<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN"><html><head><meta content="text/html;charset=UTF-8" http-equiv="Content-Type"></head><body ><div style="font-family: Verdana, Arial, Helvetica, sans-serif; font-size: 10pt;"><p style="line-height:1.2;margin-right: 9pt;margin-top:5pt;margin-bottom:9pt;" dir="ltr"><span class="highlight" style="background-color:transparent"><span class="colour" style="color:rgb(0, 0, 0)"><span class="font" style="font-family:Verdana"><span class="size" style="font-size: 12pt; font-weight: 400; font-style: normal; font-variant: normal; text-decoration: none; vertical-align: baseline; white-space: pre-wrap;">Hi All,</span></span></span></span><br></p><p style="line-height:1.44;margin-right: 9pt;margin-top:5pt;margin-bottom:9pt;" dir="ltr"><span style="font-size:12pt;font-family:Verdana;color:#000000;background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline;white-space:pre;white-space:pre-wrap;">Good discussion I wish I had more time for, but a few replies to various points made here:</span><br></p><ul style="margin-top:0;margin-bottom:0;padding-inline-start:48px;"><li aria-level="1" style="list-style-type:disc;font-size:12pt;font-family:Verdana;color:#000000;background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline;white-space:pre;" dir="ltr"><p role="presentation" style="line-height: 1.2; margin-top: 0px; margin-bottom: 0pt;" dir="ltr"><span style="font-size:12pt;font-family:Verdana;color:#000000;background-color:transparent;font-weight:700;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline;white-space:pre;white-space:pre-wrap;">Shell summarizes well:</span><span style="font-size:12pt;font-family:Verdana;color:#000000;background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline;white-space:pre;white-space:pre-wrap;"> “But I do raise questions about what it all means for the future of work, careers, creativity, etc”. Also true, “a new wave of creativity” can happen.</span><br></p></li><li aria-level="1" style="list-style-type:disc;font-size:12pt;font-family:Verdana;color:#000000;background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline;white-space:pre;" dir="ltr"><p role="presentation" style="line-height: 1.2; margin-top: 0px; margin-bottom: 0pt;" dir="ltr"><span style="font-size:12pt;font-family:Verdana;color:#000000;background-color:transparent;font-weight:700;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline;white-space:pre;white-space:pre-wrap;">Chatbot-written platforms long preceded gpt, so one can find vastly inferior website content now</span><span style="font-size:12pt;font-family:Verdana;color:#000000;background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline;white-space:pre;white-space:pre-wrap;"> from those now instantly antiquated platforms. </span><br></p></li><li aria-level="1" style="list-style-type:disc;font-size:12pt;font-family:Verdana;color:#000000;background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline;white-space:pre;" dir="ltr"><p role="presentation" style="line-height: 1.2; margin-top: 0px; margin-bottom: 0pt;" dir="ltr"><span style="font-size:12pt;font-family:Verdana;color:#000000;background-color:transparent;font-weight:700;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline;white-space:pre;white-space:pre-wrap;">Fact-checking and editing must happen, but that’s less time</span><span style="font-size:12pt;font-family:Verdana;color:#000000;background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline;white-space:pre;white-space:pre-wrap;"> than composing all, I think.</span><br></p></li><li aria-level="1" style="list-style-type:disc;font-size:12pt;font-family:Verdana;color:#000000;background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline;white-space:pre;" dir="ltr"><p role="presentation" style="line-height: 1.2; margin-top: 0px; margin-bottom: 0pt;" dir="ltr"><span style="font-size:12pt;font-family:Verdana;color:#000000;background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline;white-space:pre;white-space:pre-wrap;">I’ve not tested gpt’s writing extensively, but </span><span style="font-size:12pt;font-family:Verdana;color:#000000;background-color:transparent;font-weight:700;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline;white-space:pre;white-space:pre-wrap;">that legal letter I mentioned was truly consummate</span><span style="font-size:12pt;font-family:Verdana;color:#000000;background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline;white-space:pre;white-space:pre-wrap;">, and will be used without a single word changed. I have noticed redundancy, though usually resulting from how I asked a question. </span><br></p></li><ul style="margin-top:0;margin-bottom:0;padding-inline-start:48px;"><li aria-level="2" style="list-style-type:circle;font-size:12pt;font-family:Verdana;color:#000000;background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline;white-space:pre;" dir="ltr"><p role="presentation" style="line-height: 1.2; margin-top: 0px; margin-bottom: 0pt;" dir="ltr"><span style="font-size:12pt;font-family:Verdana;color:#000000;background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline;white-space:pre;white-space:pre-wrap;">Note that one way gpt and the like is and will be trained to improve is via sophisticated AI-human systems by which to detect chatbot text, making the writing output constantly better.