{"id":131,"date":"2011-12-15T17:56:24","date_gmt":"2011-12-16T01:56:24","guid":{"rendered":"http:\/\/facingthesing.wpengine.com\/?p=131"},"modified":"2014-12-01T13:48:07","modified_gmt":"2014-12-01T21:48:07","slug":"playing-taboo-with-intelligence","status":"publish","type":"post","link":"https:\/\/intelligenceexplosion.com\/sv\/2011\/playing-taboo-with-intelligence\/","title":{"rendered":"Att spela Tabu med &#8221;Intelligens&#8221;"},"content":{"rendered":"<p class=\"qtranxs-available-languages-message qtranxs-available-languages-message-sv\">Tyv\u00e4rr \u00e4r denna artikel enbart tillg\u00e4nglig p\u00e5 <a href=\"https:\/\/intelligenceexplosion.com\/en\/wp-json\/wp\/v2\/posts\/131\" class=\"qtranxs-available-language-link qtranxs-available-language-link-en\" title=\"English\">English<\/a>, <a href=\"https:\/\/intelligenceexplosion.com\/fr\/wp-json\/wp\/v2\/posts\/131\" class=\"qtranxs-available-language-link qtranxs-available-language-link-fr\" title=\"Fran\u00e7ais\">Fran\u00e7ais<\/a>, <a href=\"https:\/\/intelligenceexplosion.com\/ru\/wp-json\/wp\/v2\/posts\/131\" class=\"qtranxs-available-language-link qtranxs-available-language-link-ru\" title=\"\u0440\u0443\u0441\u0441\u043a\u0438\u0439\">\u0440\u0443\u0441\u0441\u043a\u0438\u0439<\/a>, <a href=\"https:\/\/intelligenceexplosion.com\/sk\/wp-json\/wp\/v2\/posts\/131\" class=\"qtranxs-available-language-link qtranxs-available-language-link-sk\" title=\"Sloven\u010dina\">Sloven\u010dina<\/a>, <a href=\"https:\/\/intelligenceexplosion.com\/zh\/wp-json\/wp\/v2\/posts\/131\" class=\"qtranxs-available-language-link qtranxs-available-language-link-zh\" title=\"\u4e2d\u6587\">\u4e2d\u6587<\/a> och <a href=\"https:\/\/intelligenceexplosion.com\/it\/wp-json\/wp\/v2\/posts\/131\" class=\"qtranxs-available-language-link qtranxs-available-language-link-it\" title=\"Italiano\">Italiano<\/a>.<\/p><p>Eliezer Yudkowsky <a href=\"http:\/\/lesswrong.com\/lw\/vb\/efficient_crossdomain_optimization\/\">recounts<\/a>:<\/p>\n<blockquote><p>Years ago when I was on a panel with <a href=\"http:\/\/en.wikipedia.org\/wiki\/Jaron_Lanier\">Jaron Lanier<\/a>, he had offered some elaborate argument that no machine could be intelligent, because it was just a machine and to call it \u201cintelligent\u201d was therefore bad poetry, or something along those lines. Fed up, I finally snapped: \u201cDo you mean to say that if I write a computer program and that computer program rewrites itself and rewrites itself and builds its own nanotechnology and zips off to Alpha Centauri and builds its own Dyson Sphere, that computer program is not intelligent?\u201d<\/p><\/blockquote>\n<p>Much of the confusion about AI comes from disagreements about the meaning of \u201cintelligence.\u201d<\/p>\n<p>Let me clear things up with a <a href=\"http:\/\/lesswrong.com\/lw\/np\/disputing_definitions\/\">parable<\/a>:<\/p>\n<blockquote><p><em>If a tree falls in the forest, and no one hears it, does it make a sound?<\/em><\/p>\n<p>A<span class=\"small-caps\">lbert<\/span>: \u201cOf course it does. What kind of silly question is that? Every time I\u2019ve listened to a tree fall, it made a sound, so I\u2019ll guess that other trees falling also make sounds. I don\u2019t believe the world changes around when I\u2019m not looking.\u201d<\/p>\n<p>B<span class=\"small-caps\">arry<\/span>: \u201cWait a minute. If no one hears it, how can it be a sound?\u201d<\/p><\/blockquote>\n<p>Albert and Barry are not arguing about facts, but about <a href=\"http:\/\/lesswrong.com\/lw\/no\/how_an_algorithm_feels_from_inside\/\">definitions<\/a>:<\/p>\n<blockquote><p>The first person is speaking as if \u201csound\u201d means acoustic vibrations in the air; the second person is speaking as if \u201csound\u201d means an auditory experience in a brain. If you ask \u201cAre there acoustic vibrations?\u201d or \u201cAre there auditory experiences?\u201d, the answer is at once obvious. And so the argument is really about the definition of the word \u201csound.