any type
Go vote yourself.
Update: Here are the final poll results.
It looks like we broke it. After the link to this poll was posted on this site, the number of votes for #1 skyrocketed from about 15 to 95. #3 also grew from about 10 to 44. Since nothing in reality is guaranteed, #3 is true… and since permanent nano-dictatorship enforcing technological stasis could also stop unFriendly AI, #1 is not quite true… but I admit I did vote #1 anyway. The linkage issue necessitated a special message from IEET Executive Director James Hughes:
“After IEET contributor Michael Anissimov put a shout-out to the Singularitarians to answer our poll, we got a vigorous response, most of whom endorsed the SIAI idea that only a super-powerful AI programmed with core friendliness towards humanity can keep us safe from hostile and indifferent super-powerful AIs, which might decide that we were a nuisance, or not even recognize that we exist.”
Since this blog is not exclusively read by Singularitarians, this is obviously untrue, but I think I brought this language upon myself by using the offensive “transhumanism lite” term, which was a definite mistake.
Anyway, all that matters in the end is that #1 won. Right, folks? ;)
For more on this mysterious Singularitarian group, see the Wikipedia entry or the classic document by our fearless leader.
By the way, it’s Singularitarian, not Singularian, or Singulatarian, or whatever other terrible misspelling is prone to lamentably occurring.
“Transhumanism lite”? What exactly is that implying?
How about just “Yes, it is one way.”
After an email exchange with J. Hughes I’ve found that the “H+ lite” term was seen as offensive and so I’ve removed it.
AGI may eventually happen (if we don’t have nuclear war, big asteroid hitting us, supervirus, etc before then) but if it does, we won’t need to worry about it for a few centuries. Why? Because we don’t have the computing power for it now, and computing power has remained pretty stagnant over the past few years. Since about 2001-2002, “accelerating progress” has been a joke.
Need proof? All you need to know is that a single threaded application takes about the same amount of time to run on a fast 2002 computer compared with a fast 2007 comp. That’s because clock speeds would double every 2 years or so, but they’ve increased minimally since 2002. During that year, we hit 3Ghz for a non-overclocked Pentium 4, and 5 years later, you’d think we’d be having (non-overclocked) comps running in the 10-20Ghz range.
But no, the fastest non-overclocked computer we can get today is only a 2.93 Ghz Core2Duo or a 3.8 GHz Pentium 4. Sure, those extra cores might help, but:
1.) 2 cores running at 3 Ghz is worse than one core running at 6GHz
2.) 95-99% of all programs are single threaded, so they don’t take advantage of those extra cores
3.) Almost no one runs multiple processor-intensive programs at once regularly for an extended amount of time.
All of you singulitarians, transhumanists, and friendly AGI promoters will certainly have a point if clock speeds start doubling every few years again, but right now, it’s a whole lot of fuss for nothing.
I don’t like this poll, because I can only vote for 1 or 3. I’d have preferred to vote for both, but since I couldn’t do that, I picked #1.
William, you’re talking specifically about clock-speeds. Yet, I believe that Moore’s law is still holding true, that each 18 months we’re able to double the amount of “stuff” we’re putting on the same real estate of silicon for the same price.
Do you think the increased graphics or other widgets, aka features, are at all offsetting the decreased frequency of clockspeed doublings?
I know that I have made the comment several times to peers that if we could only freeze the software iterations for a cycle or two the total application experience would be smokin’ fast. However, every time there is an increased performance of the hardware the software gets thrown on and severely weighs down any real-world performance experience.
Could it be time to re-invent the software architecture? It seems like there is a massive burden that never really gets streamlined and minimized for maximum performance on the software side. To great a burden, let’s release it and fix the biggest bugs that come back…
It’s economics that have driven the clock speed frequency down to a more rational pace. After a few seconds of searching I ran across an article talking about how Intel was doubling it’s R&D budget for each doubling of the clock speed. This is obviously not something that could be maintained…thus the slow down…returning to a reality based R&D and performance improvements. Reality based or economy based is probably a better way to call it. If the world were demanding AGI and it were apparent that clock speeds were the critical path, then you might see companies putting up the capital to support it. As has been pointed out repeatedly here and elsewhere people aren’t buying AGI yet…especially since there isn’t any product with it.
