Existential Risk and Public Relations
[Added 02/24/14: Some time after writing this post, I discovered that it was based on a somewhat unrepresentative picture of SIAI. I still think that the concerns therein were legitimate, but they had less relative significance than I had thought at the time. SIAI (now MIRI) has evolved substantially since 2010 when I wrote this post, and the criticisms made in the post don't apply to MIRI as presently constituted.]
A common trope on Less Wrong is the idea that governments and the academic establishment have neglected to consider, study and work against existential risk on account of their shortsightedness. This idea is undoubtedly true in large measure. In my opinion and in the opinion of many Less Wrong posters, it would be very desirable to get more people thinking seriously about existential risk. The question then arises: is it possible to get more people thinking seriously about existential risk? A first approximation to an answer to this question is "yes, by talking about it." But this answer requires substantial qualification: if the speaker or the speaker's claims have low credibility in the eyes of the audience then the speaker will be almost entirely unsuccessful in persuading his or her audience to think seriously about existential risk. Speakers who have low credibility in the eyes of an audience member decrease the audience member's receptiveness to thinking about existential risk. Rather perversely, speakers who have low credibility in the eyes of a sufficiently large fraction of their audience systematically raise existential risk by decreasing people's inclination to think about existential risk. This is true whether or not the speakers' claims are valid.
As Yvain has discussed in his excellent article titled The Trouble with "Good"
To make an outrageous metaphor: our brains run a system rather like Less Wrong's karma. You're allergic to cats, so you down-vote "cats" a couple of points. You hear about a Palestinian committing a terrorist attack, so you down-vote "Palestinians" a few points. Richard Dawkins just said something especially witty, so you up-vote "atheism". High karma score means seek it, use it, acquire it, or endorse it. Low karma score means avoid it, ignore it, discard it, or condemn it.
When Person X makes a claim which an audience member finds uncredible, the audience member's brain (semiconsciously) makes a mental note of the form "Boo for Person X's claims!" If the audience member also knows that Person X is an advocate of existential risk reduction, the audience member's brain may (semiconsciously) make a mental note of the type "Boo for existential risk reduction!"
The negative reaction to Person X's claims is especially strong if the audience member perceives Person X's claims as arising from a (possibly subconscious) attempt on Person X's part to attract attention and gain higher status, or even simply to feel as though he or she has high status. As Yvain says in his excellent article titled That other kind of status:
But many, maybe most human actions are counterproductive at moving up the status ladder. 9-11 Conspiracy Theories are a case in point. They're a quick and easy way to have most of society think you're stupid and crazy. So is serious interest in the paranormal or any extremist political or religious belief. So why do these stay popular?
[...]
a person trying to estimate zir social status must balance two conflicting goals. First, ze must try to get as accurate an assessment of status as possible in order to plan a social life and predict others' reactions. Second, ze must construct a narrative that allows them to present zir social status as as high as possible, in order to reap the benefits of appearing high status.
[...]
In this model, people aren't just seeking status, they're (also? instead?) seeking a state of affairs that allows them to believe they have status. Genuinely having high status lets them assign themselves high status, but so do lots of other things. Being a 9-11 Truther works for exactly the reason mentioned in the original quote: they've figured out a deep and important secret that the rest of the world is too complacent to realize.
I'm presently a graduate student in pure mathematics. During graduate school I've met many smart people who I wish would take existential risk more seriously. Most such people who have heard of Eliezer do not find his claims credible. My understanding is that the reason for this is that Eliezer has made some claims which they perceive to be falling under the above rubric, and the strength of their negative reaction to these has tarnished their mental image of all of Eliezer's claims. Since Eliezer supports existential risk reduction, I believe that this has made them less inclined to think about existential risk than they were before they heard of Eliezer.
There is also a social effect which compounds the issue which I just mentioned. The issue which I just mentioned makes people who are not directly influenced by the issue that I just mentioned less likely to think seriously about existential risk on account of their desire to avoid being perceived as associated with claims that people find uncredible.
I'm very disappointed that Eliezer has made statements such as:
If I got hit by a meteorite now, what would happen is that Michael Vassar would take over sort of taking responsibility for seeing the planet through to safety...Marcello Herreshoff would be the one tasked with recognizing another Eliezer Yudkowsky if one showed up and could take over the project, but at present I don't know of any other person who could do that...
which are easily construed as claims that his work has higher expected value to humanity than the work of virtually all humans in existence. Even if such claims are true, people do not have the information that they need to verify that such claims are true, and so virtually everybody who could be helping out assuage existential risk find such claims uncredible. Many such people have an especially negative reaction to such claims because they can be viewed as arising from a tendency toward status grubbing, and humans are very strongly wired to be suspicious of those who they suspect to be vying for inappropriately high status. I believe that such people who come into contact with Eliezer's statements like the one I have quoted above are less statistically likely to work to reduce existential risk than they were before coming into contact with such statements. I therefore believe that by making such claims, Eliezer has increased existential risk.
I would go further than that and say that that I presently believe that donating to SIAI has negative expected impact on existential risk reduction on account of that SIAI staff are making uncredible claims which are poisoning the existential risk reduction meme. This is a matter on which reasonable people can disagree. In a recent comment, Carl Shulman expressed the view that though SIAI has had some negative impact on the existential risk reduction meme, the net impact of SIAI on the existential risk meme is positive. In any case, there's definitely room for improvement on this point.
Last July I made a comment raising this issue and Vladimir_Nesov suggested that I contact SIAI. Since then I have corresponded with Michael Vassar about this matter. My understanding of Michael Vassar's position is that the people who are dissuaded from thinking about existential risk because of remarks like Eliezer's are too irrational for it to be worthwhile for them to be thinking about existential risk. I may have misunderstood Michael's position and encourage him to make a public statement clarifying his position on this matter. If I have correctly understood his position, I do not find Michael Vassar's position on this matter credible.
I believe that if Carl Shulman is right, then donating to SIAI has positive expected impact on existential risk reduction. I believe that that even if this is the case, a higher expected value strategy is to withold donations from SIAI and informing SIAI that you will fund them if and only if they require their staff to exhibit a high degree of vigilance about the possibility of poisoning the existential risk meme by making claims that people find uncredible. I suggest that those who share my concerns adopt the latter policy until their concerns have been resolved.
Before I close, I should emphasize that my post should not be construed as an attack on Eliezer. I view Eliezer as an admirable person and don't think that he would ever knowingly do something that raises existential risk. Roko's Aspergers Poll suggests a strong possibility that the Less Wrong community exhibits an unusually high abundance of the traits associated with Aspergers Syndrome. It would not be at all surprising if the founders of Less Wrong have a similar unusual abundance of the traits associated with Aspergers Syndrome. I believe that more likely than not, the reason why Eliezer has missed the point that I raise in this post is social naivete on his part rather than willful self-deception.
Loading…
Subscribe to RSS Feed
= f037147d6e6c911a85753b9abdedda8d)
Comments (613)
The one "uncredible" claim mentioned - about Eliezer being "hit by a meteorite" - sounds as though it is the kind of thing he might plausibly think. Not too much of a big deal, IMO.
As with many charities, it is easy to think the SIAI might be having a negative effect - simply because it occupies the niche of another organisation that could be doing a better job - but what to do? Things could be worse as well - probably much worse.
I suggested what to do about this problem in my post: withhold funding from SIAI, and make it clear to them why you're withholding funding from them, and promise to fund them if the issue is satisfactorily resolved to incentivize them to improve.
Will you do this?
I'm definitely interested in funding an existential risk organization. SIAI would have to be a lot more transparent than it is now right now for me to be interested in funding SIAI. For me personally, it wouldn't be enough for SIAI to just take measures to avoid poisoning the meme, I would need to see a lot more evidence that SIAI is systematically working to maximize its impact on existential risk reduction.
As things stand I prefer to hold out for a better organization. But if SIAI exhibited transparency and accountability of levels similar to those of GiveWell (welcoming and publically responding to criticism regularly, regularly posting detailed plans of action, seeking out feedback from subject matter specialists and making this public when possible, etc.) I would definitely fund SIAI and advocate that others do so as well.
"transparency"? I thought the point of your post was that SIAI members should refrain from making some of their beliefs easily available to the public?
I see, maybe I should have been more clear. The point of my post is that SIAI members should not express controversial views without substantiating them with abundant evidence. If SIAI provided compelling evidence that Eliezer's work has higher expected value to humanity than what virtually everybody else is doing, then I would think Eliezer's comment appropriate.
As things stand SIAI has not provided such evidence. Eliezer himself may have such evidence, but if so he's either unwilling or unable to share it.
If there really was "abundant evidence" there probably wouldn't be much of a controversy.
Now that is unfair.
Since 1997, Eliezer has published (mostly on mailing lists and blogs but also in monographs) an enormous amount (at least ten novels worth unless I am very mistaken) of writings supporting exactly that point. Of course most of this material is technical, but unlike the vast majority of technical prose, it is accessible to non-specialists and non-initiates with enough intelligence, a solid undergraduate education as a "scientific generalist" and a lot of free time on their hands because in his writings Eliezer is constantly "watching out for" the reader who does not yet know what he knows. (In other words, it is uncommonly good technical exposition.)
Can you be more specific than "it's somewhere beneath an enormous amount of 13 years of material from the very same person whose arguments are scrutinized for evidence"?
This is not sufficient to scare people up to the point of having nightmares and ask them for most of their money.
Do you want me to repeat the links people gave you 24 hours ago?
The person who was scared to the point of having nightmares was almost certainly on a weeks-long or months-long visit to the big house in California where people come to discuss extremely powerful technologies and the far future and to learn from experts on these subjects. That environment would tend to cause a person to take certain ideas more seriously than a person usually would.
It was more than one person. Anyway, I haven't read all of the comments yet so I might have missed some specific links. If you are talking about links to articles written by EY himself where he argues about AI going FOOM, I commented on one of them.
Here is an example of the kind of transparency in the form of strict calculations, references and evidence I expect.