</span><br></p></li></ul><li aria-level="1" style="list-style-type:disc;font-size:12pt;font-family:Verdana;color:#000000;background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline;white-space:pre;" dir="ltr"><p role="presentation" style="line-height: 1.2; margin-top: 0px; margin-bottom: 0pt;" dir="ltr"><span style="font-size:12pt;font-family:Verdana;color:#000000;background-color:transparent;font-weight:700;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline;white-space:pre;white-space:pre-wrap;">Regarding SEO</span><span style="font-size:12pt;font-family:Verdana;color:#000000;background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline;white-space:pre;white-space:pre-wrap;">, a ton to say, but for now:</span><br></p></li><ul style="margin-top:0;margin-bottom:0;padding-inline-start:48px;"><li aria-level="2" style="list-style-type:circle;font-size:12pt;font-family:Verdana;color:#000000;background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline;white-space:pre;" dir="ltr"><p role="presentation" style="line-height: 1.2; margin-top: 0px; margin-bottom: 0pt;" dir="ltr"><span style="font-size:12pt;font-family:Verdana;color:#000000;background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline;white-space:pre;white-space:pre-wrap;">gpt <b>expedites</b> all kinds of SEO work. </span><br></p></li><li aria-level="2" style="list-style-type:circle;font-size:12pt;font-family:Verdana;color:#000000;background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline;white-space:pre;" dir="ltr"><p role="presentation" style="line-height: 1.2; margin-top: 0px; margin-bottom: 0pt;" dir="ltr"><span style="font-size:12pt;font-family:Verdana;color:#000000;background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline;white-space:pre;white-space:pre-wrap;">My firm has long prepared for AI, esp last 2 years: </span><br></p></li><ul style="margin-top:0;margin-bottom:0;padding-inline-start:48px;"><li aria-level="3" style="list-style-type:square;font-size:12pt;font-family:Verdana;color:#000000;background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline;white-space:pre;" dir="ltr"><p role="presentation" style="line-height: 1.2; margin-top: 0px; margin-bottom: 0pt;" dir="ltr"><span style="font-size:12pt;font-family:Verdana;color:#000000;background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline;white-space:pre;white-space:pre-wrap;">It’s about “Things not stings” (entities or concepts, not text strings). But yes, keyword research, a model still prevailing in most software that can't change cost-effectively now, will be expedited by gpt. My firm is currently researching to what extent old-fashioned strings correlate to the replacement: Entities within the </span><a style="text-decoration:none;" href="https://searchengineland.com/google-mum-update-seo-future-383551" target="_blank"><span style="font-size:12pt;font-family:Verdana;color:#1155cc;background-color:transparent;font-weight:700;font-style:normal;font-variant:normal;text-decoration:underline;-webkit-text-decoration-skip:none;text-decoration-skip-ink:none;vertical-align:baseline;white-space:pre;white-space:pre-wrap;">MUM</span></a><span style="font-size:12pt;font-family:Verdana;color:#000000;background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline;white-space:pre;white-space:pre-wrap;"> revolution.</span><br></p></li><li aria-level="3" style="list-style-type:square;font-size:12pt;font-family:Verdana;color:#000000;background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline;white-space:pre;" dir="ltr"><p role="presentation" style="line-height: 1.2; margin-top: 0px; margin-bottom: 0pt;" dir="ltr"><span style="font-size:12pt;font-family:Verdana;color:#000000;background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline;white-space:pre;white-space:pre-wrap;">Web marketing pros must coordinate marketing channels efficiently; </span><span style="font-size:12pt;font-family:Verdana;color:#000000;background-color:transparent;font-weight:700;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline;white-space:pre;white-space:pre-wrap;">AI is not good at dovetailing separate silos of knowledge</span><span style="font-size:12pt;font-family:Verdana;color:#000000;background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline;white-space:pre;white-space:pre-wrap;">, like SEO, PPC, client profit logic, etc, <span style="font-size:12pt;font-family:Verdana;color:#000000;background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline;white-space:pre;white-space:pre-wrap;">and it's terrible at prioritizing (within a firm's) context</span>. Hence Shell’s “creativity.”