tag:blogger.com,1999:blog-11168555.post7780493301662739709..comments2024-03-28T03:22:24.202-04:00Comments on The Multiverse According to Ben: Will China Build AGI First?Benhttp://www.blogger.com/profile/12743597120529571571noreply@blogger.comBlogger12125tag:blogger.com,1999:blog-11168555.post-77255731146410373102009-12-14T20:20:57.934-05:002009-12-14T20:20:57.934-05:00Cost of living? Only if you do not take into accou...Cost of living? Only if you do not take into account the cost of real estate, health and education.Lihttps://www.blogger.com/profile/07988585475259144630noreply@blogger.comtag:blogger.com,1999:blog-11168555.post-17582515607840124832009-08-02T08:36:17.621-04:002009-08-02T08:36:17.621-04:00Matt: there are moral philosophers involved with ...Matt: there <i>are</i> moral philosophers involved with the AGI community already ... but as one might expect, this has not resulted in the moral issues being "solved" ;-)Ben Goertzelhttps://www.blogger.com/profile/01289041122724284772noreply@blogger.comtag:blogger.com,1999:blog-11168555.post-48963934287653481022009-08-02T07:59:31.128-04:002009-08-02T07:59:31.128-04:00Yesterday I spend a couple of hours watching onlin...Yesterday I spend a couple of hours watching online videos of some of your talks on AGI. Having a background in philosophy, I was surprised to find most of the technical stuff easy to follow. But I have some comments on AGI and ethics.<br /><br />You seem to subscribe to the view that if we can make thinking machines who are also ethical, we basically have nothing to worry about. My contention is that we don't know this, and to find out we need to know more about what morality is. Let me explain.<br /><br />We all know that in some cases it is right to kill other humans. From this it follows that a moral machine will sometimes kill humans.<br /><br />It could be argued that this is not a big problem. We want the machines to do good, and if that entails killing humans in some extreme situations then so be it. What we fear is the total annihilation of the human race by machines, and this can never be moral.<br /><br />Or can it? We would really want to know this for a fact before we build a thinking machine, just as we want to rule out any risk of inadvertantly creating a black hole before we build a high-energy particle accellerator.<br /><br />We have basically three serious hypthesis about what morality is: Kantian ethics, consequentialism and contractualism. If Kantian ethics is true, it is highly unlikely in my judgment that exterminating the human race can ever be the right thing to do. But if consequentialism is true on the other hand, it's relatively easy to specify under which conditions a holocaust should be the ethically superior solution.<br /><br />I personally subscibe to the contractualist theory of morality, which I understand as the view that moral responsibilities has it's root in cooperation and risk management. From this it could follow that machines have moral responsibilities toward humans, if but only if there's a strong (metaphysical) reason why machines will be better off if they cooperate with humans. But if the machines could create a world in which such cooperation is unnecessary, then they would not have any moral duties to preserve the human race, and if they could be much better off in a world without humans (or with a small group of humans in a zoo) they would have a moral responsibility to bring that about.<br /><br />Perhaps you should engage some moral philosophers in the AI community to help you think clearer about this, or this problem has already been solved and I just don't know about it.Mattnoreply@blogger.comtag:blogger.com,1999:blog-11168555.post-51694486514395301222009-07-27T21:32:23.272-04:002009-07-27T21:32:23.272-04:00Spiral Shell...
You asked
Has or will the Chin...Spiral Shell... <br /><br />You asked<br /><br /><i><br />Has or will the China Brain project include any planning for niceness in the AGI road-map, or will it be more like pulling a super-brain out of a hat?<br /></i><br /><br />The current plan for the "China Brain" project is to use OpenCog for cognition, and an evolved neural net for perception and action.<br /><br />OpenCog is going to use a variant of the Psi motivational/emotional framework outlined by Joscha Bach, which provides a systematic framework for handling "Friendliness" (though provides no absolute guarantees, of course).<br /><br />You asked<br /><br /><i><br />I must admit, I am a little scared of closedness and secrecy of the Chinese government, most likely uneducatedly so. But is there nothing that bothers you about the possibility of the first AGI coming to fruition under any 'less open' conditions? What can be done about this?<br /></i><br /><br />The Chinese government has some aspects I don't care for. So does the US government. <br /><br />If the code is open-source and is comprehended by a diverse international community of researchers, I'm not that concerned about where the core programming team is located.Ben Goertzelhttps://www.blogger.com/profile/01289041122724284772noreply@blogger.comtag:blogger.com,1999:blog-11168555.post-22061524677565298432009-07-27T20:53:25.792-04:002009-07-27T20:53:25.792-04:00I may be wrong, but the China Brain project does n...I may be wrong, but the China Brain project does not look like it will suddenly 'wake up', but has the potential to move ahead. Even if we think the project will not take off like a rocket, the issue of pulling a friendly AGI out of mind-space still seems important early on. Has or will the China Brain project include any planning for niceness in the AGI road-map, or will it be more like pulling a super-brain out of a hat? <br /> <br />Bullwinkle:Hey Rocky, watch me pull a *god* out of my hat! <br />Rocky:Aww bullwinkle, that trick never works.