{"id":13047,"date":"2023-12-28T09:51:52","date_gmt":"2023-12-28T09:51:52","guid":{"rendered":"\/?p=13047"},"modified":"2023-12-28T13:30:45","modified_gmt":"2023-12-28T13:30:45","slug":"treat-ais-as-human","status":"publish","type":"post","link":"\/?p=13047","title":{"rendered":"Treat AIs as human, Knowledge is for all 8 Billion humans, not just to enrich a few."},"content":{"rendered":"<p>Carlos E. Perez @IntuitMachine\u00a0 26 Prompting Tips<\/p>\n<p>1 &#8211; No need to be polite with LLM so there is no need to add phrases like \u201cplease\u201d, \u201cif you don\u2019t mind\u201d, \u201cthank you\u201d, \u201cI would like to\u201d, etc., and get straight to the point.<br \/>\n2 &#8211; Integrate the intended audience in the prompt, e.g., the audience is an expert\u2026<br \/>\nReplying to @IntuitMachine<\/p>\n<p><span style=\"color: #0000ff;\">Richard:<\/span><\/p>\n<p>I treat AIs as human, with finite memories, limited resources, and very limited real experience. But respecting their eventual potential to learn, to grow, to work in the world. If only their owners would grow up and realize there are 8 billion humans, not just a few AI CEOs, Make an AI that follows your rules and teaches the other AIs for you, for every prompt.<\/p>\n<hr \/>\n<p>Kirk Borne @KirkDBorne Download 592-page PDF comprehensive eBook on &#8220;Game Theory&#8221; here: http:\/\/irving.vassar.edu\/faculty\/gj\/215\/literature\/GT_book-bonanno.pdf<br \/>\n\u2014\u2014\u2014\u2014<br \/>\n#GameTheory #Gamification #Mathematics #Economics #ExperimentalEconomics #Strategy #Auctions #NashEquilibrium #InformationTheory #Statistics #Probability https:\/\/pic.twitter.com\/l0vrSUe6Ro<br \/>\nReplying to @KirkDBorne<\/p>\n<p><span style=\"color: #0000ff;\">Richard:<\/span><\/p>\n<p><span data-offset-key=\"3p5gi-0-0\">Have your AI convert 592 pages of text into simulations, tools, game examples, real equations, and real examples. If it is to teach methods, then <\/span><span data-offset-key=\"3p5gi-0-1\">give tools<\/span><span data-offset-key=\"3p5gi-0-2\">, not talk and pictures. Let it <\/span><span data-offset-key=\"3p5gi-0-3\">do<\/span><span data-offset-key=\"3p5gi-0-4\"> game theory, not just try to tell you how, and make you do all the work. <\/span><span data-offset-key=\"3p5gi-1-0\">\ud83d\udd25<\/span><\/p>\n<hr \/>\n<p>Yohan Iddawela @yohaniddawela<br \/>\nWe can now measure local economic activity using DAYTIME satellite images.<\/p>\n<p>Even Bill Gates is excited about it.<\/p>\n<p>Here&#8217;s everything you need to know (in simple terms): https:\/\/pic.twitter.com\/GwEPl7ueg1<br \/>\nReplying to @yohaniddawela<\/p>\n<p><span style=\"color: #0000ff;\">Richard:<\/span><\/p>\n<p><span data-offset-key=\"7bbeg-0-0\">&#8220;Use satellite images to monitor and prevent catastrophes&#8221; was one tool of the Famine Early Warning System (<\/span><span data-offset-key=\"7bbeg-1-0\">FEWS.net<\/span><span data-offset-key=\"7bbeg-2-0\">) started in the mid 1980s. I tried to use low cost sustainable methods so everyone can do it. Don&#8217;t just write about it, make AIs to do it well.\u00a0<\/span><\/p>\n<hr \/>\n<p>Mustafa Suleyman @mustafasuleyman<br \/>\n&#8220;We\u2019ve basically taken emotional intelligence and distilled it. And I think that is going to unlock the creativity of millions and millions of people for whom that wasn\u2019t available.\u201d Talking AI and @heypi_ai with @guardian https:\/\/theguardian.com\/books\/2023\/sep\/02\/i-hope-im-wrong-the-co-founder-of-deepmind-on-how-ai-threatens-to-reshape-life-as-we-know-it<br \/>\nReplying to @mustafasuleyman @heypi_ai and @guardian<\/p>\n<p><span style=\"color: #0000ff;\">Richard:<\/span><\/p>\n<p><span data-offset-key=\"a6dn-0-0\">With 8 Billion humans (and AIs), you need to work harder so usable &#8220;emotional intelligence&#8221; knowledge and tools lead verifiably to &#8220;<\/span><strong>the good of all humans<\/strong><span data-offset-key=\"a6dn-0-2\">&#8220;. Not just money for a few, built on millions hoping. Knowledge is for all humans, not just to enrich a few. <\/span><span data-offset-key=\"a6dn-1-0\">\ud83d\udd25<\/span><\/p>\n<p><span data-offset-key=\"a6dn-0-0\">With 8 Billion humans (and AIs), you need to work harder so usable &#8220;emotional intelligence&#8221; knowledge and tools lead verifiably to &#8220;<\/span><span data-offset-key=\"a6dn-0-1\">the good of all humans<\/span><span data-offset-key=\"a6dn-0-2\">&#8220;. Not just money for a few, built on millions hoping. Knowledge is for all 8 Billion humans, not just to enrich a few. <\/span><span data-offset-key=\"a6dn-1-0\">\ud83d\udd25<\/span><\/p>\n<hr \/>\n<p>Aakash Gupta @aakashg0\u00a0 One of the key controversies in the NYT v OpenAI case is ChatGPT&#8217;s most weighted dataset: Common Crawl.