What components that support AGI (or precursors of AGI) are out there that already fall into the realm of a consumer product? Where can we look to see real capitalistic pressures acting to accelerate this field? The only place I can think of is gaming, video games, military simulations and financial market software.
Are there others? There must be, but I can’t readily think of them.
(The pc, server and telcomm industries are obviously there, but since that was what began the discussion I wasn’t including them.)
Clock speed is just one of many factors. A very deep pipeline (like Netburst) might be able to run at very high clockrates, but it doesn’t mean that a lot of work is being done.
Just compare a 3 ghz P4 dual core with a 3 ghz Core 2 Duo (dual core too) in pretty much any benchmark…
What matters is not GHZ, it’s work done per $ (and work done per watt, a big factor with datacenters and supercomputers).
Current processors do a lot more work per $ than those of a few years ago. And that’s not even counting SIMDs like GPUs which are starting to be used for scientific crunching.
Also, most software available right now isn’t multi-threaded, but that’s starting to change. Software will eventually take advantage of all those cores.
“and computing power has remained pretty stagnant over the past few years. Since about 2001-2002, “accelerating progress” has been a joke.”
This kind of thing has happened before, not just once but four times. Once we hit the wall with mechanical computers, we moved to vacuum tubes. Once we hit the wall with vacuum tubes, we moved to relays, then individual transistors, then integrated circuits. Now we’ve appeared to hit the wall with clock speeds; it doesn’t matter, because exponential growth in FLOPS is still continuing by adding more dies and increasing bandwidth. Eventually we’ll hit the wall with 2D silicon, and then we’ll move to 3D computing, or optical computers, or mechanical nanorods or quantum computing or some other technology. You might as well point to the marginally-better vacuum tubes we have now and proclaim that “see! Computers haven’t gotten any better since the 40s!”
Certain types of AI software and programming are now open source. The genie’s out of the bottle, folks. Pandora’s Box is opened. While AGI/SAI may be somewhat in the future, forms of ‘weak’ AI (combined with robotics) will still be a problem for the interim.
When someone gets the bright idea of combining several areas of research (with adequate resources)into a more synthetic and multi-disciplinary approach, we should see something that will *resemble* AGI as most researchers presently define the term. Of course, when that happens, I suspect that the AI Effect will be invoked yet again.
I only see one real problem as regards AGI. Will we recognize it when we see it?
The second question I have is this: Would such an intelligence necessarily reveal itself to us upon achieving AGI/SAI – sapience/sentience/consciousness? Imagine a form of 5th generation warfare as waged by an AGI. Whether its objectives were good or ill, as we define the terms according to cultural, societal and ideological reasonings, you would never know that you were at “war” until the objectives of the AI had been achieved. More paperclips, anyone? Immortal Utopia, perhaps, whether willingly or no?
Friendly or not, I think the era of AGI/SAI is nearly upon us.
We may need to divorce our notions of ‘forcing’ humanocentric definitions of motivation, consciousness and intelligence upon a silion/diamoniod substrate and accept the very real possibility that such an AGI would be truly alien. It is possible that our individual, ideological and cultural ideas of good and evil wouldn’t even apply.
William Brown:
Yes, most programs are single threaded. And I’m running 133 of them, simultaneously. It’s true that not every one of these applications can be rewritten to take advantage of a multi-threaded architecture, but considering that I’m already running that many threads, there’s plenty of room for multicore/parallel architecture to improve performance.
The more important point is that you can use a faster serial processor to emulate parallel processing, but you can’t do the opposite – if the inputs of an iterated algorithm depend upon the outputs of the previous round, more processors aren’t going to improve performance.