\u201d<\/p><\/blockquote>\n<p>We need not argue about definitions. Wherever we might be using different meanings for words, we can cut to the chase by <a href=\"http:\/\/lesswrong.com\/lw\/nv\/replace_the_symbol_with_the_substance\/\">replacing<\/a> the confusing <em>symbol<\/em> (a word) with its intended <em>substance<\/em> (the meaning you intend).<\/p>\n<p>This is like playing the game <a href=\"http:\/\/en.wikipedia.org\/wiki\/Taboo_(game)\"><em>Taboo<\/em><\/a> (by Hasbro). In <em>Taboo<\/em>, you have to describe something to your partner <a href=\"http:\/\/lesswrong.com\/lw\/nu\/taboo_your_words\">without using a certain list of words<\/a>:<\/p>\n<blockquote><p>For example, you might have to get your partner to say \u201cbaseball\u201d without using the words \u201csport,\u201d \u201cbat,\u201d \u201chit,\u201d \u201cpitch,\u201d \u201cbase,\u201d or, of course, \u201cbaseball.\u201d<\/p><\/blockquote>\n<p>This game is good practice for a discussion about AI. If two people notice they\u2019re using different definitions of \u201cintelligence,\u201d they don\u2019t need to argue about whose definition is \u201cright.\u201d They can taboo the word \u201cintelligence\u201d and talk about \u201canalytic ability\u201d or \u201cproblem-solving ability\u201d or whatever it is they mean by \u201cintelligence.\u201d Now they are closer to arguing about facts instead of definitions.<\/p>\n<p>Shane Legg once collected seventy-one definitions of intelligence.<a id=\"fn1x10-bk\" href=\"#fn1x10\"><sup>1<\/sup><\/a> Scanning through the definitions for commonly occurring features, he notes that people seem to think intelligence is:<\/p>\n<ul>\n<li>A property that an individual agent has as it interacts with its environment or environments.<\/li>\n<li>. . . Related to the agent\u2019s ability to succeed or profit with respect to some goal or objective.<\/li>\n<li>Depend[ent] on how able [the] agent is to adapt to different objectives and environments.<\/li>\n<\/ul>\n<p>If we combine these features, we get something like this:<\/p>\n<blockquote><p>Intelligence measures an agent\u2019s ability to achieve goals in a wide range of environments.<a id=\"fn2x10-bk\" href=\"#fn2x10\"><sup>2<\/sup><\/a><\/p><\/blockquote>\n<p>This is, after all, the kind of intelligence that let humans dominate all other species on the planet, and the kind of intelligence that leaves us superior to machines (at least for now). Termites built bigger cities (relative to their body size) and whales had bigger brains, but humans had the intelligence to adapt to almost every terrestrial environment and make tools and spears and boats and farms. <a href=\"http:\/\/en.wikipedia.org\/wiki\/Watson_(computer)\">Watson<\/a> can beat us in <em>Jeopardy!<\/em> and <a href=\"http:\/\/www.wolframalpha.com\/\"><em>WolframAlpha<\/em><\/a> can process computable knowledge better than we can, but drop either of them in a lake or expect them to find their own electricity and they are helpless. Unlike Watson and <em>WolframAlpha<\/em>, humans have <em>cross-domain<\/em> goal-optimizing ability. <a href=\"http:\/\/lesswrong.com\/lw\/vb\/efficient_crossdomain_optimization\/\">Or<\/a>:<\/p>\n<blockquote><p>A bee builds hives, and a beaver builds dams; but a bee doesn\u2019t build dams and a beaver doesn\u2019t build hives. A human, watching, thinks, \u201cOh, I see how to do it\u201d and goes on to build a dam using a honeycomb structure for extra strength.<\/p><\/blockquote>\n<p>But wait a minute. Suppose Bill Gates gives me ten billion dollars. I now have <em>much<\/em> greater ability to \u201cachieve goals in a wide range of environments,\u201d but would we say my \u201cintelligence\u201d has gone up? I doubt it. If we want to measure an agent\u2019s \u201cintelligence,\u201d we should take that agent\u2019s ability to optimize for its goals in a wide range of environments\u2014its \u201coptimization power,\u201d we might say\u2014and <em>divide<\/em> that by its resources used to do so:<\/p>\n<p><a href=\"http:\/\/facingthesing.wpengine.com\/wp-content\/uploads\/2011\/12\/intelligence.png\"><img