• Claim: human-level AI will get cheap fast

    To be more precise: within five years of any AI having broadly human-level intelligence, AI will be much more cost-effective than a human for virtually all tasks

    That is to say, not long after AI can do roughly what a human can do at some cost, it will be able to do at least what a human can do, much cheaper than a human.

    This could be supported by

    AI

  • Support: even normal ratel of progress are likely to bring down the cost of human-level AI in five years

    Normal rates of AI progress should bridge a gap from ‘human-level at some cost’ to ‘much more cost effective than a human for virtually all tasks’ in less than five years

  • Support: once there is human-level AI, progress will be fast enough to bring costs way down.

    Human level AI will bring about - or at least coincide with - fast progress in AI. This progress will probably be fast enough to make human-level AI much more cost effective than a human within five years of its development.

  • Conjunct: progress will be fast once there is human-level AI

    There are various reasons this could happen. AI itself could be helpful with AI progress, though this seems somewhat unlikely if the AI in question does the tasks that a human can do at a much higher cost. But ‘human-level’ isn’t necessarily so simple. An AI which is not cost-competitive with a human on most tasks could be very good at some specific, well defined task. Such as developing AI, perhaps.

    Another reason progress might be expected to be fast just after human-level AI is developed is that at that point, AI would be on the brink of becoming much more economically important. This might be expected to spur interest and investment.

    Further, we might just expect human-level AI to be achieved at a time when AI is experiencing fast growth, as disproportionately many achievements will be reached in such times.


  • Conjunct: fast AI progress will not slow down substantially in the five years following human-level AI development

    That is, the fast AI progress needed to make AI much cheaper is very unlikely to exist when human-level AI is developed, but then quickly disappear. If things are moving fast just after human-level AI is developed, they will continue to do so for some time.

  • Conjunct: The first human-level AI is unlikely to cost more than a thousand times as much as a human to run

    [xxx]

  • Support: Not much progress will be needed to go past human-level, and creatures a little beyond us will be hugely valuable

    Suppose that the amount of thinking that AI researchers do, and the number of clever insights they come up with as a result, remains fixed. It might still be that creatures who are only a tiny bit better than us - in terms of design effort - are much more useful, in terms of their capacities for valuable work.

    By pushing past human level then, researchers would automatically have fast progress in AI, at least in practical terms.

  • Support: human-level AI will bring about faster AI progress.

  • Support: Human-level AI is likely to come abut during a period of especially fast AI progress

    There will be faster and slower bouts of progress, and disproportionately much research will happen during fast bouts.

  • Support: The economic importance of human-level AI will prompt much larger investments in research at around that time than previously in history, which will produce much faster research.

  • [xxx]