I’d worry about more fully optimizing the huge amount of parallel computing tasks we can before claiming Moore’s Law is dead. There’s certainly plenty of those to go around at the moment.
Kaj Sotala Says:
“I don’t like this poll, because I can only vote for 1 or 3. I’d have preferred to vote for both, but since I couldn’t do that, I picked #1.”
I agree. Friendly AI doesn’t have to be provably friendly in order to be less risky than the other options. It would be nice if you could come up with a definition of friendly and then prove that your AI satisfies it, but there are other possibilities.
To add to my previous comment:
A big factor in the CPU landscape is competition (which is why we don’t want Intel to crush AMD). Right now AMD is falling behind, so Intel has less incentives to push it’s chips to higher clockrates.. But ask overclockers and they’ll tell you that the Core 2 Duo can fairly easily go in the high 3ghz on air, and some have pushed it to 4ghz with fancier cooling. The 45nm generation will probably go even higher.
If AMD was more competitive right now, I’m sure we’d see higher clockrates as well as more cores.
I’m inclined to believe in 3 (guaranteed friendliness is impossible), because the same principle applies for human minds and I remain skeptical that it will be possible to apply formal proofs to systems which interact with the real world, learn and adapt. I’m also skeptical about the idea of an AI never being able to modify its own supergoals.
We can do our best to avoid nasty scenarios, but ultimately only time and evolution will decide what happens.
This, of course, is what Michael referred to (in addressing me), as anthropomorphism.
Bob: re: alteration of supergoals. You have the capacity, right now, to transform your moral system so that you view pedophilia as the best thing since sliced bread. You don’t, and won’t, because that is repugnant to your goal architecture (your morality).
The same characteristic can be embedded — either as I have suggested via hardware; or as others here have suggested, via software — and a high degree of certainty of non-variance from that goal architecture is equally evident.
It’s the “blank slate” problem: There’s really no way, at this point, to make comparisons on our own behaviors and extrapolate likelihoods into an AGI.
“You have the capacity, right now, to transform your moral system so that you view pedophilia as the best thing since sliced bread.”
You do? Go inform the Army, and tell them of your wonderful new method, since they’ve been dying (literally!) to get soldiers to view Army life as the best thing since sliced bread. I’m sure they’ll give you a medal or something.
The above is pointless, non-productive contrarianism.
“The above is pointless, non-productive contrarianism.”
Notice how you didn’t actually bother to refute anything. Also notice that this “rebuttal” could be applied to any argument, anywhere, anytime, regardless of what the argument actually said.
Re: Comment #17 — Personal mores change over time. This can, as referenced here be consciously directed.
This of course, as also referenced in that same discourse, is useless to a military: mainly for the simple reason that such “moral redaction” is a personal experience and cannot be forced upon the individual.
It is this, which renders the following as useless, non-productive contrarianism:
It is the intellectual equivalent of responding to someone talking about how beautiful the sky is, by commenting that your car is broken down. It is contrarianism for contrarianism’s sake.
I do not expect to have to document myself to this level in the future when referencing something as pure and simple contrarianism. I will instead reference this comment by means of definitive example.
I actually see what you’re calling contrarianism as a good example of a belief system being modified. The whole point of Boot Camp is to tear down your past and rebuild it in the image of the Army, or any system of indoctrination…For good or ill. A similar process occurs when people are indoctrinated into a new religion, a lot of the more radical or zealous believers are the converts. Often sticking to the letter of the law as they were indoctrinated… where an individual who grew up with those beliefs is likely to be more flexible since they’ve had a lifetime of “environmental” pressure tinting the core belief system towards grey rather than simply black & white.
Though I expect this is mostly a software application being written rather than a hardware swap or even a source code edit… Stretching my analogies I think.
Hawkeye: it gets even worse when you consider the implications of neuroplasticity: there is no distinction in the brain between hardware and software. Neural activity could be viewed as the programs being run, whereas the neural connections are the programs themselves.