As I said, I'm not sure what other links you are talking about. But if you mean the kind of LW posts dealing with antipredictions, I'm not impressed. Predicting superhuman AI to be a possible outcome of AI research is not sufficient. Where is the difference between claiming the LHC will go FOOM? I'm sure someone like EY would be able to write a thousand posts around such a scenario telling me that the high risk associated with the LHC going FOOM does outweigh its low probability. There might be sound arguments to support this conclusion. But it is a conclusion and a framework of arguments based on a assumption that is itself of unknown credibility. So is it too much to ask for some transparet evidence to fortify this basic premise? Evidence that is not somewhere to be found within hundreds of posts not directly concerned with the evidence in question but rather arguing based on the very assumption it is trying to justify?
So my impression has been that the situation is that
(i) Eliezer's writings contain a great deal of insightful material.
(ii) These writings do not substantiate the idea that [that Eliezer's work has higher expected value to humanity than what virtually everybody else is doing].
I say this having read perhaps around a thousand pages of what Eliezer has written. I consider the amount of reading that I've done to be a good "probabilistic proof" that the points (i) and (ii) apply to the portion of his writings that I haven't read.
That being said, if there are any particular documents that you would point me to which you feel do provide a satisfactory evidence for the idea [that Eliezer's work has higher expected value to humanity than what virtually everybody else is doing], I would be happy to examine them.
I'm unwilling to read the whole of his opus given how much of it I've already read without being convinced. I feel that the time that I put into reducing existential risk can be used to better effect in other ways.
It would help to know what steps in the probabilistic proof don't have high probability for you.
For example, you might think that the singularity has a good probability of being relatively smooth and some kind of friendly, even without FAI. or you might think that other existential risks may still be a bigger threat, or you may think that Eliezer isn't putting a dent in the FAI problem.
Or some combination of these and others.
Yes, I agree with you. I plan on making my detailed thoughts on these points explicit. I expect to be able to do so within a month.
But for a short answer, I would say that the situation is mostly that I think that:
This might be a convenient place to collect a variety of reasons why people are FOOM denialists. From my POV: 1. I am skeptical that safeguards against UFAI (unFAI) will not work. In part because: 2. I doubt that the "takeoff" will be "hard". Because: 3. I am pretty sure the takeoff will require repeatedly doubling and quadrupling hardware, not just autorewriting software. 4. And hence an effective safeguard would be to simply not give the machine its own credit card! 5. And in any case, the Moore's law curve for electronics does not arise from delays in thinking up clever ideas, it arises from delays in building machines to incredibly high tolerances. 6. Furthermore, even after the machine has more hardware, it doesn't yet have higher intelligence until it reads lots more encyclopedias and proves for itself many more theorems. These things take time. 7. And finally, I have yet to see the argument that an FAI protects us from a future UFAI. That is, how does the SIAI help us? 8. Oh, and I do think that the other existential risks, particularly war and economic collapse, put the UFAI risk pretty far down the priority list. Sure, those other risks may not be quite so existential, but if they don't kill us, they will at least prevent an early singularity.
Edit added two days later: Since writing this, I thought about it some more, shut up for a moment, and did the math. I still think that it is unlikely that the first takeoff will be a hard one; so hard that it gets out of control. But I now estimate something like a 10% chance that the first takeoff will be hard, and I estimate something like a 30% chance that at least one of the first couple dozen takeoffs will be hard. Multiply that by an estimated 10% chance that a hard takeoff will take place without adequate safeguards in place, and another 10% chance that a safeguardless hard takeoff will go rogue, and you get something like a 0.3% chance of a disaster of Forbin Project magnitude. Completely unacceptable.
Originally, I had discounted the chance that a simple software change could cause the takeoff; I assumed you would need to double and redouble the hardware capability. What I failed to notice was that a simple "tuning" change to the (soft) network connectivity parameters - changing the maximum number of inputs per "neuron" from 8 to 7, say, could have an (unexpected) effect on performance of several orders of magnitude simply by suppressing wasteful thrashing or some such thing.
For what definitions of "value to humanity" and "virtually everybody else"?
If "value to humanity" is assessed as in Bostrom's Astronomical Waste paper, that hugely favors effects on existential risk vs alleviating current suffering or increasing present welfare (as such, those also have existential risk effects). Most people don't agree with that view, so asserting that as a privileged frame can be seen as a hostile move (attacking the value systems of others in favor of a value system according to which one's area of focus is especially important). Think of the anger directed at vegetarians, or those who guilt-trip others about not saving African lives. And of course, it's easier to do well on a metric that others are mostly not focused on optimizing.
Dispute about what best reduces existential risk, and annoyance at overly confident statements there is a further issue, but I think that asserting uncommon moral principles (which happen to rank one's activities as much more valuable than most people would rank them) is a big factor on its own.
In case my previous comment was ambiguous, I should say that I agree with you completely on this point. I've been wanting to make a top level post about this general topic for a while. Not sure when I'll get a chance to do so.
I'm planning to fund FHI rather than SIAI, when I have a stable income (although my preference is for a different organisation that doesn't exist)
My position is roughly this.
The nature of intelligence (and its capability for FOOMing) is poorly understood
The correct actions to take depend upon the nature of intelligence.
As such I would prefer to fund an institute that questioned the nature of intelligence, rather than one that has made up its mind that a singularity is the way forward. And it is not just the name that makes me think that SIAI has settled upon this view.
And because the nature of intelligence is the largest wild card in the future of humanity, I would prefer FHI to concentrate on that. Rather than longevity etc.
FHI?
The Future of Humanity Institute.
Nick Bostrom's personal website probably gives you the best idea of what they produce.
A little too philosophical for my liking, but still interesting.
What would the charity you'd like to contribute to look like?
When I read good popular science books the people will tend to come up with some idea. Then they will test the idea to destruction. Poking and prodding at the idea until it really can't be anything but what they say it is.
I want to get the same feeling off the group studying intelligence as I do from that type of research. They don't need to be running foomable AIs, but truth is entangled so they should be able to figure out the nature of intelligence from other facets of the world, including physics and the biological examples.
Questions I hope they would be asking:
Is the g factor related to ability to absorb cultural information? I.e. is peoples increased ability to solve problems if they have a high g due to them being able to get more information about solving problems from cultural information sources?
If it wasn't then it would be further evidence for .something special in one intelligence over another and it might make sense to call one more intelligent, rather than just having different initial skill sets.
If SIAI had the ethos I'd like, we'd be going over and kicking every one of the supporting arguments for the likelihood of fooming and the nature of intelligence to make sure they were sound. Performing experiments where necessary. However people have forgotten them and moved on to decision theory and the like.
Interesting points. Speaking only for myself, it doesn't feel as though most of my problem solving or idea generating approaches were picked up from the culture, but I could be kidding myself.
For a different angle, here's an old theory of Michael Vassar's-- I don't know whether he still holds it. Talent consists of happening to have a reward system which happens to make doing the right thing feel good.
And then, hypothetically, if they found that fooming is not likely at all, and that dangerous fooming can be rendered nearly impossible by some easily enforced precautions/regulations, what then? If they found that the SIAI has no particular unique expertise to contribute to the development of FAI? An organization with an ethos you would like: what would it do then? To make it a bit more interesting, suppose they find themselves sitting on a substantial endowment when they reason their way to their own obsolescence?
How often in human history have organizations announced, "Mission accomplished - now we will release our employees to go out and do something else"?
It doesn't seem likely. The paranoid can usually find something scary to worry about. If something turns out to be not really-frightening, fear mongers can just go on to the next-most frightening thing in line. People have been concerned about losing their jobs to machines for over a century now. Machines are a big and scary enough domain to keep generating fear for a long time.
From a less cynical angle, building organizations is hard. If an organization has fulfilled its purpose, or that purpose turns out to be a mistake, it isn't awful to look for something useful for the organization to do rather than dissolving it.
I think that what SIAI works on is real and urgent, but if I'm wrong and what you describe here does come to pass, the world gets yet another organisation campaigning about something no-one sane should care about. It doesn't seem like a disastrous outcome.
I like this concept.
Assume your theory will fail in some places, and keep pressing it until it does, or you run out of ways to test it.
Right - but that's only advice for those who are already donating. Others would presumably seek reform or replacement. The decision there seems non-trivial.
The point of my post is not that there's a problem of SIAI staff making claims that you find uncredible, the point of my post is that there's a problem of SIAI making claims that people who are not already sold on taking existential risk seriously find uncredible.
Can you give a few more examples of claims made by SIAI staff that people find uncredible? Because it's probably not entirely clear to them (or to others interested in existential risk advocacy) what kind of things a typical smart person would find uncredible.
Looking at your previous comments, I see that another example you gave was that AGI will be developed within the next century. Any other examples?
Good question. I'll get back to you on this when I get a chance, I should do a little bit of research on the topic first. The two examples that you've seen are the main ones that I have in mind that have been stated in public, but there may be others that I'm forgetting.
There are some other examples that I have in mind from my private correspondence with Michael Vassar. He's made some claims which I personally do not find at all credible. (I don't want to repeat these without his explicit permission.) I'm sold on the cause of existential risk reduction, so the issue in my top level post does not apply here. But in the course of the correspondence I got the impression that he may say similar things in private to other people who are not sold on the cause of existential risk.
I second that question. I am sure there probably are other examples but they for most part wouldn't occur to me. The main examples that spring to mind are from cases where Robin has disagreed with Eliezer... but that is hardly a huge step away from SIAI mainline!
Is accepting multi-universes important to the SIAI argument? There are a very, very large number of smart people who know very little about physics. They give lip service to quantum theory and relativity because of authority - but they do not understand them. Mentioning multi-universes just slams a door in their minds. If it is important then you will have to continue referring to it but if it is not then it would be better not to sound like you have science fiction type ideas.
Definitely not, for the purposes of public relations at least. It may make some difference when actually doing AI work.
Good point. Cryonics probably comes with a worse Sci. Fi. vibe but is unfortunately less avoidable.
This is a large part of what I implicitly had in mind making my cryonics post (which I guess really rubbed you the wrong way). You might be interested in taking a look at the updated version if you haven't already done so - I hope it's more clear than it was before.
Things that stretch my credibility.
Perhaps that was a marketing effort.
After all, everyone likes to tell the tale of the forbidden topic and the apprentice being insulted. You are spreading the story around now - increasing the mystery and intrigue of these mythical events about which (almost!) all records have been deleted. The material was left in public for a long time - creating plenty of opportunities for it to "accidentally" leak out.
By allowing partly obfuscated forbidden materials to emerge, you may be contributing to the community folklaw, spreading and perpetuating the intrigue.