loading=\"lazy\" decoding=\"async\" class=\"aligncenter size-full wp-image-146\" title=\"intelligence\" alt=\"\" src=\"http:\/\/facingthesing.wpengine.com\/wp-content\/uploads\/2011\/12\/intelligence.png\" width=\"650\" height=\"155\" srcset=\"https:\/\/intelligenceexplosion.com\/wp-content\/uploads\/2011\/12\/intelligence.png 650w, https:\/\/intelligenceexplosion.com\/wp-content\/uploads\/2011\/12\/intelligence-300x71.png 300w\" sizes=\"auto, (max-width: 650px) 100vw, 650px\" \/><\/a><\/p>\n<p>This definition sees intelligence as <em>efficient<\/em> cross-domain optimization. Intelligence is what allows an agent to steer the future, a power that is amplified by the resources at its disposal.<\/p>\n<p>This may or may not match your own intuitive definition of \u201cintelligence.\u201d But it doesn\u2019t matter. I\u2019ve tabooed \u201cintelligence.\u201d I\u2019ve replaced the symbol with the substance. When discussing AI, I can speak only of \u201cefficient cross-domain optimization,\u201d and nothing about your preferred definition of \u201cintelligence\u201d will matter for anything I say.<\/p>\n<p>Now, \u201cintelligence\u201d is shorter, so I\u2019d prefer to just say <em>that<\/em>. But it\u2019s best if you always read \u201cintelligence\u201d (when I use it) as \u201cefficient cross-domain optimization.\u201d<\/p>\n<p>And, now that we understand what I mean by \u201cintelligence,\u201d we\u2019re ready to talk about AI.<\/p>\n<p class=\"footnotes\">* * *<\/p>\n<p><small><a id=\"fn1x10\" href=\"#fn1x10-bk\"><sup>1<\/sup><\/a>Shane Legg and Marcus Hutter, <em>A Collection of Definitions of Intelligence<\/em> (Manno-Lugano, Switzerland: IDSIA, July 15, 2007), <a class=\"url\" href=\"http:\/\/www.idsia.ch\/idsiareport\/IDSIA-07-07.pdf\">http:\/\/www.idsia.ch\/idsiareport\/IDSIA-07-07.pdf<\/a>.<\/small><\/p>\n<p><small><a id=\"fn2x10\" href=\"#fn2x10-bk\"><sup>2<\/sup><\/a>Shane Legg and Marcus Hutter, <em>A Formal Measure of Machine Intelligence<\/em> (Manno-Lugano, Switzerland: IDSIA, April 12, 2006), <a class=\"url\" href=\"http:\/\/www.idsia.ch\/idsiareport\/IDSIA-10-06.pdf\">http:\/\/www.idsia.ch\/idsiareport\/IDSIA-10-06.pdf<\/a>.<\/small><\/p>","protected":false},"excerpt":{"rendered":"<p>Tyv\u00e4rr \u00e4r denna artikel enbart tillg\u00e4nglig p\u00e5 English, Fran\u00e7ais, \u0440\u0443\u0441\u0441\u043a\u0438\u0439, Sloven\u010dina, \u4e2d\u6587 och Italiano.Eliezer Yudkowsky recounts: Years ago when I was on a panel with Jaron Lanier, he had offered some elaborate argument that no machine could be intelligent, because&hellip;  <a href=\"https:\/\/intelligenceexplosion.com\/sv\/2011\/playing-taboo-with-intelligence\/\">continue reading<\/a> &raquo;<\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"_monsterinsights_skip_tracking":false,"_monsterinsights_sitenote_active":false,"_monsterinsights_sitenote_note":"","_monsterinsights_sitenote_category":0,"footnotes":""},"categories":[4],"tags":[],"class_list":["post-131","post","type-post","status-publish","format-standard","hentry","category-chapter"],"_links":{"self":[{"href":"https:\/\/intelligenceexplosion.com\/sv\/wp-json\/wp\/v2\/posts\/131","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/intelligenceexplosion.com\/sv\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/intelligenceexplosion.com\/sv\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/intelligenceexplosion.com\/sv\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/intelligenceexplosion.com\/sv\/wp-json\/wp\/v2\/comments?post=131"}],"version-history":[{"count":0,"href":"https:\/\/intelligenceexplosion.com\/sv\/wp-json\/wp\/v2\/posts\/131\/revisions"}],"wp:attachment":[{"href":"https:\/\/intelligenceexplosion.com\/sv\/wp-json\/wp\/v2\/media?parent=131"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/intelligenceexplosion.com\/sv\/wp-json\/wp\/v2\/categories?post=131"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/intelligenceexplosion.com\/sv\/wp-json\/wp\/v2\/tags?post=131"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}