</span><br></p></li><li aria-level="3" style="list-style-type:square;font-size:12pt;font-family:Verdana;color:#000000;background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline;white-space:pre;" dir="ltr"><p role="presentation" style="line-height: 1.2; margin-top: 0px; margin-bottom: 0pt;" dir="ltr"><span style="font-size:12pt;font-family:Verdana;color:#000000;background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline;white-space:pre;white-space:pre-wrap;">My published prediction 10 years ago that by now AI would relieve SEO pros of lots of routine tech work (via Google Search Console) has been proven way too optimistic--we’re still far from that fine day unfortunately. </span><br></p></li></ul><li aria-level="2" style="list-style-type:circle;font-size:12pt;font-family:Verdana;color:#000000;background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline;white-space:pre;" dir="ltr"><p role="presentation" style="line-height: 1.2; margin-top: 0px; margin-bottom: 0pt;" dir="ltr"><span style="font-size:12pt;font-family:Verdana;color:#000000;background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline;white-space:pre;white-space:pre-wrap;">Extrapolating from the SEO remarks, I believe </span><span style="font-size:12pt;font-family:Verdana;color:#000000;background-color:transparent;font-weight:700;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline;white-space:pre;white-space:pre-wrap;">professionals in most fields will have to stay abreast of adjacent fields of knowledge</span><span style="font-size:12pt;font-family:Verdana;color:#000000;background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline;white-space:pre;white-space:pre-wrap;">, so as to be able to efficiently coordinate riders of separate AI horses. Broadly contextual solutions outside single domains of knowledge is something I predict AI will remain deficient in for several years. This goes to Shell’s “new wave of creativity.”</span><br></p></li></ul></ul><p style="line-height:1.2;margin-right: 9pt;margin-top:5pt;margin-bottom:9pt;" dir="ltr"><br></p><div id="Zm-_Id_-Sgn" data-zbluepencil-ignore="true" data-sigid="1535369000000073001"><div style="font-family: Verdana; font-size: 12px; line-height: normal; margin: 0px"><span class="size" style="font-size:13px"><span class="font" style="font-family:verdana, sans-serif">Best Regards,<br></span></span></div><div style="font-family: Verdana; font-size: 12px; line-height: normal; margin: 0px; min-height: 15px"><span class="size" style="font-size:13px"><span class="font" style="font-family:verdana, sans-serif"><br></span></span></div><div style="font-family: Verdana; font-size: 12px; line-height: normal; margin: 0px; color: rgb(170, 48, 37)"><b><span class="size" style="font-size:13px"><span class="font" style="font-family:verdana, sans-serif">Rob Laporte</span></span></b><span class="size" style="font-size:13px"><span class="font" style="font-family:verdana, sans-serif"><br></span></span></div><div style="font-family: Verdana; font-size: 12px; line-height: normal; margin: 0px"><span class="size" style="font-size:13px"><span class="font" style="font-family:verdana, sans-serif">CEO | R&D Manager</span><span class="font" style="font-family:verdana, sans-serif"><br></span></span></div><div style="font-family: Verdana; font-size: 12px; line-height: normal; margin: 0px"><b><span class="size" style="font-size:13px"><span class="font" style="font-family:verdana, sans-serif">DISC - Making Web Sites Make Money</span></span></b><span class="size" style="font-size:13px"><span class="font" style="font-family:verdana, sans-serif"><br><a target="_blank" href="mailto:Rob@2disc.com">Rob@2disc.com</a>, <span class="size" style="font-size:13px"><span class="font" style="font-family:verdana, sans-serif">413-584-6500</span></span></span></span><br></div><div style="font-family: Verdana; font-size: 12px; line-height: normal; margin: 0px"><span class="size" style="font-size:13px"><span class="font" style="font-family:verdana, sans-serif"><a href="https://www.2disc.com" target="_blank">www.2disc.com</a></span></span><span class="size" style="font-size:16px"><span