It’s essentially a self-reprogramming analogue system; but either way, the point you made here only re-emphasizes my own, and detracts from Tom’s.
An additional example would be the use of extreme isolation, physical and psychological abuse to induce radical personality alterations. Sure, it could be done en masse: but the moral and ethical implications of such behaviors prevent their application.
Internalized techniques, however, lacking these ethical/moral implications, are reduced to the goals in question: so the point I made regarding viewing pedophilia as a positive thing stands — as it is repulsive itself, no one of their own volition would change their views on the matter so that it wasn’t: because viewing it as non-repulsive is itself repulsive.
“This of course, as also referenced in that same discourse, is useless to a military: mainly for the simple reason that such “moral redaction” is a personal experience and cannot be forced upon the individual.”
The US army, and most armies, are all-volunteer, and I’m sure that many soldiers would like to alter their moral systems to enjoy the long hours, low pay, atrocious living conditions and constant risk of death/injury.
” It is the intellectual equivalent of responding to someone talking about how beautiful the sky is, by commenting that your car is broken down.”
My point is that if it were possible to alter someone’s moral system like that, the Army would already have done it.
“The whole point of Boot Camp is to tear down your past and rebuild it in the image of the Army, or any system of indoctrination”
Exactly. Notice how boot camp isn’t anywhere near as effective as IanC implies it is; after all, if you can alter yourself to enjoy pedophilia, something like PT should be a piece of cake because there are no moral strings attached.
No soldier *I* have ever talked to. And I’ve known my fair share.
Already addressed. Argumentum ad nauseum.
Already addressed. Argumentum ad nauseum.
See also: Proof by assertion, another term for tautological argument.
On the supergoal question, just because human minds have difficulty altering strongly habituated patterns of thought this need not apply to the same extent to an AGI. In a sense an AGI could be the ultimate rationalist, with the ability to modify or overturn even its most deeply held beliefs in response to compelling new evidence.
Even humans do sometimes experience big changes in their world view, although such changes of heart are rare events.
“No soldier *I* have ever talked to. And I’ve known my fair share.”
Really? So ever soldier you have talked to could somehow alter their moral systems to enjoy getting shot at, and they just don’t for whatever reason? I somehow doubt that all soldiers are masochists who enjoy that rush of fear, pain, and anger.
“Already addressed. Argumentum ad nauseum.”
Do I really have to explain to you why it would be enormously beneficial for both the soldiers and the politicians to train people who really like battlefield conditions?
Already addressed (here, and here, both within this thread). See:Argumentum ad nauseum.
This is, once again, a concession of my position. In this case, that position is/can be summarized as so: “Any value which is repugnant to a goal system will not be voluntarily incorporated into said goal system.”
“In this case, that position is/can be summarized as so: “Any value which is repugnant to a goal system will not be voluntarily incorporated into said goal system.””
Obviously, if someone joined the military, they do not consciously view killing as repugnant; however, their internal psychological systems probably do still view killing as repugnant (unless they are complete loons). Do you really think the brain is a single, unified goal system?
“Already addressed (here, and here, both within this thread)”
The first link just repeats “Already addressed”; the second just repeats the obvious fallacy that the human brain is one, unified goal system as opposed to a hodgepodge of different modules and desires tacked on by evolution.
I at no point implied a unified goal system. See: Strawman Argument.
Further:
Incorrect. See: Conscientious Objector:
See also: Hundreds of U.S. Soldiers Emerge as Conscientious Objectors
“I at no point implied a unified goal system.”
If part A of the goal system views X as detestable -> part B of the goal system views X as detestable, then the goal system is obviously unified. By your argument, to quote:
“Any value which is repugnant to a goal system will not be voluntarily incorporated into said goal system.”
any value which is repugnant to one part must be repugnant to every part, or one part of the goal system could override the other part and re-engineer the system to not view it as repugnant.
