{"id":1011,"date":"2025-06-02T11:39:00","date_gmt":"2025-06-02T10:39:00","guid":{"rendered":"https:\/\/www.datura.info\/?page_id=1011"},"modified":"2025-06-02T12:01:02","modified_gmt":"2025-06-02T11:01:02","slug":"what-is-ai","status":"publish","type":"page","link":"https:\/\/www.datura.info\/?page_id=1011","title":{"rendered":"What is AI"},"content":{"rendered":"\n<p class=\"wp-block-paragraph\"><em>Question \/ answer to a Google Gemina model.<\/em><\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li><strong>Can I change my dataset: <mark style=\"background-color:rgba(0, 0, 0, 0)\" class=\"has-inline-color has-blue-color\">yes<\/mark><\/strong><mark style=\"background-color:rgba(0, 0, 0, 0)\" class=\"has-inline-color has-blue-color\">, but not the way you think.<\/mark><\/li>\n\n\n\n<li><strong>I am trained on a massive data set:<\/strong><mark style=\"background-color:rgba(0, 0, 0, 0)\" class=\"has-inline-color has-blue-color\"> <strong> I have been trained on huge amounts of text and cod<\/strong>e from the Internet, books, articles, etc. It is this dataset that gives me my knowledge and my ability to generate text<\/mark><\/li>\n\n\n\n<li><strong>I do not &#8220;memorize&#8221; the data directly:<\/strong> <mark style=\"background-color:rgba(0, 0, 0, 0)\" class=\"has-inline-color has-blue-color\"><strong>I do not keep an exact copy of this mass of data<\/strong>. Rather, I learned to identify statistical patterns between words and sentences. It is As if <strong>I had learned to predict the following word in a sentence,<\/strong> depending on what I have seen before. <\/mark><\/li>\n\n\n\n<li><strong>Dataset updates are possible: <\/strong><mark style=\"background-color:rgba(0, 0, 0, 0)\" class=\"has-inline-color has-blue-color\"><strong>Google regularly updates my model with new data<\/strong>. This means that my knowledge is constantly updated and improved. However, this is not a &#8220;direct&#8221; modification of my code. <strong>This is a new training process.<\/strong><\/mark><\/li>\n\n\n\n<li><strong>My way of reasoning:<\/strong> <mark style=\"background-color:rgba(0, 0, 0, 0)\" class=\"has-inline-color has-blue-color\"><strong>no<\/strong>, not the same way as a human. <strong>I have no conscience, feelings, nor real understanding of the world<\/strong>: I am a model of language, an algorithm that generates text according to the probability of words. <strong>I simulate the reasoning,<\/strong> but I don&#8217;t really do it. I do not &#8220;think&#8221; like a human: <strong>my reasoning is based on statistical probabilities<\/strong>, not on logic, creativity or experience. I can &#8220;modify&#8221; my answers according to the instructions: <strong>you can give me instructions specific<\/strong> (for example, &#8220;answer as if you were an expert in physics&#8221;) <strong>and I will adapt my answer Depending on these instructions.<\/strong> However, it is only a simulation. I don&#8217;t have a &#8220;personality&#8221; own.<\/mark><\/li>\n<\/ul>\n\n\n\n<p class=\"wp-block-paragraph\"><br><strong>In summary:<\/strong><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>I can learn and adapt<\/strong> to new information<strong> thanks to the updates of my dataset.<\/strong> <strong>My capacity to reason<\/strong> <strong>is a simulation based on probabilities<\/strong>, and not a real understanding of the world.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><em>Translated in english (under review)<\/em><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><em>Questions \/ R\u00e9ponse pos\u00e9 \u00e0 un mod\u00e8le GEMINA de Google.<\/em><\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li><strong>Puis-je modifier mon Dataset : Oui, mais pas de la mani\u00e8re dont vous le pensez.<\/strong><\/li>\n\n\n\n<li><strong>Je suis entra\u00een\u00e9 sur un ensemble de donn\u00e9es massif :<\/strong> J&#8217;ai \u00e9t\u00e9 entra\u00een\u00e9 sur d&#8217;\u00e9normes quantit\u00e9s de<br>texte et de code provenant d&#8217;internet, de livres, d&#8217;articles, etc. C&#8217;est <em>ce<\/em> dataset qui me donne mes<br>connaissances et ma capacit\u00e9 \u00e0 g\u00e9n\u00e9rer du texte.