class="font" style="font-family:arial, helvetica, sans-serif, sans-serif"></span></span><br></div><div style="font-family: Verdana; font-size: 12px; line-height: normal"><span class="size" style="font-size:16px"><span class="font" style="font-family:arial, helvetica, sans-serif, sans-serif"></span></span><br></div><div><b style="font-family: Verdana; font-size: 12px; line-height: normal"><span class="size" style="font-size:16px"><span class="font" style="font-family:arial, helvetica, sans-serif, sans-serif"><span class="colour" style="color:rgb(153, 0, 0)">NOTE:</span></span></span></b><span class="size" style="font-size:16px"><span class="font" style="font-family:arial, helvetica, sans-serif, sans-serif"><span class="colour" style="color:rgb(153, 0, 0)"> Emails can be blocked by spam filters throughout the web. If you don’t get a reply within an expected span of time, please call.</span></span></span><br></div></div><div><br></div><div class="zmail_extra_hr" style="border-top: 1px solid rgb(204, 204, 204); height: 0px; margin-top: 10px; margin-bottom: 10px; line-height: 0px;"><br></div><div class="zmail_extra" data-zbluepencil-ignore="true"><div><br></div><div id="Zm-_Id_-Sgn1">---- On Sun, 12 Mar 2023 11:57:09 -0400 <b>Shel Horowitz via Hidden-discuss <hidden-discuss@lists.hidden-tech.net></b> wrote ---<br></div><div><br></div><blockquote id="blockquote_zmail" style="margin: 0px;"><div><div dir="ltr">I agree with Marcia. ChatGPT is to a skilled writer as clip-art is to graphics. It's adequate for basic needs but is no substitute for a designer working specifically to convey your idea.<div><br></div><div>I recently landed on a site that I could tell used a chatbot to write its copy. I was repelled and left quickly--and crossed that consultant off my list of people I'd ever hire, because she didn't disclose that a bot wrote the site. Ugh!<br><div><div dir="ltr" class="x_-590120120gmail_signature"><div dir="ltr"><div><div dir="ltr"><div dir="ltr"><div dir="ltr"><div style="font-size : small;"><br></div><div><br></div><div>Shel Horowitz - "The Transformpreneur"<br></div><div>________________________________________________<br></div><div>Contact me to bake in profitability while addressing hunger, <br></div><div>poverty, war, and catastrophic climate change<br></div><div>* First business ever to be Green America Gold Certified<br></div><div>* Inducted into the National Environmental Hall of Fame<br></div><div>* Certified speaker: International Platform Association<br></div><div><a href="https://goingbeyondsustainability.com" target="_blank">https://goingbeyondsustainability.com</a> <br></div><div>mailto:<a href="mailto:shel@greenandprofitable.com" target="_blank">shel@greenandprofitable.com</a> 413-586-2388<br></div><div>Award-winning, best-selling author of 10 books. <br></div><div>Latest: Guerrilla Marketing to Heal the World <br></div><div>(co-authored with Jay Conrad Levinson)<br></div><div><br></div><div>Watch my TEDx Talk,<br></div><div>"Impossible is a Dare: Business for a Better World"<br></div><div><a href="http://www.ted.com/tedx/events/11809" target="_blank">http://www.ted.com/tedx/events/11809</a> <br></div><div>(move your mouse to "event videos")<br></div><div>_________________________________________________<br></div><div><br></div></div></div></div></div></div></div></div><br></div></div><br><div class="x_-590120120gmail_quote"><div dir="ltr" class="x_-590120120gmail_attr">On Sat, Mar 11, 2023 at 9:45 PM Marcia Yudkin via Hidden-discuss <<a href="mailto:hidden-discuss@lists.hidden-tech.net" target="_blank">hidden-discuss@lists.hidden-tech.net</a>> wrote:<br></div><div id="zmail_block"><br></div></div>_______________________________________________<br>Hidden-discuss mailing list - home page: <a href="http://www.hidden-tech.net" target="_blank">http://www.hidden-tech.net</a> <br><a href="mailto:Hidden-discuss@lists.hidden-tech.net" target="_blank">Hidden-discuss@lists.hidden-tech.net</a> <br> <br>You are receiving this because you are on the Hidden-Tech Discussion list. <br>If you would like to change your list preferences, Go to the Members <br>page on the Hidden Tech Web site. <br><a href="http://www.hidden-tech.net/members" target="_blank">http://www.hidden-tech.net/members</a> <br></div><blockquote class="x_-590120120gmail_quote" style="margin : 0px 0px 0px 0.8ex; border-left : 1px solid rgb(204,204,204); padding-left : 1ex;">I'm not sure I would call its output "well-written," though, It's competently written, well-organized and with proper grammar, but stodgy in style and unimaginative, unless you've given it instructions like "write in the style of ____."