  • xxx

  • xxx

  • xxx

{"cards":[{"_id":"3af4fe060ef4cd8a8f000073","treeId":"3af4f8080ef4cd8a8f000071","seq":1,"position":1,"parentId":null,"content":"##Claim: human-level AI will get cheap fast\n\nTo be more precise: within **five years** of any AI having **broadly human-level** intelligence, AI will be much **more cost-effective than a human** for virtually all tasks\n\nThat is to say, not long after AI can do roughly what a human can do *at some cost*, it will be able to do at least what a human can do, much cheaper than a human. \n\nThis could be supported by \n\n![AI](http://www.govacuum.us/blog/wp-content/uploads/2012/09/Roomba-vacuums-are-hands-free-and-easy-to-use.jpeg)"},{"_id":"3af55e2f0ef4cd8a8f000075","treeId":"3af4f8080ef4cd8a8f000071","seq":1,"position":0.5,"parentId":"3af4fe060ef4cd8a8f000073","content":"##Support: even normal ratel of progress are likely to bring down the cost of human-level AI in five years\n\nNormal rates of AI progress should bridge a gap from 'human-level at some cost' to 'much more cost effective than a human for virtually all tasks' in less than five years"},{"_id":"3af5612c0ef4cd8a8f000077","treeId":"3af4f8080ef4cd8a8f000071","seq":1,"position":1,"parentId":"3af55e2f0ef4cd8a8f000075","content":"[put in stuff from #place1]"},{"_id":"3af55d180ef4cd8a8f000074","treeId":"3af4f8080ef4cd8a8f000071","seq":1,"position":1,"parentId":"3af4fe060ef4cd8a8f000073","content":"##Support: once there is human-level AI, progress will be fast enough to bring costs way down.\n\nHuman level AI will bring about - or at least coincide with - fast progress in AI. This progress will probably be fast enough to make human-level AI much more cost effective than a human within five years of its development. \n\n![](http://www.telecoms.com/wp-content/blogs.dir/1/files/2011/06/growth-chart1.jpg)"},{"_id":"3af57acf0ef4cd8a8f000078","treeId":"3af4f8080ef4cd8a8f000071","seq":1,"position":1,"parentId":"3af55d180ef4cd8a8f000074","content":"##Conjunct: progress will be fast once there is human-level AI\n\nThere are various reasons this could happen. AI itself could be helpful with AI progress, though this seems somewhat unlikely if the AI in question does the tasks that a human can do at a much higher cost. But 'human-level' isn't necessarily so simple. An AI which is not cost-competitive with a human on most tasks could be very good at some specific, well defined task. Such as developing AI, perhaps. \n\nAnother reason progress might be expected to be fast just after human-level AI is developed is that at that point, AI would be on the brink of becoming much more economically important. This might be expected to spur interest and investment. \n\nFurther, we might just expect human-level AI to be achieved at a time when AI is experiencing fast growth, as disproportionately many achievements will be reached in such times. \n\n![](http://bookriot.com/wp-content/uploads/2013/03/server-farm-shot.jpg)\n\n-----------------------------------------------"},{"_id":"3af5c3110ef4cd8a8f00007c","treeId":"3af4f8080ef4cd8a8f000071","seq":1,"position":1,"parentId":"3af57acf0ef4cd8a8f000078","content":"##Support: Not much progress will be needed to go past human-level, and creatures a little beyond us will be hugely valuable\n\nSuppose that the amount of thinking that AI researchers do, and the number of clever insights they come up with as a result, remains fixed. It might still be that creatures who are only a tiny bit better than us - in terms of design effort - are much more useful, in terms of their capacities for valuable work.\n\nBy pushing past human level then, researchers would automatically have fast progress in AI, at least in practical terms. "},{"_id":"3af5d47b0ef4cd8a8f00007e","treeId":"3af4f8080ef4cd8a8f000071","seq":1,"position":1,"parentId":"3af5c3110ef4cd8a8f00007c","content":"[from under #place2]"},{"_id":"3af5d4090ef4cd8a8f00007d","treeId":"3af4f8080ef4cd8a8f000071","seq":1,"position":2,"parentId":"3af57acf0ef4cd8a8f000078","content":"##Support: human-level AI will bring about faster AI progress."},{"_id":"3af5d6b10ef4cd8a8f00007f","treeId":"3af4f8080ef4cd8a8f000071","seq":1,"position":1,"parentId":"3af5d4090ef4cd8a8f00007d","content":"[from under #place3]"},{"_id":"3af5d6e10ef4cd8a8f000080","treeId":"3af4f8080ef4cd8a8f000071","seq":1,"position":3,"parentId":"3af57acf0ef4cd8a8f000078","content":"##Support: Human-level AI is likely to come abut during a period of especially fast AI progress\n\nThere will be faster and slower bouts of progress, and disproportionately much research will happen during fast bouts.\n\n![](http://www.eecs.umich.edu/ai/html/pics/aimain2.jpg)"},{"_id":"3af5d8930ef4cd8a8f000081","treeId":"3af4f8080ef4cd8a8f000071","seq":1,"position":1,"parentId":"3af5d6e10ef4cd8a8f000080","content":"xxx"},{"_id":"3af5dd1b0ef4cd8a8f000082","treeId":"3af4f8080ef4cd8a8f000071","seq":1,"position":2,"parentId":"3af5d6e10ef4cd8a8f000080","content":"xxx"},{"_id":"3af5dd460ef4cd8a8f000083","treeId":"3af4f8080ef4cd8a8f000071","seq":1,"position":4,"parentId":"3af57acf0ef4cd8a8f000078","content":"##Support: The economic importance of human-level AI will prompt much larger investments in research at around that time than previously in history, which will produce much faster research. \n\n![](http://coloradopeakpolitics.com/wp-content/uploads/2013/08/Money-II.jpg)"},{"_id":"3af5e22f0ef4cd8a8f000084","treeId":"3af4f8080ef4cd8a8f000071","seq":2787701,"position":1,"parentId":"3af5dd460ef4cd8a8f000083","content":"xxx"},{"_id":"3af59b7f0ef4cd8a8f00007a","treeId":"3af4f8080ef4cd8a8f000071","seq":1,"position":1.5,"parentId":"3af55d180ef4cd8a8f000074","content":"##Conjunct: fast AI progress will not slow down substantially in the five years following human-level AI development\n\nThat is, the fast AI progress needed to make AI much cheaper is very unlikely to exist when human-level AI is developed, but then quickly disappear. If things are moving fast just after human-level AI is developed, they will continue to do so for some time.\n\n![](http://stats.areppim.com/ressources/mobixsubscr_2011x25_800x574.png)"},{"_id":"3af590650ef4cd8a8f000079","treeId":"3af4f8080ef4cd8a8f000071","seq":1,"position":2,"parentId":"3af55d180ef4cd8a8f000074","content":"##Conjunct: The first human-level AI is unlikely to cost more than a thousand times as much as a human to run\n\n[xxx]\n"},{"_id":"3af5c2d10ef4cd8a8f00007b","treeId":"3af4f8080ef4cd8a8f000071","seq":1,"position":1,"parentId":"3af590650ef4cd8a8f000079","content":"[xxx]"}],"tree":{"_id":"3af4f8080ef4cd8a8f000071","name":"AI Progress","publicUrl":"ai-progress"}}