Eleven Ways to Avoid an Extremely Bad Singularity Friday, Oct 9 2009
singularity 10:12 am
See Ben Goertzel’s summary of a discussion at the post-Singularity Summit workshop, Eleven Ways to Avoid an Extremely Bad Singularity. A great post, nice to see closer analysis of the issue after so many summaries and popular articles. (In the comments, Ben says: First, I just want to clarify one thing that was stated in my original H+ blog post (you know this, but it may not be clear to your readers): that blog post of mine was a summary of some ideas tossed around in a discussion in the post-Summit workshop … it wasn’t a disquisition on my own views, nor an endorsement of the ideas mentioned!) Here is my general reaction:
I consider 1, “Human-enforced fascism”, relatively unlikely but certainly possible, especially with molecular nanotechnology (MNT) or something analogous. With sufficiently fast and flexible MNT, a small group could probably take over the world. Luckily, very few people think that MNT is possible anytime soon and it may be that it isn’t possible period. I certainly hope so, because it would make everything unnecessarily complicated. I like the fact that I seem able to understand the broad outlines of a world without MNT.
“Friendly” AGI fascism, 2, is also a concern, but less than than the above, in my opinion. It might be extremely difficult to limit AGI intelligence to a certain level once it gets going. The scenario of someone creating a superintelligent AGI and somehow being able to control it I usually just call “magic”, and the respective user “magician”. This is probably impossible, and simply leads to a classic paperclip scenario.
3, “AGI and/or upload panspermia”, is more of an optional side action than a strategy for avoiding a bad Singularity. It basically involves sending AGIs or human uploads out in every direction as fast as you can. I don’t consider it a very good strategy for running away from an unfriendly AI because you eventually have to stop and colonize somewhere, and that’s when they get you. I also very much doubt that you will be able to build a spaceship that travels at near the speed of light without randomly exploding or bathing you in cosmic rays unless you yourself are extremely superintelligent and near-omniscient anyway. At that point you’ve already passed the last relevant hurdle between humanity and a beneficial Singularity. (Unless you yourself are malevolent or selfish.)
4 is creating a virtual AI world sandbox. Great idea, I just hope that programmers don’t get lazy because they think they have a “leakproof” situation. David Chalmers talked about this a lot in his talk at the Summit, and I think he’s simply wrong about the relative difficulty of containing and/or fooling an AGI. Only a fool tries to fool a superintelligence.
5 is “Build an oracular question-answering AGI system, not an autonomous AGI agent”. I doubt this would work — there is no solid difference between question-answering and taking real actions in the world. For instance, an AGI would need to be allotted some intelligence-gathering action-potential to answer questions. Preventing these actions from profoundly impacting the world would be difficult. This falls into the family of containment/magic scenarios whose plausibility is very speculative. Since the whole human race (including you) goes bye-bye if you mess up, it’s probably a bad idea to even think in these terms. Just Say No to Constructing Superintelligent AI Oracles.
6 is “Create upgraded human uploads or brain-enhanced humans first”. Good luck. Historical examples of individual humans getting ultra-powerful usually involve them going slightly crazy and conquering everything around them. We can actually talk about Friendly AI and come up with some consensus on what it should be, then implement it, but with intelligence enhancement you are entirely subject to the whims and beliefs of some random dude or gal. With Friendly AI you can make an entity that is almost entirely unbiased with respect to human cultures and institutions, with an enhanced human you can’t. There is no one I trust enough, including myself, to do this. The first substantially enhanced human intelligence is more likely to be the sort of person that can claw their way to the top of some transhumanist-engineer social group, not the type of person bred from birth to do it. (Which would be more appropriate, and still might fall short.) If we had to pick anyone to upgrade, I’d choose an emotionally stable 10-year-old boy or girl, because this is around as smart and developed humans get before being hit by a wave of pubescent cognitive changes, including lust, adult-like social competition, and more Machiavellian crap that makes people start thinking in zero-sum terms. Power and lust in men leads to rape, and rape leads to chaos. I don’t know how it works with women.