<br> <br> <br> <br> <br> <br> <br> On Saturday, March 11, 2023 at 04:22:19 PM HST, Alan Frank <<a href="mailto:alan@8wheels.org" target="_blank">alan@8wheels.org</a>> wrote: <br> <br> <br> <br> <br> <br> I would use it for composing text, but not for facts at all. And if I <br> asked it "How would you introduce George Takei at a political <br> convention"," I would expect well-written text, but would also <br> fact-check everything.<br> <br> <br> -------- Original Message --------<br> Subject: Re: [Hidden-tech] Question about ChatGPT and machine learning<br> Date: 2023-03-11 15:55<br> From: Marcia Yudkin via Hidden-discuss <br> <<a href="mailto:hidden-discuss@lists.hidden-tech.net" target="_blank">hidden-discuss@lists.hidden-tech.net</a>><br> To: Rob Laporte <<a href="mailto:rob@2disc.com" target="_blank">rob@2disc.com</a>><br> <br> Rob,<br> <br> Those are very interesting examples. It's a mixed track record, <br> though. Based on your experience, what would you say ChatGPT should and <br> shouldn't be used for, or how it should or shouldn't be used?<br> <br> For example, based on the errors in bios you saw, would you still use it <br> for those artist bios given that you'd have to meticulously fact check <br> everything it wrote?<br> <br> Marcia Yudkin<br> <br> <br> <br> <br> <br> <br> On Saturday, March 11, 2023 at 10:19:20 AM HST, Rob Laporte <br> <<a href="mailto:rob@2disc.com" target="_blank">rob@2disc.com</a>> wrote:<br> <br> <br> <br> <br> <br> I've noticed that humans simplify complex, especially threatening, new <br> things, by using dichotomies of good and evil, red and blue, etc, and <br> that conviction is often inversely proportionate to knowledge. I've <br> worked in search marketing for 28 years, have digested a lot about the <br> kind of tech and processes underlying chatGPT (gpt), and I have no sound <br> basis for strong conviction on any of the issues broached here. But this <br> I can tell you from my mere 6 hours novice use so far:<br> <br> * In under an hour it solved a complex HR-financial legal question, <br> and provided the letter the plaintiff should write to the corporation's <br> HR department, with quality of writing, sensitivity to workplace <br> politics, and effective brevity way better than anything I can recall in <br> my 50 years of adulthood, decade teaching college lit and writing, and 3 <br> decades in search marketing. Truly stunning. Save at least $2000 in <br> legal fees that might have gone to a local law firm.<br> * A few times over the years I researched best email spam blocking <br> solutions, and considered an aggressive form of it. gpt explained the <br> problem with that solution, and did so way faster than my past and still <br> inclusive Google searches, saving me a few hundred dollars in IT <br> consulting.<br> * It completely conflated my semi-famous lawyer grandad's bio with <br> that of his stellar but less accomplished son of the same name. Both are <br> years deceased (most gpt data ends Sept '21), yet totally wrong.<br> * So too it got the bio of a decade-deceased famous scholar of Roman <br> architecture (friend's dad) wrong on a few points, most notably that <br> most his career was at Smith college, not Princeton as gpt said. <br> * It produced strikingly eloquent spreadsheet solutions for two <br> different complex purposes. I asked it for the actual spreadsheet, and <br> cell references were off, but in a second half hour of work, I'm sure it <br> wold have gotten it right or I could correct it myself. A few hours of <br> work time saved there, and one of the two tasks was billable. <br> * My firm had a prospective writing project for a client, involving <br> bios of famous and notable painters sold by the client. I say "had" <br> because now gpt or its structuring within services like NeuroFash will <br> cut both client and my firm's copywriter time substantially. <br> * I've not tried but viewed a YouTube of good, arguably very good, <br> graphic design for a marketing campaign done in well under half a day.<br> Outside of broad ideological judgements, there's much to consider in how <br> gpt will change work and incomes. <br> <br> The current version 3.5 will be upgraded to 4.0 within weeks. Think of <br> gpt like the web in 1994 or personal PCs in 1981, with advancements <br> happening 10x faster. <br> <br> Best Regards,<br> <br> Rob Laporte<br> CEO | R&D Manager<br> DISC - Making Web Sites Make Money<br> <a href="mailto:Rob@2disc.com" target="_blank">Rob@2disc.com</a>, 413-584-6500<br> <a href="http://www.2disc.com" target="_blank">www.2disc.com</a><br> <br> NOTE: Emails can be blocked by spam filters throughout the web. If you <br> don’t get a reply within an expected span of time, please call.