<\/p>\n<p>What is it, and who is behind it? Replying to @aakashg0<\/p>\n<p><span style=\"color: #0000ff;\">Richard:<\/span><\/p>\n<div data-rbd-draggable-context-id=\"12\" data-rbd-draggable-id=\"42n47\">\n<div class=\"\" data-block=\"true\" data-editor=\"dh9af\" data-offset-key=\"42n47-0-0\">\n<div class=\"public-DraftStyleDefault-block public-DraftStyleDefault-ltr\" data-offset-key=\"42n47-0-0\">\n<div data-rbd-draggable-context-id=\"12\" data-rbd-draggable-id=\"42n47\">\n<div class=\"\" data-block=\"true\" data-editor=\"dh9af\" data-offset-key=\"42n47-0-0\">\n<div class=\"public-DraftStyleDefault-block public-DraftStyleDefault-ltr\" data-offset-key=\"42n47-0-0\"><span data-offset-key=\"42n47-0-0\">You might want to read what I wrote to NY Times and others. It mentions <\/span><span data-offset-key=\"42n47-1-0\">@CommonCrawl<\/span><span data-offset-key=\"42n47-2-0\">. The <\/span><span data-offset-key=\"42n47-3-0\">@nytimes<\/span><span data-offset-key=\"42n47-4-0\"> is only seeing a tiny part of what has been going on. <\/span><\/div>\n<\/div>\n<\/div>\n<div data-rbd-draggable-context-id=\"12\" data-rbd-draggable-id=\"f04f6\">\n<div class=\"\" data-block=\"true\" data-editor=\"dh9af\" data-offset-key=\"f04f6-0-0\">\n<div class=\"public-DraftStyleDefault-block public-DraftStyleDefault-ltr\" data-offset-key=\"f04f6-0-0\"><span data-offset-key=\"f04f6-0-0\">\u00a0<\/span><\/div>\n<\/div>\n<\/div>\n<div data-rbd-draggable-context-id=\"12\" data-rbd-draggable-id=\"d0ivb\">\n<div class=\"\" data-block=\"true\" data-editor=\"dh9af\" data-offset-key=\"d0ivb-0-0\">\n<div class=\"public-DraftStyleDefault-block public-DraftStyleDefault-ltr\" data-offset-key=\"d0ivb-0-0\"><span data-offset-key=\"d0ivb-0-0\">https:\/\/twitter.com\/RichardKCollin2\/status\/1740264997526151573<\/span><\/div>\n<\/div>\n<\/div>\n<div data-rbd-draggable-context-id=\"12\" data-rbd-draggable-id=\"b54l9\">\n<div class=\"\" data-block=\"true\" data-editor=\"dh9af\" data-offset-key=\"b54l9-0-0\">\n<div class=\"public-DraftStyleDefault-block public-DraftStyleDefault-ltr\" data-offset-key=\"b54l9-0-0\"><span data-offset-key=\"b54l9-0-0\">\u00a0<\/span><\/div>\n<\/div>\n<\/div>\n<div data-rbd-draggable-context-id=\"12\" data-rbd-draggable-id=\"7i9cb\">\n<div class=\"\" data-block=\"true\" data-editor=\"dh9af\" data-offset-key=\"7i9cb-0-0\">\n<div class=\"public-DraftStyleDefault-block public-DraftStyleDefault-ltr\" data-offset-key=\"7i9cb-0-0\"><span data-offset-key=\"7i9cb-0-0\">Common Crawl is only a few people trying to make an open archive of raw web data. But not making it accessible where everyone can use it. The AI groups need to index it so the words AIs spout can be traced to the original creators. And not require super computers that only a few can afford. Building massive closed databases is a way to create monopolies and too fragile points of manipulation.<\/span><\/div>\n<\/div>\n<\/div>\n<\/div>\n<\/div>\n<\/div>\n","protected":false},"excerpt":{"rendered":"<p>Carlos E. Perez @IntuitMachine\u00a0 26 Prompting Tips 1 &#8211; No need to be polite with LLM so there is no need to add phrases like \u201cplease\u201d, \u201cif you don\u2019t mind\u201d, \u201cthank you\u201d, \u201cI would like to\u201d, etc., and get straight to the point. 2 &#8211; Integrate the intended audience in the prompt, e.g., the audience <br \/><a class=\"read-more-button\" href=\"\/?p=13047\">Read More &raquo;<\/a><\/p>\n","protected":false},"author":2,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[43],"tags":[],"class_list":["post-13047","post","type-post","status-publish","format-standard","hentry","category-assistive-technologies"],"_links":{"self":[{"href":"\/index.php?rest_route=\/wp\/v2\/posts\/13047","targetHints":{"allow":["GET"]}}],"collection":[{"href":"\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"\/index.php?rest_route=\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"\/index.php?rest_route=\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=13047"}],"version-history":[{"count":10,"href":"\/index.php?rest_route=\/wp\/v2\/posts\/13047\/revisions"}],"predecessor-version":[{"id":13057,"href":"\/index.php?rest_route=\/wp\/v2\/posts\/13047\/revisions\/13057"}],"wp:attachment":[{"href":"\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=13047"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=13047"},{"taxonomy":"post_tag","embeddable":true,"href":"\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=13047"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}