7 is CEV, Coherent Extrapolated Volition. I personally think that this is the best idea out there. I disagree with Ben that this isn’t well-definable in practice, though it certainly isn’t well-defined as it stands. It is clearly fleshed out enough that you could have someone (a team of AGI programmers) specify it in more detail until they have something. Still, would it be well-specified enough to work? It looks like a better idea than most others that have been proposed. The one-sentence version is, “In poetic terms, our coherent extrapolated volition is our wish if we knew more, thought faster, were more the people we wished we were, had grown up farther together; where the extrapolation converges rather than diverges, where our wishes cohere rather than interfere; extrapolated as we wish that extrapolated, interpreted as we wish that interpreted.” Coding this as a goal system sounds like a pain in the ass in terms of the technical challenges, but perhaps necessary if we want to avoid scenario 6. It would be nice to have a toy system of this that actually worked on a limited scale. I have no idea what that would cost, probably a lot. (A lot of the costs would be creating an underlying intelligence powerful enough to model human preferences in detail and “interpret them as we wish that interpreted”.) Its value and convincingness would obviously scale with the underlying intelligence. Places like SecondLife would be the test grounds, and hopefully we would have lots of pre-human AGI behavior data before ascending to truly general intelligence.
8 is “Individual Extrapolated Volition” (IEV). Unfortunately, human beings are inherently social beings, so you have to model social interaction in the extrapolation process. Not a minute doesn’t go by for most of us that we aren’t submerged in human artifacts. Only a man naked in the wilderness is truly alone. The individual aspect is slightly appealing due to the fact that it would be apparently easier than CEV, but wouldn’t it be “stealing the Singularity” to have just one person’s volition extrapolated? I feel morally uncomfortable about the idea. I’m not sure whether it’s rational or irrational, but CEV “feels” better to me. Of course, if it’s technically much more difficult, we might be forced to pick IEV whether we like it or not.
9 is “Make a machine that puts everyone in their personal dream world” is what I call an “Evermore scenario” because it is the basis of the plot of the video game Secret of Evermore. It also happens in the truly terrifying P.K. Dick story Eye in the Sky. Anyone thinking about messing around with advanced AI should read the latter. In that book you get to see what would happen if fundamentalist Christians got their hands on the AI, as well as other scenarios. Basically, I think this is a bad idea. I think it should be strongly discouraged, in fact. Social connections and consensus and the need for stakeholder politics (to steal a line from Dale Carrico) could go away after the Singularity if we wanted them to, but that would lead to scenario #9, which I think is a soulless one. The only way to make politics go away would be to make the connections between people go away. You could do that if you surrounded yourself with puppets and zombies. After living in a world like this for a few years, a lot of people would probably simply forget about the real one and start “whacking off” (metaphorically or literally) for decades on end. Not a happy scenario. A beautiful world can’t be created out of one man’s imagination.
10 is “Engineer a very powerful nonhuman AGI that has a beneficial goal system”. Various people have various thoughts on this, which I have consolidated as a self-styled Friendly AI librarian. Feel free to mention any links not in that list. Essentially, while I think various people have good ideas, including Ben, Eliezer’s ideas really stand out for me in their rigor and plausibility. People might say, “of course you would say that, you work for SIAI”, but the thing is, if I started to change my mind, it would make a load of sense to talk about it, because without a good Friendly AI theory, we might all be screwed. I’d rather live to 1000 than please an intellectual figure or social group. Still, Eliezer has avoided many obvious pitfalls that practically everyone else seems to fall into. Ben’s concept of Voluntary Joyous Growth and other thoughts are interesting but even more poorly specified than CEV. Am I missing a major paper here? Ben also references Joyous Growth on SL4 here. I really hate mailing lists — why don’t people talk on blogs more so that there is more of a public record and discussion? Most mailing lists are not private anyway.