<br> <br> <br> <br> ---- On Fri, 10 Mar 2023 20:36:26 -0500 Marcia Yudkin via Hidden-discuss <br> <<a href="mailto:hidden-discuss@lists.hidden-tech.net" target="_blank">hidden-discuss@lists.hidden-tech.net</a>> wrote ---<br> <br> > David,<br> > <br> > Some great points there. I especially like this one:<br> > <br> >>> it is *ALL* made up.<<<br> > <br> > That helps me to dimly understand that everything the chat says is <br> > simply plausible, no more than that.<br> > <br> > Maybe we should think of it as no more authoritative than the cocktail <br> > party chatter of someone who reads indiscriminately and can't shut up <br> > until they've spewed five paragraphs.<br> > <br> > Marcia<br> > <br> > <br> > <br> > <br> > <br> > <br> > <br> > <br> > On Friday, March 10, 2023 at 12:34:46 PM HST, R. David Murray via <br> > Hidden-discuss <<a href="mailto:hidden-discuss@lists.hidden-tech.net" target="_blank">hidden-discuss@lists.hidden-tech.net</a>> wrote:<br> > <br> > <br> > <br> > <br> > <br> > From what I understand (admittedly from only a *basic* understanding of<br> > machine learning), it is not so much that ChatGPT is "making errors",<br> > but rather that it is "making stuff up", and does not admit that it is<br> > making stuff up.<br> > <br> > I'm going to brain dump what I think here, but I'm not an expert in <br> > this<br> > by any stretch, so don't take me as an authority. Perhaps this can <br> > help<br> > you reason about ChartGPT until you find a better expert to consult ;)<br> > <br> > One thing to understand is that this is a *trained* model. That means<br> > that it was given a set of questions and answers and told "these are<br> > good, these are bad", probably with a rating of *how* good or bad. <br> > Then<br> > it was given a lot of other data (and how exactly this gets turned into<br> > questions and answers is *way* beyond my knowledge level). Then a team<br> > of model trainers started asking questions. The trainers would look at<br> > the answers it came up with and rate them, thus adding to the "trained"<br> > data set. When you tell ChatGPT that its answer was good or bad, you<br> > are also potentially adding to that training data, by the way.<br> > <br> > I'm guessing that the way the system works there is actually no way for<br> > it to "know" that it has made something up. The output that it <br> > produces<br> > is generated based on what you can think of as a very advanced version<br> > of statistical language modelling: given a certain input, what are the<br> > most likely kinds of things that would follow as a response? And like<br> > any statistical model, when you get enough standard deviations out,<br> > things get weird. At no point in the model output are things tagged as<br> > "made up" or "not made up": it is *ALL* made up.<br> > <br> > In the middle of the bell curve the made up things are *much* more<br> > likely to be "correct" than out at the edges of the bell curve. But<br> > oh those edges...<br> > <br> > It is of course more sophisticated than a statistical model, but the<br> > same principle applies: if there are few examples of *exactly* the <br> > kind<br> > of data your input contains, then it is going to draw from stuff that <br> > is<br> > a lot less closely related to your input for its response. But, and<br> > here is the important part, it is going to make up *something* to <br> > answer<br> > with. If a source is mentioned multiple times in the context of your<br> > input, it will use it. If there are no sources mentioned in the <br> > context<br> > of your input, it will generate an output that looks like the *kind of<br> > thing* that would be a response to that *kind of input*. In this case<br> > that included a list of articles. It generated at least one of them<br> > from an author whose name was probably mentioned in the context of your<br> > input, but never with an actual article name attached. Or maybe that<br> > author was mentioned in the context of conversations containing a<br> > subset of the *words* in your input (rather than logically formed<br> > sentences), depending on just how fuzzy the match was. Then it<br> > effectively made up a plausible sounding article name to go with the<br> > author name, because that's what responses to other similar questions <br> > in<br> > its training data looked like (not similar in content, but similar in<br> > *form*).