Comments on: Analysis of Spielberg’s move, AI http://www.foresight.org/nanodot/?p=641 examining transformative technology Fri, 03 May 2013 19:22:42 +0000 hourly 1 http://wordpress.org/?v=3.0.4 By: Elan Horsfield http://www.foresight.org/nanodot/?p=641#comment-40056 Elan Horsfield Mon, 18 Sep 2006 19:06:07 +0000 http://www.foresight.org/nanodot/?p=641#comment-40056 There is the possibility that "AI" is not about AI at all, but about human relationships, about our relations to different groups of humans, about our relationships to individuals. Judging it on how near it comes to achieving a good explanation of Artificial Intelligence may be off course. Judging it on the muddled ending may be unfair, since many great films have had muddled plots. "Once upon a Time in America" for example. The boy comes accross as a human, yet we are constantly reminded during the film that he is a robot. He is as rejected by humans as much as he seems at first to reject his fellow robots. A parallel might be someone born black, brought up by whites, then rejected as he loses his child appeal, and from his point of view, puzzlement at his view of fellow blacks who are treated as lesser humans. I think I've been quoting from several plots here. There is the possibility that “AI” is not about AI at all, but about human relationships,
about our relations to different groups of humans, about our relationships to individuals.

Judging it on how near it comes to achieving a good explanation of Artificial Intelligence
may be off course.

Judging it on the muddled ending may be unfair, since many great films have had muddled
plots. “Once upon a Time in America” for example.

The boy comes accross as a human, yet we are constantly reminded during the film that he
is a robot. He is as rejected by humans as much as he seems at first to reject his fellow
robots.
A parallel might be someone born black, brought up by whites, then rejected as he loses
his child appeal, and from his point of view, puzzlement at his view of fellow blacks who are
treated as lesser humans.

I think I’ve been quoting from several plots here.

]]> By: fred http://www.foresight.org/nanodot/?p=641#comment-5309 fred Fri, 24 Jun 2005 18:36:29 +0000 http://www.foresight.org/nanodot/?p=641#comment-5309 You are missing the point of the entire movie. What Spielberg was creating was a story of genesis. AI is the bible story for the next group of dominent creatures on Earth. Forget about sci-fi theories and think about how much we humans would embrace the one and only true story behind our creation if it could be proven (although we may not like what we find). This is what we were witnessing as an audience. To compare AI to ET is really simplistic. You are missing the point of the entire movie.
What Spielberg was creating was a story of genesis. AI is the bible story for the next group of dominent creatures on Earth. Forget about sci-fi theories and think about how much we humans would embrace the one and only true story behind our creation if it could be proven (although we may not like what we find). This is what we were witnessing as an audience.
To compare AI to ET is really simplistic.

]]> By: MarkGubrud http://www.foresight.org/nanodot/?p=641#comment-1754 MarkGubrud Mon, 09 Jul 2001 02:57:01 +0000 http://www.foresight.org/nanodot/?p=641#comment-1754 <p><strong>Verbose AI</strong></p> <p>If you go to Yudkowsky's website, you find about a billion words, and if you start trying to wade through them, after about half an hour you find that you still don't have a clue how "friendly AI" is supposed to work. So you give up.</p> <p>Could someone write about a four-page paper, perhaps with a flow chart or two, explaining the overall architecture which is supposed to enable self-improving artificial intelligence while guaranteeing that it will always be "friendly" to humans?</p> <p>The impression I get is that what you really have is a very confused tangle of fragmentary thoughts, presented in a way which demands so much of readers that it either forces them to back down (thus achieving intimidation) or else make such an investment of time that they are motivated to sign on as true believers. I am prepared to believe that Eliezer may have something, but if it can't be boiled down to a succinct top-level description, and decomposed hierarchically, then I suspect that maybe it actually isn't there.</p> Verbose AI

If you go to Yudkowsky's website, you find about a billion words, and if you start trying to wade through them, after about half an hour you find that you still don't have a clue how "friendly AI" is supposed to work. So you give up.

Could someone write about a four-page paper, perhaps with a flow chart or two, explaining the overall architecture which is supposed to enable self-improving artificial intelligence while guaranteeing that it will always be "friendly" to humans?

The impression I get is that what you really have is a very confused tangle of fragmentary thoughts, presented in a way which demands so much of readers that it either forces them to back down (thus achieving intimidation) or else make such an investment of time that they are motivated to sign on as true believers. I am prepared to believe that Eliezer may have something, but if it can't be boiled down to a succinct top-level description, and decomposed hierarchically, then I suspect that maybe it actually isn't there.

]]> By: MarkGubrud http://www.foresight.org/nanodot/?p=641#comment-1753 MarkGubrud Mon, 09 Jul 2001 02:41:11 +0000 http://www.foresight.org/nanodot/?p=641#comment-1753 <p><strong>Trustworthy systems?</strong></p> <blockquote> <p>regardless of smartness, the AI will face a philosophical crises that will probably be the end of it. For example, David is taught to love humans, but humans, from the start, did not love him by forcing him to love rather than trusting him.</p> </blockquote> <p>I can agree that AIs will face philosophical crises, just as humans do, only probably much worse in proportion to how much more intelligent they are. In my view, love is the only way out of the fundamental crisis, "Why even exist?" I also agree that the idea of making an AI love is silly, unless the purpose is to create a love doll. What I don't understand is why you think humans should trust AIs. I know Yudkowsky claims to have a way of guaranteeing that AI's will be "friendly," but I don't trust his claim any more than I trust a superintelligent robot in a philosophical crisis.</p> Trustworthy systems?

regardless of smartness, the AI will face a philosophical crises that will probably be the end of it. For example, David is taught to love humans, but humans, from the start, did not love him by forcing him to love rather than trusting him.

I can agree that AIs will face philosophical crises, just as humans do, only probably much worse in proportion to how much more intelligent they are. In my view, love is the only way out of the fundamental crisis, "Why even exist?" I also agree that the idea of making an AI love is silly, unless the purpose is to create a love doll. What I don't understand is why you think humans should trust AIs. I know Yudkowsky claims to have a way of guaranteeing that AI's will be "friendly," but I don't trust his claim any more than I trust a superintelligent robot in a philosophical crisis.

]]> By: redbird http://www.foresight.org/nanodot/?p=641#comment-1752 redbird Mon, 02 Jul 2001 15:53:12 +0000 http://www.foresight.org/nanodot/?p=641#comment-1752 <p><strong>Eratta</strong></p> <p>Well, after I submitted this I realized that I had made an error: mecca should be mecha. Oh well, I guess I can't catch every error before posting to the world.<br /> <br /> Also, you should check out <a href="http://sysopmind.com/archive-sl4/0107/0006.html">Eliezer's analysis</a>. It convers a lot more stuff than mine does.</p> Eratta

Well, after I submitted this I realized that I had made an error: mecca should be mecha. Oh well, I guess I can't catch every error before posting to the world.

Also, you should check out Eliezer's analysis. It convers a lot more stuff than mine does.

]]>