CEV is exciting to me because it “solves” the problem of morality by avoiding it almost entirely. Instead of trying to specify the necessary parameters for Friendliness acquisition directly (before I just said “a morality” here, which Ben pointed out was a strawman. Instead, now I am saying “90% of all non-careful Friendly AI programmers” would believe that directly programming the “necessary parameters for Friendliness acquisition” would work better than letting coherent human preferences dictate those parameters) it just points to human minds and says, “suck the morality out of that thing, then average it with others”. The approach of using an extrapolition algorithm inherently makes it difficult if not impossible to manipulate it in directions that you personally like, at the expense of other people’s goals. With most other goal systems, including Joyous Growth, it seems to rely on us trusting the programmer to provide good “basic values”. With CEV, the programmer is forced only to use what can be extracted from real human preferences. No one gets more than one “vote”. It is the most democratic scheme yet presented for an AI goal system, for those who get all excited by the word “democratic”. I think that some things are best done democratically and some aren’t. Providing goal system content for an AGI whose goal is to manifest our collective volition is definitely something best done democratically.
11, “Let humanity die the good death” through gradual intelligence enhancement seems like another one of those optional side issues and not really a way to avoid a bad Singularity. The way it is phrased here seems especially distracting and likely to confuse people new to this stuff. (More than they already are.)

Thanks for the reactions Mike!
First, I just want to clarify one thing that was stated in my original H+ blog post (you know this, but it may not be clear to your readers): that blog post of mine was a summary of some ideas tossed around in a discussion in the post-Summit workshop … it wasn’t a disquisition on my own views, nor an endorsement of the ideas mentioned!
Now some random reactions to some of your reactions…
I continue to think CEV is an ill-defined and non-workable idea ;-) … and I agree that #10 is not yet adequately well-defined, yet it feels to me more achievable than CEV or even IEV (though I can’t prove that, yet).
Also, I don’t know ANY AGI programmers who believe that programming an AGI with morality directly will work… except one: Selmer Bringsjord. It’s pretty much a “straw man” perspective. Most of the AGI developers I know are in favor of programming an AGI with a rational goal system and then providing it with specific (morally acceptable) goals mainly through teaching the system, as one does with a human child. I’m not claiming this is a foolproof methodology, just pointing out that it’s a more common view among AGI developers than the notion of a “preprogrammed morality.”
About “AGI Fascism”, I think this could be quite feasible, actually. One could design an AGI system with a very non-human-like intelligence, but a great amount of analytical power, designed to rule the world. This seems vastly easier than building a CEV machine, for example…. I dislike this outcome for other reasons, but I suspect it would be feasible…
– Ben Goertzel
Hi Ben, I made various updates to the text to respond to some of your concerns and clarify that you do not necessarily endorse these ideas.
Could you make 1 or 2 primary criticisms of the implementability of CEV, just so I have a general idea, or point me to where you already wrote that? You mention it a little in your papers, and I will reread your AI morality papers today to check. I agree that CEV is poorly specified but am not sure why, at the very least, a lazy, fuzzy implementation wouldn’t work at all.
OK, I qualified my comment about “directly programmming morality” to directly programming “parameters for Friendliness acquisition”. I think this statement more directly grasps what I was trying to say. To say that an AI acquires morality by “learning” still means that it will have programming that cuts down the combinatorially large space of possible moralities to something commonly seen as human-friendly. There are two options for this programming — it will either by programmed relatively directly (unless learning also changes the way the AI learns in a structured way), or it will be set by some processed products of human preferences (CEV or any other extrapolation approach).