<br> > <br> > So while I agree that making up all the sources seems like an extreme<br> > example of this, ChatGPT is what Science Fiction calls an "Artificial<br> > Stupid" (something that can't actually *reason*), and thus I think my<br> > explanation is plausible. It just depends on how fuzzy the match was<br> > that it made on the input. If the match was very fuzzy, then it would<br> > have come back with material from its data that generally followed at<br> > least some of your input, and then since responses the trainers<br> > considered "good" to questions like that usually included some sources,<br> > it made some up based on how the answers to other, less related,<br> > questions looked.<br> > <br> > Anyone want to bet that four sources was the average number that was<br> > accepted as "a good answer" by the people who did the training? I know<br> > I've seen "four things" in a couple of ChatGPT answers, and I haven't<br> > asked it very many questions :)<br> > <br> > Given all this, there are only two things you can do, one of which is<br> > exactly what you did: ask it for the sources. Given *that* input, it<br> > should be able to come up with the most likely response being the <br> > actual<br> > source. If it can't, then it has probably made up the source (note: I<br> > have not tested this technique myself, but it follows logically from <br> > how<br> > I think the system works).<br> > <br> > The second thing you can do (which you probably also already did) is to<br> > rephrase your input, giving it different amounts and kinds of context,<br> > and see how the output changes. If your altered input results in a <br> > less<br> > fuzzy match, you will get better answers.<br> > <br> > The big takeaway, which you clearly already know, is to never trust<br> > anything ChatGPT produces. Use it as a rough draft, but verify all the<br> > facts.<br> > <br> > My fear is that there are going to be a lot of people who aren't as<br> > diligent, and we'll end up with a lot of made up information out on the<br> > web adding to all of the maliciously bad information that is already <br> > out<br> > there. I have read that the ChatGPT researchers are worried about how<br> > to avoid using ChatGPT's output as input to a later ChatGPT model, and <br> > I<br> > have no idea how they are going to achieve that!<br> > <br> > And keep in mind that that maliciously bad information *is part of<br> > ChatGPT's data set*. Some of it the people who did the training will <br> > have<br> > caught, but I'm willing to bet they missed a lot of it because *they*<br> > didn't know it was bad, or it never came up during training.<br> > <br> > --David<br> > <br> > On Fri, 10 Mar 2023 03:14:21 +0000, Marcia Yudkin via Hidden-discuss <br> > <<a href="mailto:hidden-discuss@lists.hidden-tech.net" target="_blank">hidden-discuss@lists.hidden-tech.net</a>> wrote:<br> >> Yes, I know that people have been pointing out "ridiculous factual <br> >> errors" from ChatGPT. However, to make up sources that sound <br> >> completely plausible but are fake seems like it belongs in a whole <br> >> other category.<br> >> <br> >> <br> >> <br> >> <br> >> <br> >> <br> >> On Thursday, March 9, 2023 at 04:10:43 PM HST, Alan Frank <br> >> <<a href="mailto:alan@8wheels.org" target="_blank">alan@8wheels.org</a>> wrote:<br> >> <br> >> <br> >> <br> >> <br> >> <br> >> ChatGPT is a conversation engine, not a search engine. It is designed<br> >> to provide plausible responses based on similarity of questions and<br> >> answers to existing material on the internet, without attempting to<br> >> correlate its responses with actual facts. Pretty much every social<br> >> media space I follow has had multiple posts from people pointing out<br> >> ridiculous factual errors from ChatGPT.<br> >> <br> >> --Alan<br> >> <br> >> <br> >> -------- Original Message --------<br> >> Subject: [Hidden-tech] Question about ChatGPT and machine learning<br> >> Date: 2023-03-09 15:29<br> >> From: Marcia Yudkin via Hidden-discuss<br> >> <<a href="mailto:hidden-discuss@lists.hidden-tech.net" target="_blank">hidden-discuss@lists.hidden-tech.net</a>><br> >> To: "<a href="mailto:Hidden-discuss@lists.hidden-tech.net" target="_blank">Hidden-discuss@lists.hidden-tech.net</a>"<br> >> <<a href="mailto:Hidden-discuss@lists.hidden-tech.net" target="_blank">Hidden-discuss@lists.hidden-tech.net</a>><br> >> <br> >> This question is for anyone who understands how the machine learning <br> >> in<br> >> ChatGPT works.<br> >> <br> >> I've been finding ChatGPT useful for summarizing information that is<br> >> widely dispersed around the web, such as questions like "what are the<br> >> most popular objections to X?" However, the other day for a blog post <br> >> I<br> >> was writing I asked it "What are some sources on the relationship of X<br> >> to Y?" It gave me four sources of information, including the article<br> >> title, where it was published and who wrote it. <br> >> <br> >> This looked great, especially since I recognized two of the author <br> >> names<br> >> as authorities on X. However, when I then did a Google search, I <br> >> could<br> >> not track down any of the four articles, either by title, author or<br> >> place of publication. I tried both in Google and in Bing. Zilch!