Please stop doing this. You are adding spaced repetition to something that I, and others, positively do not want to think about. That is a real harm and you do not appear to have taken it seriously.
I'm sorry, but people like Wei force me to do this as they make this whole movement look like being completely down-to-earth, when in fact most people, if they knew about the full complexity of beliefs within this community, would laugh out loud.
The "laugh test" is not rational. I think that, if the majority of people fully understood the context of such statements, they would not consider them funny.
The context asked 'what kind of things a typical smart person would find uncredible'. This is a perfect example of such a thing.
A typical smart person would find the laugh test credible? We must have different definitions of "smart."
(Voted you back up to 0 here.)
I think you are right about the laugh test itself.
The topic was the banned topic and the deleted posts - not the laugh test. If you explained what happened to an outsider - they would have a hard time believing the story - since the explanation sounds so totally crazy and ridiculous.
You have a good point. It would be completely unreasonable to ban topics in such a manner while simultaneously expecting to maintain an image of being down to earth or particularly credible to intelligent external observers. It also doesn't reflect well on the SIAI if their authorities claim they cannot consider relevant risks because due to psychological or psychiatric difficulties. That is incredibly bad PR. It is exactly the kind of problem this post discusses.
I still have a hard time believing it actually happened. I have heard that there's no such thing as bad publicity - but surely nobody would pull this kind of stunt deliberately. It just seems to be such an obviously bad thing to do.
You don't seem to realize that claims like the ones in the post in question are a common sort of claim to make people vulnerable to neuroses develop further problems. Regardless whether or not the claims are at all reasonable, repeatedly referencing them this way is likely to cause further psychological harm. Please stop.
JoshuaZ:
However, it seems that in general, the mere fact that certain statements may cause psychological harm to some people is not considered a sufficient ground for banning or even just discouraging such statements here. For example, I am sure that many religious people would find certain views often expressed here shocking and deeply disturbing, and I have no doubt that many of them could be driven into serious psychological crises by exposure to such arguments, especially if they're stated so clearly and poignantly that they're difficult to brush off or rationalize away. Or, to take another example, it's very hard to scare me with hypotheticals, but the post "The Strangest Thing An AI Could Tell You" and the subsequent thread came pretty close; I'm sure that at least a few readers of this blog didn't sleep well if they happened to read that right before bedtime.
So, what exact sorts of potential psychological harm constitute sufficient grounds for proclaiming a topic undesirable? Is there some official policy about this that I've failed to acquaint myself with?
That's a very valid set of points and I don't have a satisfactory response.
The form of blanking out you use isn't secure. Better to use pure black rectangles.
That document was knocking around on a public website for several days.
Using very much security would probably be pretty pointless.
Pure black rectangles are not necessarily secure, either.
Amusing anecdote: There was a story about this issue on Slashdot one time, where someone possessing kiddy porn had obscured the faces by doing a swirl distortion, but investigators were able to sufficiently reverse this by doing an opposite swirl and so were able to identify the victims.
Then someone posted a comment to say that if you ever want to avoid this problem, you need to do something like a Gaussian blur, which deletes the information contained in that portion of the image.
Somebody replied to that comment and said, "Yeah. Or, you know, you could just not molest children."
Brilliant.
Nice link. (It's always good to read articles where 'NLP' doesn't refer, approximately, to Jedi mind tricks.)
I like your post. I wouldn't go quite so far as to ascribe outright negative utility to SIAI donations - I believe you underestimate just how much potential social influence money provides. I suspect my conclusion there would approximately mirror Vassar's.
(Typo: I think you meant to include 'traits' or similar in there.)
While Eliezer occasionally takes actions that seem clearly detrimental to his cause I do suggest that Eliezer is at least in principle aware of the dynamics you discuss. His alter ego "Harry Potter" has had similar discussion with his Draco in his fanfiction.
Also note that appearing too sophisticated would be extremely dangerous. If Eliezer or SIAI gains the sort of status and credibility you would like them to seek they open themselves to threats from governments and paramilitary organisations. If you are trying to take over the world it is far better to be seen as an idealistic do gooder who writes fanfic than as a political power player. You don't want the <TLA of choice> to raid your basement, kill you and run your near complete FAI with the values of the TLA. Obviously there is some sort of balance to be reached here...
They took down the "SIAI will not enter any partnership that compromises our values" "commitment" from their web site. Maybe they are more up for partnerships these days.
Thanks for pointing this out :-)
Interesting, I'll check this out at some point.
I raised a similar point on the IEET existential risk mailing list in a reply to James J. Hughes:
— James J. Hughes (existential.ieet.org mailing list, 2010-07-11)
I replied:
Seeing you quote James Hughes makes me wonder if I didn't realize where you were getting your ideas when I said the anti-Summit should be technical minded and avoid IEET-style politics.
This is a perfect example of where the 'outside view' can go wrong. Even the most basic 'inside view' of the topic would make it overwhelmingly obvious why a "75% certain of death by AI" folks could be allied (or the same people!) as the "solve all problems through AI" group. Splitting the two positions prematurely and trying to make a simple model of political adversity like that is just naive.
I personally guess >= 75% for AI death and also advocate FAI research. Preventing AI development indefinitely via desperate politico-military struggle would just not work in the long term. Trying would be utter folly. Nevermind the even longer term which would probably result in undesirable outcomes even if humanity did manage to artificially stunt its own progress in such a manner.
(The guy also uses 'schizophrenic' incorrectly.)
Greater folly than letting happen something with greater than a 75% chance of destroying the human race?
If you estimate a high chance of this action destroying humanity then trying to get through that bottleneck with slightly better than 75% chance of surviving is almost certainly better than trying to stamp out such research and buying a few years in exchange for replacing 75% with a near certainty. The only argument against that I can see if one accepts the 75% number is that forced delaying until we have uploads might help matters since uploads would have moral systems close to those of their original humans, and uploads can will have a better chance at solving the FAI problem or if not solving it being able to counteract any unFriendly or unfriendly AI.
AI research is hard. It's not clear to me that a serious, global ban on AI would only delay the arrival of AGI by a few years. Communication, collaboration, recruitment, funding... all of these would be much more difficulty. Even more, since current AI researchers are open about their work they would be the easiest to track after a ban, thus any new AI research would have to come from green researchers.
That aside, I agree that a ban whose goal is simply indefinite postponement of AGI is unlikely to work (and I'm dubious of any ban in general). Still, it isn't hard for me to imagine that a ban could buy us 10 years, and that a similar amount of political might could also greatly accelerate an upload project.
The biggest argument against, in my opinion, is that the only way the political will could be formed is if the threat of AGI was already so imminent that a ban really would be worse than worthless.
A ban seems highly implausible to me. What is the case for considering it? Do you really think that enough people will become convinced that there is a significant danger?
I agree, it seems highly implausible to me as well. However, the subject at hand (AI, AGI, FAI, uploads, etc) is riddled with extremes, so I'm hesitant to throw out any possibility simply because it would be incredibly difficult.
See the last line of the comment you responded to.
The other thing to consider is just what the ban would achieve. I would expect it to lower the 75% chance by giving us the opportunity to go extinct in another way before making a mistake with AI. When I say 'extinct' I include (d)evolving to an equilibrium (such as those described by Robin Hanson from time to time).
I don't think James Hughes would present or believe in that particular low-quality analysis himself either, if he didn't feel that SIAI is an organization competing with his IEET for popularity within the transhumanist subculture.
So mostly that statement is probably just about using "divide and conquer" towards transhumanists/singularitarians who are currently more popular within the transhumanist subculture than he is.
James Hughes seems like a fine fellow to me - and his SIAI disagreements seem fairly genuine. It is much of the rest of IEET that is the problem.
Can you tell us more about how you've seen people react to Yudkowsky? That these negative reactions are significant is crucial to your proposal, but I have rarely seen negative reactions to Yudkowsky (and never in person) so my first availability-heuristic-naive reaction is to think it isn't a problem. But I realize my experience may be atypical and there could be an abundance of avoidable Yudkowsky-hatred where I'm not looking, so would like to know more about that.
Did that objectionable Yudkowsky-meteorite comment get widely disseminated? YouTube says the video has only 500 views, and I imagine most of those are from Yudkowsky-sympathizing Less Wrong readers.
Yudkowsky-hatred isn't the risk, Yudkowsky-mild-contempt is. People engage with things they hate, sometimes it brings respect and attention to both parties (by polarizing a crowd that would otherwise be indifferent.) But you never want to be exposed to mild contempt.
I can think of some examples of conversations about Eliezer that would fit the category but it is hard to translate them to text. The important part of the reaction was non-verbal. Cryonics was one topic and the problem there wasn't that it was uncredible but that it was uncool. Another topic is the old "thinks he can know something about Friendly AIs when he hasn't even made an AI yet" theme. Again, I've seen that reaction evident through mannerisms that in no way translate to text. You can convey that people aren't socially relevant without anything so crude as saying stuff.
[insert the obvious bad pun here]
I know, I couldn't think of worthy witticism to lampshade it so I let it slide. :P
I haven't seen examples of Yudkowsky-hatred. But I have regularly seen people ridicule him. Recalling Hanson's view that a lot human behavior is really signaling and vying for status, I interpret this ridicule as an functioning to lower Eliezer's status to compensate for what people perceive as inappropriate status grubbing on his part.
Most of the smart people who I know (including myself) perceive him as exhibiting a high degree of overconfidence in the validity of his views about the world.
This leads some of them conceptualize him as a laughingstock; as somebody who's totally oblivious and feel that the idea that we should be thinking about artificial intelligence is equally worthy of ridicule. I personally am quite uncomfortable with these attitudes, agreeing with Holden Karnofsky's comment
"I believe that there are enormous risks and upsides associated with artificial intelligence. Managing these deserves serious discussion, and it’s a shame that many laugh off such discussion."
I'm somewhat surprised that you appear not to have notice this sort of thing independently. Maybe we hang out in rather different crowds.
Yes, I think that you're right. I just picked it out as a very concrete example of a statement that could provoke a substantial negative reaction. There are other qualitatively similar things (but more mild) things that Eliezer has said that have been more widely disseminated.