To my mind, it would be very difficult to separate “analytics” from “experimentation and action”. “Analytical power” as something separate from “optimization pressure in general” is a human invention. (Sure, everything is a human invention, including all morality, but it seems easier to say “take this large, coherent, knowably consistent-under-reflection moral system” rather than “take this moral system based on interpreting my (and only my) commands the way I want them interpreted, and act as an oracle or magic engine without actually taking my “agency” (poorly defined) away from me”. How would the AI distinguish its master from other people? If its master lost those defining features (by dying, say), then what does the AI do? Just keep performing the last command it was left with? Shut down neatly? Go critical? The challenge of those master-definition problem and power-transfer problems seem sufficiently large that I doubt anyone would mess with it, but I could be wrong.
-MA
Responding to your response to 6:
I doubt the first upload will be an important person. It’s more likely to be someone terminally ill, in my opinion. Not that it matters, because I doubt the first upload will have unrestricted access (or any access at all) to its own code.
It’s not at all clear to me that self modification will be easy for an upload. Unlike in an AI scenario, the upload’s code will likely be very opaque (spaghetti code), preventing easy insight into making it better, and perhaps ultimately limiting the extent to which enhancement can occur without abandoning the inherent architecture of the human mind. This suggests a much slower ‘FOOM’ to me. Of course, a FOOM is still a FOOM, and at some point becomes uncontrollable from a human vantage point, but the slower speed might be enough to allow multiple uploads to stay relatively even with each other.
In short: of course I wouldn’t trust the future to a single upload, but I would potentially trust it to a group of thousands of uploads.
See Magister Goertzel’s summary of a discussion at the post-Examancy summit conclave, Eleven Ways to Avoid an Extremely Bad Examantic Event. A great scroll, nice to see closer analysis of the issue after so many summaries and imps sending out hue-and-calls. (In the comments, Magister G. says: First, I just want to clarify one thing that was stated in my original ascension guild main hall posting (you know this, but it may not be clear to your readers): that guild post of mine was a summary of some ideas tossed around in a discussion in the post-Summit workshop … it wasn’t a disquisition on my own views, nor an endorsement of the ideas mentioned!) Here is my general reaction:
I consider 1, “Society turning all Lawful Evil” relatively unlikely but certainly possible, especially by means of widespread summoning and binding of Elementals (S&BE) or something analogous. With sufficiently fast and flexible S&BE, a small group of elementalism-specialized wizards could probably take over the world. Luckily, very few people think that such a degree of elementalism is possible anytime soon and it may be that it isn’t possible period. I certainly hope so, because it would make everything unnecessarily complicated. I like the fact that I seem able to understand the broad outlines of a world without robust elementalism.
Governance by “Friendly” Djinn, 2, is also a concern, but less than than the above, in my opinion. It might be extremely difficult to limit Djinn intelligence to a certain level once it gets going – they can wish one another more smart all the time. The scenario of someone calling forth a really high level Djinn and somehow being able to control it I usually just call “gullible”, and the respective spellcaster “cleric”. This is probably impossible, and simply leads to a classic pot of gold scenario.
3, “Lichcraft and/or necromancy epidemic”, is more of an optional side action than a strategy for avoiding a bad Examancia. It basically involves sending Liches, Vampires and infectious eating zombies out in every direction as fast as you can. I don’t consider it a very good strategy for running away from something like Demons because you eventually have to stop and settle down somewhere, and that’s when the demons turn out to have high levels in turning undead. I also very much doubt that you will be able to build a skygalleon that travels at near the speed of a Dragon without running stuck in really awful random encounters or cosmic sargasso realms, unless you yourself are extremely
high level, and have attributes over 25. At that point you’ve already passed the last relevant hurdle between humanity and Demons. Or Slaadi. (Unless you yourself have become evil in turning undead).
4 is becoming one of the fey and hiding in faerie hills. Great idea, I just hope that the elve queens don’t get lazy because they think they have racial and realm bonuses when facing lower planar creatures. Great-Sorceror Chalmers talked about this a lot in his talk at the Summit, and I think he’s simply wrong about the relative difficulty of containing and/or fooling the minions of Demogorgon. Only a fool tries to fool a a Greater Demon.