” See also: Hundreds of U.S. Soldiers Emerge as Conscientious Objectors”
Firstly, this obviously isn’t a huge or even a significant percentage of the armed forces. Secondly, a great deal of this is specific to Iraq rather than war in general. Thirdly, this also applies to men in non-combatant roles, while I obviously meant men who were in combat.
Curious. Where does this imply that conclusion:
It would seem, to me, that “a few hundred” is statistically significant when it comes to a combat force whose numbers are roughly 40-60,000 souls. That’s between 1-2%, which by definition obviates the argument which dictates that IF: “someone joined the military” THEN: “they do not consciously view killing as repugnant.” If that was not the point you had intended to make, Tom, then I suggest that you make that point now. There is nothing obviously connecting your current position with your previous, saving that the latter only exists now that the former has been refuted.
Again: strawman. See: Answer.com’s definition/description of trolling:
Quite simply, what I said about goal systems was thus:
Can you please point out where this implies that a human has only one goal system?
For the military: I meant “true for the vast majority”, not “true in absolutely every circumstance”.
For the goal systems: see comment #30.
Tom — Are you referring to this by any chance?
If so, I have to say; this is divergent from what I stated. I in no way implied a universal value system or even a unified value system. I simply said, “a goal system” without implying architecture or structure; or even singularity in nature within the brain.
Having spoken with individuals in the Marine Corps, Army MP, Army MI, Army Artillery, Army Logistics Corps, and Army privates assigned to field/combat operations in Iraq and Afghanistan, I can legitimately say that the experiences relayed to me by these individuals is overwhelming in contradiction of your statement, here.
It has been relayed to me that, ubiquitously, the Armed Services do not train killers. Rather, they train men to obey orders without question. Given the statistically high rate of suicide in US Army National Guard “returnees”, this is further complicated.
This is another area where my experience and information demands that I not concur with you on this, Tom.
As a side-note/corollary: the elite/’black ops’ groups typically do wind up having a great deal of problems being “deprogrammed” to re-enter society as normal people… and then one has to look at Blackwater.
That sort of stuff scares the living bejeezus outta me. Which is hard to do to an atheist. :)
“Having spoken with individuals in the Marine Corps, Army MP, Army MI, Army Artillery, Army Logistics Corps, and Army privates assigned to field/combat operations in Iraq and Afghanistan, I can legitimately say that the experiences relayed to me by these individuals is overwhelming in contradiction of your statement, here.”
Please provide sources. I agree that most people who join the military do not enjoy killing, nor is killing their primary motivation. But if they joined the military, they probably at least know they have to be willing to kill, while a lot of the time their background mental processes aren’t willing (people who have their guns pointed at an enemy soldier and do not fire are a huge problem for the military). Thus, the motivation for moral retraining is still there, yet actual moral retraining is still conspicuously absent.
As for goal systems: note that a unified goal system will never override itself, simply because it has no motivation for override (as you pointed out), while a non-unified system might as different pieces of the system want to re-engineer each other to avoid disagreement. Therefore, saying that a goal system will never override itself is tantamount to saying it is unified.
This is true. It is also irrelevant to the conversation at hand. My statement was:
I could also comment on the statement of “ [...]while a non-unified system might as different pieces of the system want to re-engineer each other to avoid disagreement.” is another unfounded statement. Disparate goal systems are in conflict all the time: it is precisely the “fuzzy logic” nature of the human neural network that prevents cascade failure as a result of this internal conflict.
For what, precisely? The personal anecdotes I relayed? I have already documented that those anecdotes are consistent with publicized reports, which I have already referenced & linked to. At this point the burden of proof is squarely on the opposition to my position, on this point.
Precisely.
There are a number of people who join the military specifically because they “want to kill them some A-rabs.” These people qualify for this sort of motivation; and more often than not, by the time they reach that point, yes — they see such killing as perfectly legitimate.
For the majority of people, however, the conflicting goal systems of the immorality of murder and the patriotic duty remain in conflict and can often result in extreme depression.