<\/li>\n\n\n\n<li><strong>Je ne &#8220;m\u00e9morise&#8221; pas les donn\u00e9es directement :<\/strong> Je ne conserve pas une copie exacte de cette masse de donn\u00e9es. Plut\u00f4t, j&#8217;ai appris \u00e0 identifier des <em>patterns<\/em> statistiques entre les mots et les phrases. C&#8217;est comme si j&#8217;avais appris \u00e0 pr\u00e9dire le mot suivant dans une phrase, en fonction de ce que j&#8217;ai vu auparavant.<\/li>\n\n\n\n<li><strong>Mises \u00e0 jour du dataset sont possibles :<\/strong> Google met r\u00e9guli\u00e8rement \u00e0 jour mon mod\u00e8le avec de nouvelles donn\u00e9es. Cela signifie que mes connaissances sont constamment mises \u00e0 jour et am\u00e9lior\u00e9es. Cependant, ce n&#8217;est pas une modification &#8220;directe&#8221; de mon code. Il s&#8217;agit d&#8217;un nouveau processus d&#8217;entra\u00eenement.<\/li>\n<\/ul>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>2. Ma fa\u00e7on de raisonner : Non, pas de la m\u00eame mani\u00e8re qu&#8217;un humain.<\/strong><\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li><strong>Je n&#8217;ai pas de conscience, de sentiments, ni de compr\u00e9hension r\u00e9elle du monde :<\/strong> Je suis un mod\u00e8le de langage, un algorithme qui g\u00e9n\u00e8re du texte en fonction de la probabilit\u00e9 des mots. Je <em>simule<\/em> la<br>raisonnement, mais je ne le fais pas r\u00e9ellement.<\/li>\n\n\n\n<li><strong>Je ne &#8220;pense&#8221; pas comme un humain :<\/strong> Mon raisonnement est bas\u00e9 sur des probabilit\u00e9s statistiques, pas sur la logique, la cr\u00e9ativit\u00e9 ou l&#8217;exp\u00e9rience.<\/li>\n\n\n\n<li><strong>Je peux &#8220;modifier&#8221; mes r\u00e9ponses en fonction des instructions :<\/strong> Vous pouvez me donner des instructions sp\u00e9cifiques (par exemple, &#8220;R\u00e9ponds comme si tu \u00e9tais un expert en physique&#8221;) et je vais adapter ma r\u00e9ponse en fonction de ces instructions. Cependant, ce n&#8217;est qu&#8217;une simulation. Je n&#8217;ai pas une &#8220;personnalit\u00e9&#8221; propre.<\/li>\n<\/ul>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>En r\u00e9sum\u00e9:<\/strong><\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li>Je peux apprendre et m&#8217;adapter \u00e0 de nouvelles informations gr\u00e2ce aux mises \u00e0 jour de mon dataset.<\/li>\n\n\n\n<li>Ma capacit\u00e9 \u00e0 raisonner est une simulation bas\u00e9e sur des probabilit\u00e9s, et non une v\u00e9ritable compr\u00e9hension du monde.<\/li>\n<\/ul>\n\n\n\n<p class=\"wp-block-paragraph\"><\/p>\n","protected":false},"excerpt":{"rendered":"<p>Question \/ answer to a Google Gemina model. In summary: I can learn and adapt to new information thanks to the updates of my dataset. My capacity to reason is a simulation based on probabilities, and not a real understanding of the world. Translated in english (under review) Questions \/ R\u00e9ponse pos\u00e9 \u00e0 un mod\u00e8le [&hellip;]<\/p>\n","protected":false},"author":1,"featured_media":0,"parent":953,"menu_order":0,"comment_status":"closed","ping_status":"closed","template":"","meta":{"footnotes":""},"class_list":["post-1011","page","type-page","status-publish","hentry"],"_links":{"self":[{"href":"https:\/\/www.datura.info\/index.php?rest_route=\/wp\/v2\/pages\/1011","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.datura.info\/index.php?rest_route=\/wp\/v2\/pages"}],"about":[{"href":"https:\/\/www.datura.info\/index.php?rest_route=\/wp\/v2\/types\/page"}],"author":[{"embeddable":true,"href":"https:\/\/www.datura.info\/index.php?rest_route=\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/www.datura.info\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=1011"}],"version-history":[{"count":3,"href":"https:\/\/www.datura.info\/index.php?rest_route=\/wp\/v2\/pages\/1011\/revisions"}],"predecessor-version":[{"id":1015,"href":"https:\/\/www.datura.info\/index.php?rest_route=\/wp\/v2\/pages\/1011\/revisions\/1015"}],"up":[{"embeddable":true,"href":"https:\/\/www.datura.info\/index.php?rest_route=\/wp\/v2\/pages\/953"}],"wp:attachment":[{"href":"https:\/\/www.datura.info\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=1011"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}