5 is “Build a mirror of question-answering, not try and create some extremely mighty Golem”. I doubt this would work — there is no solid difference between question-answering and taking real actions in the world. For instance, a Golem would need to be allotted some intelligence-gathering action-potential to answer questions. Preventing these actions from profoundly impacting the world would be difficult. This falls into the family of containment/magic scenarios whose plausibility is very speculative. Since the whole human race (including you) goes bye-bye if you mess up, it’s probably a bad idea to even think in these terms. Just Say No to Magic Mirrors or Golems.
6 is “Polymorph humans into superior species, such as Githyanki, or Drow or Draenai.” Good luck. Historical examples of individual humans getting ultra-powerful usually involve them going slightly crazy and conquering everything around them. We can actually talk about level advancement in the ideal combination of classes and come up with some consensus on what it should be, then implement it, but with
these freaky humanoid races you are entirely subject to the whims and beliefs of some random dude or gal. With Friendly Dragons you can make an entity that is almost entirely unbiased with respect to human cultures and institutions, with permanently polymorphed humans you can’t. There is no one I trust enough, including myself, to do this. (Well except this succubus I know, she is really nice.) The first substantially enhanced human intelligence is more likely to be the sort of person that can claw their way to the top of some magical guild, not the type of person bred from birth to do it. In fact I heard of a Beholder that did just that and it was a mess. (Which would be more appropriate, and still might fall short.) If we had to pick anyone to turn into a Higher Being, I’d choose an emotionally stable 10-year-old boy or girl, because this is around as smart and developed humans get before being hit by a alignment and original sin, including lust, adult-like social competition, and more Machiavellian crap that makes people start thinking in zero-sum terms. Power and lust in men leads to rape, and rape leads to chaos. I don’t know how it works with women. Well I do know, ask my succubus friend, the rumor I have not cast level-9 spells for a while is highly overrated, nothing a Restoration won’t cure, and the hiatus was well worth it.
7 is CEV, Simple Old Fashioned Levelling. I personally think that this is the best idea out there. I disagree with Magister G. that this isn’t well-definable in practice, though it certainly isn’t well-defined as it stands. It is clearly fleshed out enough that you could have someone (a team of friendly NPC guildmaster, master thieves, high-clerics all giving out quests) specify it in more detail until they have something. Still, would it be well-specified enough to work? It looks like a better idea than most others that have been proposed. The one-sentence version is, “In poetic terms, us being higher level is our wish if we knew more, thought faster, were more the people we wished we were, had grown up farther together; where the extrapolation converges rather than diverges, where our wishes cohere rather than interfere; extrapolated as we wish that extrapolated, interpreted as we wish that interpreted.” Coding this as a goal system sounds like a pain in the ass in terms of the technical challenges, but perhaps necessary if we want to avoid scenario 6. It would be nice to have a toy system of this that actually worked on a limited scale. I have no idea what that would cost, probably a lot. (A lot of the costs would be creating an underlying intelligence powerful enough to model human preferences in detail and “interpret them as we wish that interpreted”.) Its value and convincingness would obviously scale with the underlying intelligence. Places like the Forgotten Realms would be the test grounds, and hopefully we would have lots of level 15+ or so dungeon data before ascending to truly epic levels.
8 is characters taking these really nasty class combo’s and minmaxing.
Unfortunately, human beings are inherently social beings, so you have to model social interaction in the class process. Where are you going to learn all these classes and where will you have to travel? Where can I learn the arts of the ninja? Where can I learn decent Psionics? Not a minute doesn’t go by for most of us that we aren’t submerged in magical artifacts. Only a ranger naked in the wilderness is truly alone. The munchkin/twink aspect is slightly appealing due to the fact that it would be apparently easier than leveling in one class, but wouldn’t it be “stealing the Examancy” to have just one person’s preferences extrapolated? I feel morally uncomfortable about the idea. I’m not sure whether it’s rational or irrational, but being just a Mage “feels” better to me. Of course, if it’s technically much more difficult, we might be forced to pick shaman-ninja-barbarian-warlock multiclasses whether we like it or not.