<br> >> <br> >> Could ChatGPT have totally made up these sources? If so, how does <br> >> that<br> >> work?<br> >> <br> >> I am baffled about the explanation of this. One of the publications<br> >> involved was Psychology Today, so we are not talking about obscure<br> >> corners of the Internet or sites that would have disappeared recently.<br> >> <br> >> Thanks for any insights.<br> >> <br> >> Marcia Yudkin<br> >> Introvert UpThink<br> >> Introvert UpThink | Marcia Yudkin | Substack<br> >> <br> >> <br> >> <br> >> <br> >> <br> >> Introvert UpThink | Marcia Yudkin | Substack<br> >> Marcia Yudkin<br> >> Exploring how introverts are misunderstood, maligned and<br> >> underappreciated in our culture - yet still thrive. Cli...<br> >> <br> >> <br> >> _______________________________________________<br> >> Hidden-discuss mailing list - home page: <a href="http://www.hidden-tech.net" target="_blank">http://www.hidden-tech.net</a><br> >> <a href="mailto:Hidden-discuss@lists.hidden-tech.net" target="_blank">Hidden-discuss@lists.hidden-tech.net</a><br> >> <br> >> You are receiving this because you are on the Hidden-Tech Discussion<br> >> list.<br> >> If you would like to change your list preferences, Go to the Members<br> >> page on the Hidden Tech Web site.<br> >> <a href="http://www.hidden-tech.net/members" target="_blank">http://www.hidden-tech.net/members</a><br> >> _______________________________________________<br> >> Hidden-discuss mailing list - home page: <a href="http://www.hidden-tech.net" target="_blank">http://www.hidden-tech.net</a><br> >> <a href="mailto:Hidden-discuss@lists.hidden-tech.net" target="_blank">Hidden-discuss@lists.hidden-tech.net</a><br> >> <br> >> You are receiving this because you are on the Hidden-Tech Discussion <br> >> list.<br> >> If you would like to change your list preferences, Go to the Members<br> >> page on the Hidden Tech Web site.<br> >> <a href="http://www.hidden-tech.net/members" target="_blank">http://www.hidden-tech.net/members</a><br> > _______________________________________________<br> > Hidden-discuss mailing list - home page: <a href="http://www.hidden-tech.net" target="_blank">http://www.hidden-tech.net</a><br> > <a href="mailto:Hidden-discuss@lists.hidden-tech.net" target="_blank">Hidden-discuss@lists.hidden-tech.net</a><br> > <br> > You are receiving this because you are on the Hidden-Tech Discussion <br> > list.<br> > If you would like to change your list preferences, Go to the Members<br> > page on the Hidden Tech Web site.<br> > <a href="http://www.hidden-tech.net/members" target="_blank">http://www.hidden-tech.net/members</a><br> > _______________________________________________<br> > Hidden-discuss mailing list - home page: <a href="http://www.hidden-tech.net" target="_blank">http://www.hidden-tech.net</a><br> > <a href="mailto:Hidden-discuss@lists.hidden-tech.net" target="_blank">Hidden-discuss@lists.hidden-tech.net</a><br> > <br> > You are receiving this because you are on the Hidden-Tech Discussion <br> > list.<br> > If you would like to change your list preferences, Go to the Members<br> > page on the Hidden Tech Web site.<br> > <a href="http://www.hidden-tech.net/members" target="_blank">http://www.hidden-tech.net/members</a><br> <br> <br> _______________________________________________<br> Hidden-discuss mailing list - home page: <a href="http://www.hidden-tech.net" target="_blank">http://www.hidden-tech.net</a><br> <a href="mailto:Hidden-discuss@lists.hidden-tech.net" target="_blank">Hidden-discuss@lists.hidden-tech.net</a><br> <br> You are receiving this because you are on the Hidden-Tech Discussion <br> list.<br> If you would like to change your list preferences, Go to the Members<br> page on the Hidden Tech Web site.<br> <a href="http://www.hidden-tech.net/members" target="_blank">http://www.hidden-tech.net/members</a><br> _______________________________________________<br> Hidden-discuss mailing list - home page: <a href="http://www.hidden-tech.net" target="_blank">http://www.hidden-tech.net</a><br> <a href="mailto:Hidden-discuss@lists.hidden-tech.net" target="_blank">Hidden-discuss@lists.hidden-tech.net</a><br> <br> You are receiving this because you are on the Hidden-Tech Discussion list.<br> If you would like to change your list preferences, Go to the Members<br> page on the Hidden Tech Web site.<br> <a href="http://www.hidden-tech.net/members" target="_blank">http://www.hidden-tech.net/members</a><br></blockquote></blockquote></div><div><br></div></div><br></body></html>