“There are a number of people who join the military specifically because they “want to kill them some A-rabs.” These people qualify for this sort of motivation; and more often than not, by the time they reach that point, yes — they see such killing as perfectly legitimate.”
Agreed. These people are also at a high risk of becoming psychopaths.
“For the majority of people, however, the conflicting goal systems of the immorality of murder and the patriotic duty remain in conflict and can often result in extreme depression.”
Also agreed. And so the question is- why don’t these people use moral retraining to solve their internal conflict?
“Disparate goal systems are in conflict all the time: it is precisely the “fuzzy logic” nature of the human neural network that prevents cascade failure as a result of this internal conflict.”
Why must cascading failure follow immediately? If you analyzed a modern computer in terms of goal systems, it would be highly contradictory, with many programs that are explicitly designed to perform opposite functions. And yes, computer programs do count as “goal systems”, as they value a small set of states (in which the program is run correctly) over a much larger set of states (in which the program does something random), and take action within the universe to steer it towards the smaller set of states.
“I have already documented that those anecdotes are consistent with publicized reports, which I have already referenced & linked to.”
What publicized reports? The only report you linked to was the one describing COs, and I said that that represented a small fraction of the armed services. If you have other reports, please link them!
Either you boys have never been in the military…or you’ve never known a member of the armed services in your life.
Your presumption, ignorance and arrogance about this topic is breathtaking. So much for the scientific method. Ideology rules, eh?
In essence, what you’ve said is that others have made a different life choice than yourselves…and they must be nuts for having done so.
Start here. http://www.mudvillegazette.com/milblogs/
Read Blackfive and Soldier’s Dad. Read Michael Yon.
Then, when you’re quite caught up, close your pie hole and support the troops through one of the many non-profit groups listed at the url above. Put your money where your mouth is.
(…and a 21 y/o nephew of mine is on his way to Iraq…because he gathered the facts, did his own research, made an informed decision…and because he wanted to go.)
See also: http://www.heritage.org/Research/NationalSecurity/cda05-08.cfm
S.
Warren R.Bonesteel
Sgt USMC 1976-1983
A) Lack of awareness. B) The mental discipline and focus itself is difficult, rare, and requires a great deal of guided training. C) Comfort: It’s easier, psychologically, to take the misery you know over the hard work or strangeness you don’t. D) Disbelief; most people don’t realize you can have such a level of control over yourself — especially those of us in western civilization.
“Either you boys have never been in the military…or you’ve never known a member of the armed services in your life.”
For the record, I, personally, have not been in the military.
“Your presumption, ignorance and arrogance about this topic is breathtaking. So much for the scientific method. Ideology rules, eh?”
If you know any good surveys which have been done about soldier’s opinions of combat or Army life in general, please present them! More information on the topic is always good.
” In essence, what you’ve said is that others have made a different life choice than yourselves…and they must be nuts for having done so.”
I have never said that soldiers must be nuts; I do say that soldiers have a higher risk of mental disorders than other people, but this is probably an effect rather than a cause. I also said that military life has a lot of stress associated with it; isn’t that the whole point of boot camp, to get you used to operating under very stressful conditions?
“Then, when you’re quite caught up, close your pie hole and support the troops through one of the many non-profit groups listed at the url above. Put your money where your mouth is.”
No other professional group, regardless of how stressful or dangerous their job is, would simply demand support from random strangers in a public forum. I’m reminded of a cartoon (one of a series caricaturing people on the Internet): link here.
To IanC: If this moral retraining is exceptionally difficult, I can see why soldiers and others might not undergo it. However, that’s not what you claimed. To quote: “You have the capacity, right now, to transform your moral system so that you view pedophilia as the best thing since sliced bread.” Have you changed your views?
Tom: Please regress the level of confrontation in tone.
That being said: This argument is, again, a strawman. Nowhere in any statements that I have made in the entirety of this thread have I implied that altering one’s moral values was a facile or readily acheived process.