9 is “Make a spell that puts everyone in their personal dream world” is what I call an “Dante scenario” because this is what happened to this Wizard called Dante, and he just dissappeared one day and we haven’t seen him since. It also happened in those dreams those crazy people had a few years ago – and then thet dissolved into nothing, screaming. Anyone thinking about messing around with otherplanars and elementals should remember the latter. The Necronomicon tells you what would happen if the followers of St.Cuthbert got their hands on these spells, as well as other scenarios. Basically, I think this is a bad idea. I think it should be strongly discouraged, in fact. Social connections and consensus and the need for stakeholder politics (to steal a line from Magister Darth Quixote – what is his spellcasting class, btw?) could go away after the Examantic event if we wanted them to, but that would lead to scenario #9, which I think is a soulless one. The only way to make politics go away would be to make the connections between people go away. You could do that if you surrounded yourself with puppets and zombies. After living in a world like this for a few years, a lot of people would probably simply forget about the real one and start “whacking off” (metaphorically or literally) for decades on end. Not a happy scenario. A beautiful world can’t be created out of one man’s imagination. Well, my succubus does a fair job, but that’s not the same.
10 is “Summon a very powerful Dragon that is Lawful Good”. Or if you really insist, Neutral Good.” Various people have various thoughts on this, which I have consolidated as a self-styled Golden Dragon librarian. Feel free to mention any links not in that list. Essentially, while I think various people have good ideas, including Magister G, Magister Eliezer’s ideas really stand out for me in their rigor and plausibility. People might say, “of course you would say that, you work for Hogwarts”, but the thing is, if I started to change my mind, it would make a load of sense to talk about it, because without a good Golden Dragon scenario, we might all be screwed. I’d rather polymorph and live with the wood elves to 1000 years old than please an savant or bard. Still, Magister Eliezer has avoided many obvious pitfalls that practically everyone else seems to fall into. Magister G’s concept of Making Everyone Chaotic Good and other thoughts are interesting but even more poorly specified than everyone rolling up Githzerai characters. Am I missing a major paper here? Ben also references Forced Alignment Change in the chantry here. I really hate imp messengers — why don’t people talk using winged scrolls more so that there is more of a public record and discussion? Most Imps are terrible gossips anyway.
Leveling classes that are only allowed to be good aligned is exciting to me because it “solves” the problem of free will by avoiding it almost entirely. Instead of trying to specify the necessary parameters for good alignment acquisition directly (before I just said “a morality” here, which Magister G. pointed out was the straw man. Instead, now I am saying “90% of all summoners, conjurers and elementalists (and necromancers)” would believe that directly programming the “necessary parameters for good aligned adventure” would work better than letting coherent human preferences dictate the strategies of the Kingdom) it just points to human minds and says, “make them all true Neutral, then average its attributes with others”.
Khannea, I assume you are making fun of my post because the way it uses niche jargon and concepts? If you actually read the papers behind the words, you would find various reasonable arguments referring to real ideas, not fantasies like D&D. It is natural to develop new jargon to describe concepts in areas new to human knowledge. Numerous branches of science and philosophy do the same thing all the time. The key is precisely defining the words and concepts used. If you want to make fun of a post, next time, make it shorter. The point you are making can be made in a few paragraphs only.
Don’t be silly, I am well aware of the jargon and the concepts behind it (and agree more or less), it just popped in my mind and well, you know my lame (lack of any sense of) humor.
Mike …
Yes I and most other AGI-ers think it’s likely feasible to program an AGI goal framework whose structure and dynamics implicitly constitutes a (high-dimensional) space of “parameters for Friendliness acquisition.” So I accept your rephrasing on that point, even though it’s not exactly the way I would say it…
About CEV: some obvious problems are…
1) it seems like it will require much more advanced technology than creating an AGI with, say, 2x human intelligence. So even if it were possible and sensible, it seems unlikely to happen at the stage where humans are still in control. Maybe the Guardian AGI will be the guy to perform the CEV experiment.