Existential risk reduction too! Charities are mostly used for signalling purposes - and to display affiliations and interests. Those caught up in causes use them for social networking with like-minded individuals - to signal how much they care, to signal how much spare time and energy they have - and so on. The actual cause is usually not irrelevant - but it is not particularly central either. It doesn't make much sense to expect individuals to be actually attempting to SAVE THE WORLD! This is much more likely to be a signalling phenomenon, making use of a superstimulus for viral purposes.
Ditto.
I know of a lot of very smart people (ok, less than 10, but still, more than 1) who essentially read Eliezer's AI writings as a form of entertainment, and don't take them even slightly seriously. This is partly because of the Absurdity Heuristic, but I think it's also because of Eliezer's writing style, and statements like the one in the initial post.
I personally fall somewhere between these people and, say, someone who has spent a summer at the SIAI on the 'taking Eliezer seriously' scale - I think he (and the others) probably have a point, and I at least know that they intend to be taken seriously, but I've never gotten round to doing anything about it.
Why do they find them entertaining?
I said the same before. It's mainly good science fiction. I'm trying to find out if there's more to it though.
Just saying this as evidence that there is a lot doubt even within the LW community.
As XiXiDu says - pretty much the same reason they find Isaac Asimov entertaining.
Oh, definitely. I have no real-life friends who are interested enough in these topics to know who Yudkowsky is (except, possibly, for what little they hear from me, and I try to keep the proselytizing to acceptable levels). So it's just me and the internet.
I have seen some ridicule of Yudkowsky (on the internet) but my impression had been that it wasn't a reaction to his tone, but rather that people were using the absurdity heuristic (cryonics and AGI are crazy talk) or reacting to surface-level status markers (Yudkowsky doesn't have a PhD). That is to say, it didn't seem the kind of ridicule that was avoidable by managing one's tone. I don't usually read ridicule in detail so it makes sense I'd be mistaken about that.
If it hasn't happened yet, that's at least some evidence it won't happen. Do you have reason to imagine a scenario which makes things very much worse than they already are based on such an effect, which means we must take care to tiptoe around these possibilities without allowing even one to happen? Because if not, we should probably worry about the things that already go wrong more than the things that might go wrong but haven't yet.
I'm confused - I feel like I already address most of your remarks in the comment that you're responding to?
Negative reactions to Yudkowsky from various people (academics concerned with x-risk), just within the past few weeks:
...
...
...
...
I was told that the quotes above state some ad hominem falsehoods regarding Eliezer. I think it is appropriate to edit the message to show that indeed some person might not be have been honest, or clueful. Otherwise I'll unnecessary end up perpetuating possible ad hominem attacks.
That N negative reactions about issue S exist only means that issue S is sufficiently popular.
Not if the polling is of folk in a position to have had contact with S, or is representative.
I don't like to, but if necessary I can provide the indentity of the people who stated the above. They all directly work to reduce x-risks. I won't do so in public however.
The last one was from David Pearce.
Identity of these people is not the issue. The percentage of people in given category that have negative reactions for given reason, negative reactions for other reason, and positive reactions would be useful, but not a bunch of filtered (in unknown way) soldier-arguments.
I know. I however just wanted to highlight that there are negative reactions, including not so negative critique. If you look further, you'll probably find more. I haven't saved all I saw over the years, I just wanted to show that it's not like nobody has a problem with EY. And in all ocassion I actually defended him by the way.
The context is also difficult to provide as some of it is from private e-Mails. Although the first one is from here and after thinking about it I can also provide the name since he was anyway telling this Michael Anissimov. It is from Sean Hays:
Sure, but XiXiDu's quotes bear no such framing.
Is it likely that someone who's doing interesting work that's publicly available wouldn't attract some hostility?
Just take the best of anybody and discard the rest. Yudkowsky has some very good points (about 80% of his writings, by my view) - take them and say thank you.
When he or the SIAI missed the point, to put it mildly, you know it better anyway, don't you?
I agree that Yudkowsky has some very good points.
My purpose in making the top level post is as stated: to work against poisoning the meme.
My point is, that you can "poise the meme" only for the fools. A wise individual can see for zerself what to pick and what not to pick.
But there are (somewhat) wise individuals who have not yet thought carefully about existential risk. They're forced to heuristically decide whether or not thinking about it more makes sense. Given what they know at present, it may be rational for them to dismiss Eliezer as being like a very smart version of the UFO conspiracy theorists or something like that. Because of the halo effect issue that Yvain talks about, this may lower their willingness to consider existential risk at all.
Most people do not systematically go through every statement that some particular person has made. If someone has heard primarily negative things about somebody else, then that reduces the chance of them even bothering to look at the person's other writings. This is quite rational behavior, since there are a lot of people out there and one's time is limited.
You must not go through the entire "memesphere" by the people, but by the memes! The current internet makes that quite easy to do.
You must mean "quite easy" in the sense that it would only require a few millions of man-hours, rather than a few quadrillions.
I don't find persuasive your arguments that the following policy suggestion has high impact (or indeed is something to worry about, in comparison with other factors):
(Note that both times I qualified the suggestion to contact SIAI (instead of starting a war) with "if you have a reasonable complaint/usable suggestion for improvement".)
Understood. Here too, we may have legitimate grounds for difference of opinion rooted in the fact that our impressions have emerged from thousands of data points which we don't have conscious access to.
Do you feel that I've misrepresented you? I'd be happy to qualify my reference to you however you'd like if you deem my doing so suitable.
I'm new to all this singularity stuff - and as an anecdotal data point, I'll say a lot of it does make my kook bells go off - but with an existential threat like uFAI, what does the awareness of the layperson count for? With global warming, even if most of any real solution involves the redesign of cities and development of more efficient energy sources, individuals can take some responsibility for their personal energy consumption or how they vote. uFAI is a problem to be solved by a clique of computer and cognitive scientists. Who needs to put thought into the possibility of misbuilding an AI other than people who will themselves engage in AI research? (This is not a rhetorical question - again, I'm new to this.)
There is, of course, the question of fundraising. ("This problem is too complicated for you to help with directly, but you can give us money..." sets off further alarm bells.) But from that perspective someone who thinks you're nuts is no worse than someone who hasn't heard of you. You can ramp up the variance of people's opinions and come out better financially.
Awareness on the part of government funding agencies (and the legislators and executive branch people with influence over them), technology companies and investors, and political and military decisionmakers (eventually) could all matter quite a lot. Not to mention bright young people deciding on their careers and research foci.
The people who do the real work. Utlimately it doesn't matter if the people who do the AI research care about existential risk or not (if we make some rather absolute economic assumptions). But you've noticed this already and you are right about the 'further alarm bells'.
Ultimately, the awareness of the layperson matters for the same reason that the awareness of the layperson matters for any other political issue. While with AI people can't get their idealistic warm fuzzies out of barely relevant things like 'turning off a light bulb' things like 'how they vote' do matter. Even if it is at a lower level of 'voting' along the lines of 'which institutions do you consider more prestigious'?
Good point!
But what if you're increasing existential risk, because encouraging SIAI staff to censor themselves will make them neurotic and therefore less effective thinkers? We must all withhold karma from multifoliaterose until this undermining stops! :-)
Have we seen any results (or even progress) come from the SIAI Challenge Grants, which included a Comprehensive Singularity FAQ and many academic papers dealing directly with the topics of concern? These should hopefully be less easy to ridicule and provide an authoritative foundation after the peer review process.
Edit: And if they fail to come to fruition, then we have some strong evidence to doubt SIAI's effectiveness.
I am one of those who haven't been convinced by the SIAI line. I have two main objections.
First, EY is concerned about risks due to technologies that have not yet been developed; as far as I know, there is no reliable way to make predictions about the likelihood of the development of new technologies. (This is also the basis of my skepticism about cryonics.) If you're going to say "Technology X is likely to be developed" then I'd like to see your prediction mechanism and whether it's worked in the past.
Second, shouldn't an organization worried about the dangers of AI be very closely in touch with AI researchers in computer science departments? Sure, there's room for pure philosophy and mathematics, but you'd need some grounding in actual AI to understand what future AIs are likely to do.
I think multifoliaterose is right that there's a PR problem, but it's not just a PR problem. It seems, unfortunately, to be a problem with having enough justification for claims, and a problem with connecting to the world of professional science. I think the PR problems arise from being too disconnected from the demands placed on other scientific or science policy organizations. People who study other risks, say epidemic disease, have to get peer-reviewed, they have to get government funding -- their ideas need to pass a round of rigorous criticism. Their PR is better by necessity.
As was mentioned in other threads, SIAI's main arguments rely on disjunctions and antipredictions more than conjunctions and predictions. That is, if several technology scenarios lead to the same broad outcome, that's a much stronger claim than one very detailed scenario.
For instance, the claim that AI presents a special category of existential risk is supported by such a disjunction. There are several technologies today which we know would be very dangerous with the right clever 'recipe'– we can make simple molecular nanotech machines, we can engineer custom viruses, we can hack into some very sensitive or essential computer systems, etc. What these all imply is that a much smarter agent with a lot of computing power is a severe existential threat if it chooses to be.
I agree completely. The reason why I framed my top level post in the way that I did was so that it would be relevant to readers of a variety of levels of confidence in SIAI's claims.
As I indicate here, I personally wouldn't be interested in funding SIAI as presently constituted even if there was no PR problem.
I think there are ways to make these predictions. On the most layman level I would point out that IBM build a robot that beats people at Jeopardy. Yes, I am aware that this is a complete machine-learning hack (this is what I could gather from the NYT coverage) and is not true cognition, but it surprised even me (I do know something about ML). I think this is useful to defeat the intuition of "machines cannot do that". If you are truly interested I think you can (I know you're capable) read Norvig's AI book, and than follow up on the parts of it that most resemble human cognition; I think serious progress is made in those areas. BTW, Norvig does take FAI issues seriously, including a reference to EY paper in the book.
I think they should, I have no idea if this is being done; but if I would do it I would not do it publicly, as it may have very counterproductive consequences. So until you or I become SIAI fellows we will not know, and I cannot hold such lack of knowledge against them.
First, I'm not really claiming "machines cannot do that." I can see advances in machine learning and I can imagine the next round of advances being pretty exciting. But I'm thinking in terms of maybe someday a machine being able to distinguish foreground from background, or understand a sentence in English, not being a superintelligence that controls Earth's destiny. The scales are completely different. One scale is reasonable; one strains credibility, I'm afraid.