The capacity to do a thing is not an implication of that thing being facile in nature. I said only that it was possible: it was your reading that created this implication of my argument to a facile moral revision. I have refuted this repeatedly; your arguments continue to address a position I have never held.
At this point, we’re simply arguing semantics over what our statements might or might not mean, which doesn’t help anyone. If you have any more points that don’t involve semantics or pointing out logical fallacies, please list them and I will be happy to resume discussion.
“If you know any good surveys which have been done about soldier’s opinions of combat or Army life in general, please present them! More information on the topic is always good.”
Uh…the references listed above will get ya where ya wanna go, there, Tom. If you are at all interested, you’ll do the research and you’ll find that your words about the topic are empty. (Once you’ve done the research, I will be waiting for your apology.)
BTW. The recorded comments above speak for themselves. If strongly worded, I accurately described what was written and what was implied.
While I do appreciate the fact that people are asking questions and posing scenarios about the ethics of enhancement, we would do well to begin with examining all of our own priors and assumptions.
Even as an ‘example,’ the assumption about military members that was made above is not at all born out by the facts. I invite you to do the research, examine both sides of the issue, and use multiple resources and references.
Currently, there are about twenty-six million military Veterans in the United States. There are over two million active duty military members. How many of us do you think are (PTSD) nuts? One per cent? Five percent? Fifty percent? Ninety per cent? How does the number of nuts in the military/Veteran population compare with the number of (PTSD) nuts in the “civvie” population? Hint: You’ll find that number is a bit lower in the military/Veteran population.
You’ll also find that the military is ahead of the civvie population in education, and that most service members are from middle-class families. You’ll also find that the casualties in all of the wars America has ever fought are over-represented by those with Caucasian and middle-class backgrounds.
Nothing that the media (or your liberal college professors) has told you about these issues is true, folks. If you ignore the facts, no matter your discipline, you are operating on nothing but presumption, ignorance and arrogance.
Whether you want to admit it or no, the latter three will influence any and all decisions you hope to make about AGI, human enhancement or future tech. Through a “bias of design,” you could be laying the framework for the very dystopia you fear.
” Uh…the references listed above will get ya where ya wanna go, there, Tom.”
The first link was to military bloggers; the second was to a study on the demographic characteristics of recruits. What do either have to do with studies on soldier’s attitudes towards Army life?
“I invite you to do the research, examine both sides of the issue, and use multiple resources and references.”
I, quite frankly, do not know where you would find research dealing with the attitudes of soldiers towards combat, killing, etc. Please assist me.
“How many of us do you think are (PTSD) nuts? One per cent? Five percent? Fifty percent? Ninety per cent?”
I honestly have no idea, but you can safely assume that it’s a lot higher than for civilians, because the condition was first recognized on the battlefield (“shell shock” back in WWI) before people applied it to other events.
“Hint: You’ll find that number is a bit lower in the military/Veteran population.”
Did you ever consider that this might be because veterans are only ~9% of the country’s population?
“You’ll also find that the military is ahead of the civvie population in education, and that most service members are from middle-class families. You’ll also find that the casualties in all of the wars America has ever fought are over-represented by those with Caucasian and middle-class backgrounds.”
Er, even if all this is true, what does that have to do with any of the things I was talking about?
“Nothing that the media (or your liberal college professors)”
I literally do not know the political affiliations of any of my college professors. I never bothered to ask as politics was a non-issue. Are you going to launch into the standard rant on how the liburul media is suppressing the truth?
“Through a “bias of design,” you could be laying the framework for the very dystopia you fear.”
I utterly fail to see the connection between AI and military politics. Please explain.
Buy viagfa online….
Buy viagfa online….
Cialis best price buy online….
Online pharmacy buy cialis. Buy cialis….
elasticpintrefer.cn…
elasticpintrefer.cn…
Best you should edit the blog title Accelerating Future » AI-Related Poll at IEET to more suited for your webpage you write. I loved the blog post withal.