2) it’s not clear that “the person I wish I were” is a well-defined concept. At the post-Singularity workshop, an actual professional moral philosopher was present, and he opined that it is NOT a well-defined concept, thus rendering CEV non-workable and ill-defined. It may be that minor variations in the way one defines this ill-defined concept, cause large variations in the “extrapolated volition” obtained. In fact I suspect this is the case for many people though not all.
3) different people have wildly incoherent ideas about what’s good. I doubt the extrapolated Ben has compatible ideas about the world as the extrapolation of the average Islamic mullah, for example … or even the average Walmart clerk. I’m a weird dude.
IEV solves complaint 3, and partially solves complaint 1, but doesn’t touch complaint 2.
I’m sure these are not the only objections to the fanciful but fascinating CEV idea, but I don’t have more time to search my memory or re-think the issues at the moment…
– Ben Goertzel
Khannea, I find it trite and mocking. It’s hard to evaluate over the Internet where exactly your social intuitions are. Your comments are always so long and often rambling — can you try to be more concise? You also seem to enjoy theatrics in your comments — why not just drill directly to the core of the issue and make the simplest statements possible? It’s important to me that comments be as meaningful and thought-provoking as possible.
Ben, thanks. I think the differences we are talking about are subtle enough that there has to be a programmatic distinction. Presumably one could derive a basic-order description of the person I wish I were through a questionnaire asking me about that very issue. It could thereby be precisely defined, though not at a great level of resolution. The crucial question is whether the resolution could be improved enough to lead to an AI that creates a model of my volition that comes up with interesting and surprising new conclusions that make sense.
I don’t think it’s too fanciful to discuss extrapolations — in my opinion, anything that works must be an extrapolation of some sort. The alternative, copying an existing human morality, seems like a bad idea. The question is not to use an extrapolation or not, but which extrapolation we should use.
-MA
I’d be careful about dismissing molecular manufacturing, particularly if the implications make you uncomfortable. That’s the same kind of reasoning (if we can call it that) used against artificial intelligence and life extension. Indeed, I see no reason to consider those two things more likely than nanofactories. Plenty of futurists seen MNT coming as quickly as AGI, if not sooner. See Drexler, Hall, Freitas, the CRN folks, Kurzweil, and so on. Each technology has the potential to enable the other.
Hi Ben, there is no connection between my emotional appraisal of MNT and the thoughts that I use to estimate its plausibility. I weave them together in the text partially just to make it sound more interesting. MNT has the potential to enable brute-force AGI where we are able to implement intelligence without thoroughly understanding its goal system, leading to global destruction.
The parameters you are discussing are interesting. Keep in mind though, the real danger is NOT the AGI, but the human.
Not long ago, I relayed to Ben it comes down to the three ‘B’s'… booze, booty and bounty. Without going into the humorous anecdotes, any dangers will be more attached to that because of the human, not the machine.
Starting with the computer, if an AGI level unit were sitting on the desk, it will not drink and/or feel the need associated with procreation. To build a bot that can drink is, well a waste of time, for the only liquid it may need is in the form of lubrication if that is the case.
Both the computer and bot will not need the money, for as long as they have the power source, it can function.
It is the human who has the drive that centers on the three B’s and how that drive often moves toward bad outcomes. As long as resources are few among a nation, the leader will take from everyone and have his drinking and sex parties.
In the case of the nation with many resources, the same thing happens, for due to the weakness and insecurities of man, we often try to have others think we have lots of bounty that pays for the booze to consume while the booty dances.
The vanity associated with that leads to a lot of bounty that could be put to a useful purpose, wasted at best.
My suggestion is placing debate regarding the need of a Guardian AGI more toward protecting all of us from the AGI used by the person who has bad motive, rather than the AGI coming up with having fun via pillage and plunder on it’s own.