Thanks for the book recommendation; I'll be sure to check it out.
I'm not sure what you refer to by "actual AI." There is a sub-field of academic computer science which calls itself "Artificial Intelligence," but it's not clear that this is anything more than a label, or that this field does anything more than use clever machine learning techniques to make computer programs accomplish things that once seemed to require intelligence (like playing chess, driving a car, etc.)
I'm not sure why it is a requirement that an organization concerned with the behavior of hypothetical future engineered minds would need to be in contact with these researchers.
Yes, the subfield of computer science is what I'm referring to.
I'm not sure that the difference between "clever machine learning techniques" and "minds" is as hard and fast as you make it. A machine that drives a car is doing one of the things a human mind does; it may, in some cases, do it through a process that's structurally similar to the way the human mind does it. It seems to me that machines that can do these simple cognitive tasks are the best source of evidence we have today about hypothetical future thinking machines.
I gave the wrong impression here. I actually think that machine learning might be a good framework for thinking about how parts of the brain work, and I am very interested in studying machine learning. But I am skeptical that more than a small minority of projects where machine learning techniques have been applied to solve some concrete problem have shed any light on how (human) intelligence works.
In other words, I largely agree with Ben Goertzel's assertion that there is a fundamental difference between "narrow AI" and AI research that might eventually lead to machines capable of cognition, but I'm not sure I have good evidence for this argument.
Although one should be very, very careful not to confuse the opinions of someone like Goertzel with those of the people (currently) at SIAI, I think it's fair to say that most of them (including, in particular, Eliezer) hold a view similar to this. And this is the location -- pretty much the only important one -- of my disagreement with those folks. (Or, rather, I should say my differing impression from those folks -- to make an important distinction brought to my attention by one of the folks in question, Anna Salamon.) Most of Eliezer's claims about the importance of FAI research seem obviously true to me (to the point where I marvel at the fuss that is regularly made about them), but the one that I have not quite been able to swallow is the notion that AGI is only decades away, as opposed to a century or two. And the reason is essentially disagreement on the above point.
At first glance this may seem puzzling, since, given how much more attention is given to narrow AI by researchers, you might think that someone who believes AGI is "fundamentally different" from narrow AI might be more pessimistic about the prospect of AGI coming soon than someone (like me) who is inclined to suspect that the difference is essentially quantitative. The explanation, however, is that (from what I can tell) the former belief leads Eliezer and others at SIAI to assign (relatively) large amounts of probability mass to the scenario of a small set of people having some "insight" which allows them to suddenly invent AGI in a basement. In other words, they tend to view AGI as something like an unsolved math problem, like those on the Clay Millennium list, whereas it seems to me like a daunting engineering task analogous to colonizing Mars (or maybe Pluto).
This -- much more than all the business about fragility of value and recursive self-improvement leading to hard takeoff, which frankly always struck me as pretty obvious, though maybe there is hindsight involved here -- is the area of Eliezer's belief map that, in my opinion, could really use more public, explicit justification.
Note that allowing for a possibility of sudden breakthrough is also an antiprediction, not a claim for a particular way things are. You can't know that no such thing is possible, without having understanding of the solution already at hand, hence you must accept the risk. It's also possible that it'll take a long time.
I don't think this is a good analogy. The problem of colonizing Mars is concrete. You can make a TODO list; you can carve the larger problem up into subproblems like rockets, fuel supply, life support, and so on. Nobody knows how to do that for AI.
OK, but it could still end up being like colonizing Mars if at some point someone realizes how to do that. Maybe komponisto thinks that someone will probably carve AGI in to subproblems before it is solved.
Well, it seems we disagree. Honestly, I see the problem of AGI as the fairly concrete one of assembling an appropriate collection of thousands-to-millions of "narrow AI" subcomponents.
Perhaps another way to put it would be that I suspect the Kolmogorov complexity of any AGI is so high that it's unlikely that the source code could be stored in a small number of human brains (at least the way the latter currently work).
EDIT: When I say "I suspect" here, of course I mean "my impression is". I don't mean to imply that I don't think this thought has occurred to the people at SIAI (though it might be nice if they could explain why they disagree).
I'm not a member of SIAI but my reason for thinking that AGI is not just going to be like lots of narrow bits of AI stuck together is that I can see interesting systems that haven't been fully explored (due to difficulty of exploration). These types of systems might solve some of the open problems not addressed by narrow AI.
These are problems such as
Now I also doubt that these systems will develop quickly when people get around to investigating them. And they will have elements of traditional narrow AI in as well, but they will be changeable/adaptable parts of the system, not fixed sub-components. What I think needs is exploring is primarily changes in software life-cycles rather than a change in the nature of the software itself.
The portion of the genome coding for brain architecture is a lot smaller than Windows 7, bit-wise.
Thanks, this is useful to know. Will revise beliefs accordingly.
I don't think AGI in a few decades is very farfetched at all. There's a heckuvalot of neuroscience being done right now (the Society for Neuroscience has 40,000 members), and while it's probably true that much of that research is concerned most directly with mere biological "implementation details" and not with "underlying algorithms" of intelligence, it is difficult for me to imagine that there will still be no significant insights into the AGI problem after 3 or 4 more decades of this amount of neuroscience research.
Of course there will be significant insights into the AGI problem over the coming decades -- probably many of them. My point was that I don't see AGI as hard because of a lack of insights; I see it as hard because it will require vast amounts of "ordinary" intellectual labor.
I'm having trouble understanding how exactly you think the AGI problem is different from any really hard math problem. Take P != NP, for instance the attempted proof that's been making the rounds on various blogs. If you've skimmed any of the discussion you can see that even this attempted proof piggybacks on "vast amounts of 'ordinary' intellectual labor," largely consisting of mapping out various complexity classes and their properties and relations. There's probably been at least 30 years of complexity theory research required to make that proof attempt even possible.
I think you might be able to argue that even if we had an excellent theoretical model of an AGI, that the engineering effort required to actually implement it might be substantial and require several decades of work (e.g. Von Neumann architecture isn't suitable for AGI implementation, so a great deal of computer engineering has to be done).
If this is your position, I think you might have a point, but I still don't see how the effort is going to take 1 or 2 centuries. A century is a loooong time. A century ago humans barely had powered flight.
By no means do I want to downplay the difficulty of P vs NP; all the same, I think we have different meanings of "vast" in mind.
The way I think about it is: think of all the intermediate levels of technological development that exist between what we have now and outright Singularity. I would only be half-joking if I said that we ought to have flying cars before we have AGI. There are of course more important examples of technologies that seem easier than AGI, but which themselves seem decades away. Repair of spinal cord injuries; artificial vision; useful quantum computers (or an understanding of their impossibility); cures for the numerous cancers; revival of cryonics patients; weather control. (Some of these, such as vision, are arguably sub-problems of AGI: problems that would have to be solved in the course of solving AGI.)
Actually, think of math problems if you like. Surely there are conjectures in existence now -- probably some of them already famous -- that will take mathematicians more than a century from now to prove (assuming no Singularity or intelligence enhancement before then). Is AGI significantly easier than the hardest math problems around now? This isn't my impression -- indeed, it looks to me more analogous to problems that are considered "hopeless", like the "problem" of classifying all groups, say.
...but you don't really know - right?
You can't say with much confidence that there's no AIXI-shaped magic bullet.
That's right; I'm not an expert in AI. Hence I am describing my impressions, not my fully Aumannized Bayesian beliefs.
One obvious piece of evidence is that many forms of narrow learning are mathematically incapable of doing much. There are for example a whole host of theorems about what different classes of neural networks can actually recognize, and the results aren't very impressive. Similarly, support vector machine's have a lot of trouble learning anything that isn't a very simple statistical model, and even then humans need to decide which stats are relevant. Other linear classifiers run into similar problems.
Prediction is hard, especially about the future.
One thing that intrigues me is snags. Did anyone predict how hard to would be to improve batteries, especially batteries big enough for cars?
Let's keep in mind that your estimated probabilities of various technological advancements occurring and your level of confidence in those estimates are completely distinct... In particular, here you seem to express low estimated probabilities of various advancements occurring, and you justify this by saying "we really have no idea". This seems like a complete non sequitur. Maybe you have a correct argument in your mind, but you're not giving us all the pieces.
If you haven't demonstrated 1 -- if it's still unknown -- you can't expect me to believe 3. The burden of proof is on whoever's asking for money for a new risk-mitigating venture, to give strong evidence that the risk is real.
I see "burden of proof" as a misconcept in the same way that someone "deserving" something is. A better way of thinking about this: "You seem to be making a strong claim. Mind sharing the evidence for your claim for me? ...I disagree that the evidence you present justifies your claim."
For what it's worth, I also see "must _" as a misconcept--although "must _ to _" is not. It's an understandable usage if the "to _*" clause is implicit, but that doesn't seem true in this case. So to fix up SIAI's argument, you could say that these are the statements whose probabilities are being contested:
And depending on their probabilities, the following may or may not be true:
Pretty much anything you say that's not relevant to one of statements 1 or 2 (including statements that certain people haven't been "responsible" enough in supporting their claims) is completely irrelevant to the question of whether you want to take action Y. You already have (or ought to be able to construct) probability estimates for each of 1 and 2.
Your grasp of decision theory is rather weak if you are suggesting that when Technology X is developed is irrelevant to SarahC's decision. Similarly, you seem to suggest that the ratio of value to cost is irrelevant and that all that matters is which is bigger. Wrong again.
But your real point was not to set up a correct decision problem, but rather to suggest that her questions about whether "certain people" have been "responsible" are irrelevant. Well, I have to disagree. If action Y is giving money to "certain people", then their level of "responsibility" is very relevant.
I did enjoy your observations regarding "burden of proof" and "must", though probably not as much as you did.
Of course that is important. I didn't want to include a lot of qualifiers.
I'm not trying to make a bulletproof argument so much as concisely give you an idea of why I think SarahC's argument is malformed. My thinking is that should be enough for intellectually honest readers, as I don't have important insights to offer beyond the concise summary. If you think I ought to write longer posts with more qualifications for readers who aren't good at taking ideas seriously feel free to say that.