<br />Bullwinkle:This time for sure!...Presto!!! <br />god-agi-artillect:RRROOOOOAAARRR!!!<br /><br />It would be great there a Chinese version or branch of the Singularity Institute or Foresight Institute. Perhaps this would also help to decrease the "China Isolation" factor...<br />I must admit, I am a little scared of closedness and secrecy of the Chinese government, most likely uneducatedly so. But is there nothing that bothers you about the possibility of the first AGI coming to fruition under any 'less open' conditions? What can be done about this?<br /><br />*p.s. I really want to travel to China, it looks like it would be a great experience!<br /><br />-- Cheers, Adamspiral_shellhttps://www.blogger.com/profile/17378584752759768681noreply@blogger.comtag:blogger.com,1999:blog-11168555.post-39312110565982271032009-07-27T20:51:01.723-04:002009-07-27T20:51:01.723-04:00This comment has been removed by the author.spiral_shellhttps://www.blogger.com/profile/17378584752759768681noreply@blogger.comtag:blogger.com,1999:blog-11168555.post-32291871201095858172009-07-27T07:44:54.878-04:002009-07-27T07:44:54.878-04:00Hi Stefan,
I don't know much about the diffi...Hi Stefan, <br /><br />I don't know much about the difficulty of finding quality commercial software programmers in China.<br /><br />But I found a bunch of very good *AI research students* at Wuhan University and Xiamen University ... with a lot of creative and interesting ideas pertinent to AGI.<br /><br />Commercial software engineering is different, of course.<br /><br />Part of the problem you're seeing may be due to the way their education system is structured, with Software Engineering (at least at both Xiamen and Wuhan universities) totally separate from Computer Science. So, the professional software engineers don't have much background outside software engineering (lacking the diverse input that would foster creativity), whereas the computer science and AI researchers don't have that much software engineering background.<br /><br />At first I thought the Chinese grad students lacked creativity, but once I got to know them better I found I was mistaken. They were (with some exceptions) more shy about sharing their ideas than Americans would be; but once I got to know them better, they turned out to have a load of great ideas.<br /><br />So ... as with anything else in China, the conclusion has to be "it's complex" ;-)<br /><br />-- BenBen Goertzelhttps://www.blogger.com/profile/01289041122724284772noreply@blogger.comtag:blogger.com,1999:blog-11168555.post-7660843088786037412009-07-27T02:01:32.613-04:002009-07-27T02:01:32.613-04:00Interesting take on China and AGI. Having lived in...Interesting take on China and AGI. Having lived in China for the better part of the past 25 years and having been the operations director of a China based IT outsourcing company, I do not share your positive outlook.<br /><br />It is exceedingly hard to find good programmers and creativity is a big weakness of Chinese programmers as well. The latter being a particularly difficult hurdle for such a highly complex and uncharted problem of AGI development.<br /><br />Ultimately however I do agree that I would not mind the Chinese developing AGI... That has more to do with my thinking on the nature of AGI though than that on the Chinese government :-)Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-11168555.post-63902470791967672852009-07-26T19:36:27.364-04:002009-07-26T19:36:27.364-04:00Oh I wouldn't worry about Alfred E. Newman I f...Oh I wouldn't worry about Alfred E. Newman I found that suitably chillingflamoothttps://www.blogger.com/profile/00743008986393837515noreply@blogger.comtag:blogger.com,1999:blog-11168555.post-61069715541138112022009-07-26T16:48:41.653-04:002009-07-26T16:48:41.653-04:00Oh, and, TransAlchemy: the game I am playing is no...Oh, and, TransAlchemy: the game <b>I</b> am playing is not one of "waiting" ;-)Ben Goertzelhttps://www.blogger.com/profile/01289041122724284772noreply@blogger.comtag:blogger.com,1999:blog-11168555.post-17111389956421912552009-07-26T16:47:56.070-04:002009-07-26T16:47:56.070-04:00TransAlchemy: The idea that "the best path to...TransAlchemy: The idea that "the best path to AGI is to have the right elite group secretly develop it" is sort of like the idea that "the best government is a benevolent dictatorship" ;-)Ben Goertzelhttps://www.blogger.com/profile/01289041122724284772noreply@blogger.comtag:blogger.com,1999:blog-11168555.post-74750561302539559912009-07-26T16:45:18.532-04:002009-07-26T16:45:18.532-04:00Cheers on doing your best to keep AGI from governm...Cheers on doing your best to keep AGI from governments, though Im not sure how long this may last.<br /><br />Yet in your effort to keep AGI from governments, your doing something I fear more. Your exposing AGI to the internet, in essence letting the internet be it's womb. <br /><br />In either case centralized or uncentralized creation of AGI has it pitfalls. <br /><br />Now we play the waiting game ;)TransAlchemyhttp://www.transalchemy.comnoreply@blogger.com