Really? So in some circumstances it is rational to take an action for which the expected cost is greater than the expected value? Or it is irrational to take an action for which the expected value exceeds the expected cost? (I'm using "rational" to mean "expected utility maximizing", "cost" to refer to negative utility, and "value" to refer to positive utility--hopefully at this point my thought process is transparent.)
It would be a well-formed argument to say that because SIAI folks make strong claims without justifying them, they won't use money SarahC donates well. As far as I can tell, SarahC has not explicitly made that argument. (Recall I said that she might have a correct argument in her mind but she isn't giving us all the pieces.)
Please no insults, this isn't you versus me is it?
No, your error was in the other direction. If you look back carefully, you will notice that the ratio is being calculated conditionally on Technology X being developed. Given that the cost is sunk regardless of whether the technology appears, it is possible that SarahC should not act even though the (conditionally) expected return exceeds the cost.
Shouldn't be. Nor you against her. I was catty only because I imagined that you were being catty. If you were not, then I surely apologize.
I edited my post before I saw your response :-P
I'm sorry, I don't see any edits that matter for the logic of the thread. What am I missing?
OK, my mistake.
I didn't say what SarahC should do with the probabilities once she had them. All I said was that they were pretty much all was relevant to the question of whether she should donate. Unless I didn't, in which case I meant to.
So you think a danger needs to likely arrive in a few decades for it to merit attention?
I think that is quite irresponsible. No law of physics states that all problems can certainly be solved very well in a few decades (the solutions for some problems might even necessarily involve political components, btw), so starting preparations earlier can be necessary.
There needs to be an article on this point. In the absence of a really good way of deciding what technologies are likely to be developed, you are still making a decision. You haven't signed up yet; whether you like it or not, that is a decision. And it's a decision that only makes sense if you think technology X is unlikely to be developed, so I'd like to see your prediction mechanism and whether it's worked in the past. In the absence of really good information, we sometimes have to decide on the information we have.
EDIT: I was thinking about cryonics when I wrote this, though the argument generalizes.
What is it that is making you think that whatever SarahC hasn't "signed up" to is having a positive effect - and that she can't do something better with her resources?
whpearson mentioned this already, but if you think that the most important thing we can be doing right now is publicizing an academically respectable account of existential risk, then you should be funding the Future of Humanity Institute.
Funding SIAI is optimal only if you think that the pursuit of Friendly AI is by far the most important component of existential risk reduction, and indeed they're focusing on persuading more people of this particular claim. As you say, by focusing on something specific, radical and absurd, they run more of a risk of being dismissed entirely than does FHI, but their strategy is still correct given the premise.
But who does the evaluation? It seems that it's better to let specialists think about whether a given cause is important, and they need funding just to get that running. This argues for ensuring minimum funding of organizations that research important uncertainties, even the ones where your intuitive judgment says probably lead nowhere. Just as most people shouldn't themselves research FAI, and instead fund its research, similarly most people shouldn't research feasibility of research of FAI, and instead fund the research of that feasibility.
I think you claim too much. If I decided I couldn't follow the relevant arguments, and wanted to trust a group to research the important uncertainties of existential risk, I'd trust FHI. (They could always decide to fund or partner with SIAI themselves if its optimality became clear.)
This seems to work as an argument for much greater marginal worth of ensuring minimal funding, so that there is at least someone who researches the uncertainties professionally (to improve on what people from the street can estimate in their spare time), before we ask about the value of researching the stuff these uncertainties are about. (Of course, being the same organization that directly benefits from a given answer is generally a bad idea, so FHI might work in this case.)
The FHI's pitch: http://www.fhi.ox.ac.uk/get_involved
My only worry about funding FHI exclusively is that they are primarily philosophical and academic. I'd worry that the default thing they would do with more money would be to produce more philosophical papers. Rather than say doing/funding biological research or programming, if that was what was needed.
But as the incentive structures for x-risk reduction organisations go, those of an academic philosophy department aren't too bad at this stage.
This seems to assume that existential risk reduction is the only thing people care about. I doubt I am the only person who wants more from the universe than eliminating risk of humans going extinct. I would trade increased chance of extinction for a commensurate change in the probable outcomes if we survive. Frankly I would consider it insane not to be willing make such a trade.
It seems pretty clear that very few care much about existential risk reduction.
That makes perfect sense from an evolutionary perspective. Organisms can be expected to concentrate on producing offspring - not indulging paranoid fantasies about their whole species being wiped out!
The bigger puzzle is why anyone seems to care about it at all. The most obvious answer is signalling. For example, if you care for the fate of everyone in the whole world, that SHOWS YOU CARE - a lot! Also, the END OF THE WORLD acts as a superstimulus to people's warning systems. So - they rush and warn their friends - and that gives them warm fuzzy feelings. The get credit for raising the alarm about the TERRIBLE DANGER - and so on.
Disaster movies - like 2012 - trade on people's fears in this area - stimulating and fuelling their paranoia further - by providing them with fake memories of it happening. One can't help wondering whether FEAR OF THE END is a healthy phenomenon - overall - and if not, whether it realy sensible to stimulate those fears.
Does the average human - on being convinced the world is about to end - behave better - or worse? Do they try and hold back the end - or do they rape and pillage? If their behaviour is likely to be worse then responsible adults should think very carefully before promoting the idea that THE END IS NIGH on the basis of sketchy evidence.
I meant "optimal within the category of X-risk reduction", and I see your point.
Upvoted.
We've had agreements and disagreements here. This is one of the agreements.
I disagree. If we can avoid being wiped out, or otherwise have our potential permanently limited, our eventual outcome is very likely to be good beyond our potential to imagine. I really think the "maxipok" term of our efforts toward the greater good can't fail to absolutely dominate all other terms.
Lack of interest in existential risk reduction makes perfect sense from an evolutionary perspective. As I have previously explained:
"Organisms can be expected to concentrate on producing offspring - not indulging paranoid fantasies about their whole species being wiped out!"
Most people are far more concerned about other things - for perfectly sensible and comprehensible reasons.
That sounds very optimistic. I just don't see any reason for us to expect the future should be so bright if human genetic, cultural and technological go on under the usual influence of competition. Unless we do something rather drastic (eg. FAI or some other kind of positive singleton) in the short term then it seems inevitable that we end up in Malthusian hell.
Most of what I consider 'good' is, for the purposes of competition, a complete waste of time.
Solid, bold post.
Eliezer's comments on his personal importance to humanity remind me of the Total Perspective Device from Hitchhiker's. Everyone who gets perspective from the TPD goes mad; Zaphod Beeblebrox goes in and finds out he's the most important person in human history.
Eliezer's saying he's Zaphod Beeblebrox. Maybe he is, but I'm betting heavily against that for the reasons outlined in the post. I expect AI progress of all sorts to come from people who are able to dedicate long, high-productivity hours to the cause, and who don't believe that they and only they can accomplish the task.
I also don't care if the statements are social naivete or not; I think the statements that indicate that he is the most important person in human history - and that seems to me to be what he's saying - are so seriously mistaken, and made with such a high confidence level, as to massively reduce my estimated likelihood that SIAI is going to be productive at all.
And that's a good thing. Throwing money into a seriously suboptimal project is a bad idea. SIAI may be good at getting out the word of existential risk (and I do think existential risk is serious, under-discussed business), but the indicators are that it's not going to solve it. I won't give to SIAI if Eliezer stops saying these things, because it appears he'll still be thinking those things.
I expect AI progress to come incrementally, BTW - I don't expect the Foomination. And I expect it to come from Google or someone similar; a large group of really smart, really hard-working people.
I could be wrong.
--JRM
I'd like to point out that it's not either/or: it's possible (likely?) that it will take decades of hard work and incremental progress by lots of really smart people to advance AI science to a point where an AI could FOOM.
I would say likely, conditional on eventual FOOM. The alternative means both a concentration of probability mass in the next ten years and that the relevant theory and tools are almost wholly complete.
I disagree strongly with this post. In general, it is a bad idea to refrain from making claims that one believes are true simply because those claims will make people less likely to listen to other claims. That direction lies the downwards spiral of emotional manipulation, rhetoric, and other things not conducive to rational discourse.
Would one under this logic encourage the SIAI to make statements that are commonly accepted but wrong in order to make people more likely to listen to the SIAI? If not, what is the difference?
It seems as though the latter strategy could backfire - if the false statements were exposed. Keeping your mouth shut about controversial issues seems safer.
I believe that there are contexts in which the right thing to do is to speak what one believes to be true even if doing so damages public relations.
These things need to be decided on a case-by-case basis. There's no royal road to instrumental rationality.
As I say here, in the present context, a very relevant issue in my mind is that Eliezer & co. have not substantiated their most controversial claims with detailed evidence.
It's clichéd to say so, but extraordinary claims require extraordinary evidence. A claim of the type "I'm the most important person alive" is statistically many orders of magnitude more likely to be made by a poser than by somebody for whom the claim is true. Casual observers are rational to believe that Eliezer is a poser. The halo effect problem is irrational, yes, but human irrationality must be acknowledged, it's not the sort of thing that goes away if you pretend that it's not there.
I don't believe that Eliezer's outlandish and unjustified claims contribute to rational discourse. I believe that Eliezer's outlandish and unjustified claims lower the sanity waterline.
To summarize, I believe that in this particular case the costs that you allude to are outweighed by the benefits.
Come on - he never actually claimed that.
Besides, many people have inflated views of their own importance. Humans are built that way. For one thing, It helps them get hired, if they claim that they can do the job. It is sometimes funny - but surely not a big deal.
To the extent that people really want what you argue against, perhaps they should pursue an alternate organization than SIAI that promotes only the more palatable subset. I agree with you that somebody should be making all the claims, popular or not, that bear on x-risk.
Seems like a reasonable position to me.
An important part of existential risk reduction is making sure that people who are likely to work on AI, or fund it, have read the sequences, and are at least aware of how most possible minds are not minds we would want, and of how dangerous recursive self-improvement could be.
Really? I don't understand this position at all. The vast majority of the planet isn't very rational and the people with lots of resources are often not rational. If one can get some of those people to direct their resources in the right directions then that's still a net win for preventing existential risk even if they aren't very rational. If say a hundred million dollars more gets directed to existential risk even if much of that goes to the less likely existential risks that's still an overall reduction in existential risk and a general increase to the sanity waterline.
I find it impossible to believe that the author of Harry Potter and the Methods of Rationality is oblivious to the first impression he creates. However, I can well believe that he imagines it to be a minor handicap which will fade in importance with continued exposure to his brilliance (as was the fictional case with HP). The unacknowledged problem in the non-fictional case, of course, is in maintaining that continued exposure.
I am personally currently skeptical that the singularity represents existential risk. But having watched Eliezer completely confuse and irritate Robert Wright, and having read half of the "debate" with Hanson, I am quite willing to hypothesize that the explanation of what the singularity is (and why we should be nervous about it) ought to come from anybody but Eliezer. He speaks and writes clearly on many subjects, but not that one.
Perhaps he would communicate more successfully on this topic if he tried a dialog format. But it would have to be one in which his constructed interlocutors are convincing opponents, rather than straw men.
It depends on exactly what you mean by "existential risk". Development will likely - IMO - create genetic and phenotypic takeovers in due course - as the bioverse becomes engineered. That will mean no more "wild" humans.
That is something which some people seem to wail and wave their hands about - talking about the end of the human race.
The end of earth-originating civilisation seems highly unlikely to me too - which is not to say that the small chance of it is not significant enough to discuss.
Eliezer's main case for that appears to be on http://lesswrong.com/lw/y3/value_is_fragile/
I think that document is incoherent.
Damnit! My smug self assurance that I could postpone thinking about these issues seriously because I'm an SIAI donor .... GONE! How am I supposed to get any work done now?
Seriously though, I do wish the SIAI toned down its self importance and incredible claims, however true they are. I realize, of course, that dulling some claims to appear more credible is approaching a Dark Side type strategy, but... well, no buts. I'm just confused.
Edit: I misunderstood what Jordan was trying to say - the previous version of this comment is irrelevant to the present discussion and so I've deleted it.
No one is claiming that honesty deserves top priority. I would lie to save someone's life, or to make a few million dollars, etc. In the context of SIAI though, or any organization, being manipulative can severely discredit you.
If he were to go back on his incredible claims, or even only make more credible claims in the future, how would he reconcile the two when confronted? If someone new to Eliezer read his tame claims, then went back and read his older, more extreme claims, what would they think? To many people this would enforce the idea that SIAI is a cult, and that they are refining their image to be more attractive.
All of that said, I do understand where you're coming from intuitively, and I'm not convinced that scaling back some of the SIAI claims would ever have a negative effect. Certainly, though, a public policy conversation about it would cast a pretty manipulative shade over SIAI. Hell, even this conversation could cast a nasty shade to some onlookers (to many people trying to judge SIAI, the two of us might be a sufficiently close proxy, even though we have no direct connections).
Okay, I misunderstood where you were coming from earlier, I thought you were making a general statement about the importance of stating one's beliefs. Sorry about that.
In response to your present comments, I would say that though the phenomenon that you have in mind may be a PR issue, I think it would be less of a PR issue than what's going on right now.
One thing that I would say is that I think that Eliezer would come across as much more credible simply by accompanying his weird sounding statements with disclaimers of the type "I know that what I'm saying probably sounds pretty 'out there' and understand if you don't believe me, but I've thought about this hard, and I think..." See my remark here.
I mostly agree, although I'm still mulling it and think the issue is more complicated than it appears. One nitpick:
Personally these kind of qualifiers rarely do anything to allay my doubt, and can easily increase them. I prefer to see incredulity. For instance, when a scientist has an amazing result, rather than seeing that they fully believe it but recognizing it's difficult for me to believe, I'd rather see them doubtful of their own conclusion but standing by it nonetheless because of the strength of the evidence.
"I know it's hard to believe, but it's likely an AI will kill us all in the future."
could become
"It's hard for me to come to terms with, but there doesn't seem to be any natural safeguards preventing an AI from doing serious damage."
Sure, I totally agree with this - I prefer your formulation to my own. My point was just that there ought to be some disclaimer - the one that I suggested is a weak example.
Edit: Well, okay, actually I prefer:
"It took me a long time to come to terms with, but there don't seem to be any natural safeguards preventing an AI from doing serious damage."
If one has actually become convinced of a position, it sounds disingenuous to say that it's hard for one to come to terms with at present, but any apparently absurd position should at some point have been hard to come to terms with.
Adding such a qualifier is a good caution against appearing to be placing oneself above the listener. It carries the message "I know how you must be feeling about these things, I've been there too."
I am a relative newbie commenter here, and my interest in this site has so far been limited to using it as a fun forum where it's possible to discuss all kinds of sundry topics with exceptionally smart people. However, I have read a large part of the background sequences, and I'm familiar with the main issues of concern here, so even though it might sound impertinent coming from someone without any status in this community, I can't resist commenting on this article.
To put it bluntly, I think the main point of the article is, if anything, an understatement. Let me speak from personal experience. From the perspective of this community, I am a sort of person who should be exceptionally easy to get interested and won over to its cause, considering both my intellectual background and my extreme openness to contrarian viewpoints and skepticism towards the official academic respectability as a criteron of truth and intellectual soundness. Yet, to be honest, even though I find a lot of the writing and discussion here extremely interesting, and the writings of Yudkowsky (in addition to others such as Bostrom, Hanson, etc.) have convinced me that technology-related existential risks should be taken much more seriously than they presently are, I still keep encountering things in this community that set off various red flags, which are undoubtedly taken by many people as a sign of weirdness and crackpottery, and thus alienate huge numbers of potential quality audience.
Probably the worst such example I've seen was the recent disturbance in which Roko was subjected to abuse that made him leave. When I read the subsequent discussions, it surprised me that virtually nobody here appears to be aware what an extreme PR disaster it was. Honestly, for someone unfamiliar with this website who has read about that episode, it would be irrational not to conclude that there's some loony cult thing going on here, unless he's also presented with enormous amounts of evidence to the contrary in the form of a selection of the best stuff that this site has to offer. After these events, I myself wondered whether I want to be associated with an outlet where such things happen, even just as an occasional commenter. (And not to even mention that Roko's departure is an enormous PR loss in its own right, in that he was one of the few people here who know how to write in a way that's interesting and appealing to people who aren't hard-core insiders.)
Even besides this major PR fail, I see many statements and arguments here that may be true, or at least not outright unreasonable, but should definitely be worded more cautiously and diplomatically if they're given openly for the whole world to see. I'm not going to get into details of concrete examples -- in particular, I do not concur unconditionally with any of the specific complaints from the above article -- but I really can't help but conclude that lots of people here, including some of the most prominent individuals, seem oblivious as to how broader audiences, even all kinds of very smart, knowledgeable, and open-minded people, will perceive what they write and say. If you want to have a closed inner circle where specific background knowledge and attitudes can be presumed, that's fine -- but if you set up a large website attracting lots of visitors and participants to propagate your ideas, you have to follow sound PR principles, or otherwise its effect may well end up being counter-productive.
Agreed.
One good sign here is that LW, unlike most other non-mainstream organizations, doesn't really function like a cult. Once one person starts being critical, critics start coming out of the woodwork. I have my doubts about this place sometimes too, but it has a high density of knowledgeable and open-minded people, and I think it has a better chance than anyone of actually acknowledging and benefiting from criticism.
I've tended to overlook the weirder stuff around here, like the Roko feud -- it got filed under "That's confusing and doesn't make sense" rather than "That's an outrage." But maybe it would be more constructive to change that attitude.
Looking at my own posts I see a lot of this problem; that is, the problem of addressing only far too small an audience. Thank you for pointing it out.
What are the scenarios where someone unfamiliar with this website would hear about Roko's deleted post?
I suppose it could be written about dramatically (because it was dramatic!) but I don't think anyone is going to publish such an account. It was bad from the perspective of most LWers -- a heuristic against censorship is a good heuristic.
This whole thing is ultimately a meta discussion about moderation policy. Why should this discussion about banned topics be that much interesting than a post on Hacker News that is marked as dead? Hacker News generally doesn't allow discussion of why stories were marked dead. The moderators are anonymous and have unquestioned authority.
If Less Wrong had a mark as dead function (on HN unregistered users don't see dead stories, but registered users can opt-in to see them), I suspect Eliezer would have killed Roko's post instead of deleting it to avoid the concerns of censorship, but no one has written that LW feature yet.
As a solid example of what a not-PR disaster it was, I doubt that anyone at the Singularity Summit that isn't a regular Less Wrong reader (the majority of attendees) has heard that Eliezer deleted a post. It's just not the kind of thing that actually makes a PR disaster... honestly if this was a PR issue it might be a net positive because it would lead some people to hear of LW that otherwise would never have heard of Less Wrong. Please don't take that as a reason to make this a PR issue.
Eliezer succeeded in the sense that it is very unlikely that people in the future on Less Wrong are going to make stupid emotionally abhorrent posts about weird decision theory torture scenarios. He failed in that he could have handled the situation better.
If anyone would like to continue talking about Less Wrong moderation policy, the place to talk about it is the Meta Thread (though you'd probably want to make a new one (good for +[20,50] karma!) instead of discussing it in an out of season thread)
I agree completely. I still read LessWrong because I am a relatively long-time reader, and thus I know that most of the people here are sane. Otherwise, I would conclude that there is some cranky process going on here. Still, the Roko affair caused me to significantly lower my probabilities assigned to SIAI success and forced me to seriously consider the hypothesis that Eliezer Yudkowsky went crazy.
By the way, I have a little bit disturbing feeling that too little of the newer material here is actually devoted to refining the art of human rationality, as the blog's header proudly states, while instead the posts often discuss relatively narrow list of topics which are only tangentially related to rationality. E.g. cryonics, AI stuff, evolutionary psychology, Newcomb-like scenarios.
Part of that mission is to help people overcome the absurdity heuristic, and to help them think carefully about topics that normally trigger a knee-jerk reflex of dismissal on spurious grounds; it is in this sense that cryonics and the like are more than tangentially related to rationality.
I do agree with you that too much of the newer material keeps returning to those few habitual topics that are "superstimuli" for the heuristic. This perhaps prevents us from reaching out to newer people as effectively as we could. (Then again, as LW regulars we are biased in that we mostly look at what gets posted, when what may matter more for attracting and keeping new readers is what gets promoted.)
A site like YouAreNotSoSmart may be more effective in introducing these ideas to newcomers, to the extent that it mostly deals with run-of-the-mill topics. What makes LW valuable which YANSS lacks is constructive advice for becoming less wrong.
Thanks for the link, I haven't known YANSS.
As for overcoming absurdity heuristics, more helpful would be to illustrate its inaproppriateness (is this a real word?) on thoughts which are seemingly absurd while having a lot of data proving them right, rather than predictions like Singularity which are mostly based on ... just different heuristics.
Aside from the body of the article, which is just "common" sense, given the author's opinion against the current policies of SIAI, I found the final paragraph interesting because I also exhibit "an unusually high abundance of the traits associated with Aspergers Syndrome." Perhaps possessing that group of traits gives one a predilection to seriously consider existential risk reduction by being socially detached enough to see the bigger picture. Perhaps LW is somewhat homogenously populated with this "certain kind" of people. So, how do we gain credibility with normal people?
I'll state my own experience and perception, since it seems to be different from that of others, as evidenced in both the post and the comments. Take it for what it's worth; maybe it's rare enough to be disregarded.
The first time I heard about SIAI -- which was possibly the first time I had heard the word "singularity" in the technological sense -- was whenever I first looked at the "About" page on Overcoming Bias, sometime in late 2006 or early 2007, where it was listed as Eliezer Yudkowsky's employer. To make this story short, the whole reason I became interested in this topic in the first place was because I was impressed by EY -- specifically his writings on rationality on OB (now known as the Sequences here on LW). Now of course most of those ideas were hardly original with him (indeed many times I had the feeling he was stating the obvious, albeit in a refreshing, enjoyable way) but the fact that he was able to write them down in such a clear, systematic, and readable fashion showed that he understood them thoroughly. This was clearly somebody who knew how to think.
Now, when someone has made that kind of demonstration of rationality, I just don't have much problem listening to whatever they have to say, regardless of how "outlandish" it may seem in the context of most human discourse. Maybe I'm exceptional in this respect, but I've never been under the impression that only "normal-sounding" things can be true or important. At any rate, I've certainly never been under that impression to such an extent that I would be willing to dismiss claims made by the author of The Simple Truth and A Technical Explanation of a Technical Explanation, someone who understands things like the gene-centered view of evolution and why MWI exemplifies rather than violates Occam's Razor, in the context of his own professional vocation!
I really don't understand what the difference is between me and the "smart people" that you (and XiXiDu) know. In fact maybe they should be more inclined to listen to EY and SIAI; after all, they probably grew up reading science fiction, in households where mild existential risks like global warming were taken seriously. Are they just not as smart as me? Am I unusually susceptible to following leaders and joining cults? (Don't think so.) Do I simply have an unusual personality that makes me willing to listen to strange-sounding claims? (But why wouldn't they as well, if they're "smart"?)
Why can't they just read the darn sequences and pick up on the fact that these people are worth listening to?
Thanks for your thoughtful comment.
I know some people who have had this sort of experience. My claim is not that Eliezer has uniformly repelled people from thinking about existential risk. My claim is that on average Eliezer's outlandish claims repel people from thinking about existential risk.
My guess would be that this is it. I'm the same way.
It's not clear that willingness to listen to strange-sounding claims exhibits correlation with instrumental rationality, or what the sign of that correlation is. People who are willing to listen to strange-sounding claims statistically end up hanging out with UFO conspiracy theorists, New Age people, etc. more often than usual. Statistically, people who make strange-sounding claims are not worth listening to. Too much willingness to listen to strange-sounding claims can easily result in one wasting large portions of one's life.
See my remarks above.
Also, keep in mind that reading the sequences requires nontrivial effort-- effort which even moderately skeptical people might be unwilling to expend. Hopefully Eliezer's upcoming rationality book will solve some of that problem, though. After all, even if it contains largely the same content, people are generally much more willing to read one book rather than hundreds of articles.
For my part, I keep wondering how long it's going to be before someone throws his "If you don't sign up your kids for cryonics then you are a lousy parent" remark at me, to which I will only be able to say that even he says stupid things sometimes.
(Yes, I'd encourage anyone to sign their kids up for cryonics; but not doing so is an extremely poor predictor of whether or not you treat your kids well in other ways, which is what the term should mean by any reasonable standard).
Yes, this is the sort of thing that I had in mind in making my cryonics post - as I said in the revised version of my post, I have a sense that a portion of the Less Wrong community has the attitude that cryonics is "moral" in some sort of comprehensive sense.
Thank you for your thoughtful reply; although, as will be evident, I'm not quite sure I actually got the point across.
I didn't realize at all that by "smart" you meant "instrumentally rational"; I was thinking rather more literally in terms of IQ. And I would indeed expect IQ to correlate positively with what you might call openness. More precisely, although I would expect openness to be only weak evidence of high IQ, I would expect high IQ to be more significant evidence of openness.
The point of my comment was that reading his writings reveals a huge difference between Eliezer and UFO conspiracy theorists, a difference that should be more than noticeable to anyone with an IQ high enough to be in graduate school in mathematics. Yes, of course, if all you know about a person is that they make strange claims, then you should by default assume they're a UFO/New Age type. But I submit that the fact that Eliezer has written things like these decisively entitles him to a pass on that particular inference, and anyone who doesn't grant it to him just isn't very discriminating.
My own experience is that the correlation is not very high. Most of the people who I've met who are as smart as me (e.g. in the sense of having high IQ) are not nearly as open as I am.
I did not intend to equate intelligence with instrumental rationality. The reason why I mentioned instrumental rationality is that ultimately what matters is to get people with high instrumental rationality (whether they're open minded or not) interested in existential risk.
My point is that people who are closed minded should not be barred from consideration as potentially useful existential risk researchers, that although people are being irrational to dismiss Eliezer as fast as they do, that doesn't mean that they're holistically irrational. My own experience has been that my openness has both benefits and drawbacks.
Math grad students can see a huge difference between Eliezer and UFO conspiracy theorists - they recognize that Eliezer's intellectually sophisticated. They're still biased to dismiss him out of hand. See bentram's comment
Edit: You might wonder where the bias to dismiss Eliezer comes from. I think it comes mostly from conformity, which is, sadly, very high even among very smart people.
You may be right about this; perhaps Eliezer should in fact work on his PR skills. At the same time, we shouldn't underestimate the difficulty of "recruiting" folks who are inclined to be conformists; unless there's a major change in the general sanity level of the population, x-risk talk is inevitably going to sound "weird".
This is a problem; no question about it.
I agree with this. It's all a matter of degree. Maybe at present one has to be in the top 1% of the population in nonconformity to be interested in existential risk and with better PR one could reduce the level of nonconformity required to the top 5% level.
(I don't know whether these numbers are right, but this is the sort of thing that I have in mind - I find it very likely that there are people who are nonconformist enough to potentially be interested in existential risk but too conformist to take it seriously unless the people who are involved seem highly credible.)
I would perhaps expand 'conformity' to include neighbouring social factors - in-group/outgroup, personal affiliation/alliances, territorialism, etc.
One more point - though I could immediately recognize that there's something important to some of what Eliezer says, the fact that he makes outlandish claims did make me take longer to get around to thinking seriously about existential risk. This is because of a factor that I mention in my post which I quote below.
I'm not proud that I'm so influenced, but I'm only human. I find it very plausible that there are others like me.
Warning: Shameless Self Promotion ahead
Perhaps part of the difficulty here is the attempt to spur a wide rationalist community on the same site frequented by rationalists with strong obscure positions on obscure topics.
Early in Lesswrong discussion of FAI was discouraged so that it didn't just become a site about FAI and the singularity, but a forum about human rationality more generally.
I can't track down an article[s] from EY about how thinking about AI can be too absorbing, and how in order properly create a community, you have to truly put aside the ulterior motive of advancing FAI research.
It might be wise for us to again deliberately shift our focus away from FAI and onto human rationality and how it can be applied more widely (say to science in general.)
Enter the SSP: For months now I've been brainstorming a community to educate people on the creation and use of 3D printers, with the eventual goal of making much better 3D printers. So this is a different big complicated problem with a potential high payoff, and it ties into many fields, provides tangible previews of the singularity, can benefit from the involvement of people with almost any skill set, and seems to be much safer than advancing AI, nanotech, or genetic engineering.
I had already intended to introduce rationality concepts where applicable and link a lot to Lesswrong. but if a few LWers were willing to help, It could become a standalone community of people committed to thinking clearly about complex technical and social problems, with a latent obsession with 3D printers.
Just to check... have I said any "naughty" things analogous to the Eliezer quote above?
Not to my knowledge... but Eliezer makes his words far more prominent that you do.
This post reminds me of the talk at this year's H+ summit by Robert Tercek. Amongst other things, he was pointing out how the PR battle over transhumanist issues was already lost in popular culture, and that the transhumanists were not helping matters by putting people with very freaky ideas in the spotlight.
I wonder if there are analogous concerns here.
When I'm talking to someone I respect (and want to admire me), I definitely feel an urge to distance myself from EY. I feel like I'm biting a social bullet in order to advocate for SIAI-like beliefs or action.
What's more, this casts a shadow over my actual beliefs.
This is in spite of the fact that I love EY's writing, and actually enjoy his fearless geeky humor ("hit by a meteorite" is indeed more fun than the conventional "hit by a bus").
The fear of being represented by EY is mostly due to what he's saying, not how he's saying it. That is, even if he were always dignified and measured, he'd catch nearly as much flak. If he'd avoided certain topics entirely, that would have made a significant difference, but on the other hand, he's effectively counter-signaled that he's fully honest and uncensored in public (of course he is probably not, exactly), which I think is also valuable.
I think EY can win by saying enough true things, convincingly, that smart people will be persuaded that he's credible. It's perhaps true that better PR will speed the process - by enough for it to be worth it? That's up to him.
The comments in this diavlog with Scott Aaronson - while some are by obvious axe-grinders - are critical of EY's manner. People appear to hate nothing more than (what they see as) undeserved confidence. Who knows how prevalent this adverse reaction to EY is, since the set of commenters is self-selecting.
People who are floundering in a debate with EY (e.g. Jason Lanier) seem to think they can bank on a "you crazy low-status sci-fi nerd" rebuttal to EY. This can score huge with lazy or unintellectual people if it's allowed to succeed.