1import{r as m,c as i,i as e,a as g,b,s as p,t as d}from"./web-3lO_5tTB.js";var f=d('<section class="bg-white dark:bg-gray-900"><div class="py-8 px-4 mx-auto max-w-screen-xl lg:py-16 lg:px-6"><h2 class="text-4xl font-extrabold dark:text-white">Timeline</h2><p class="my-4 text-lg font-normal text-gray-500 dark:text-gray-400">Large Language Models have not been actively used by adtech companies besides some initial experiments. However, everyone sees the potential and we believe we will soon see a new kind of ads.</p><ol class="relative border-s border-gray-200 dark:border-gray-700"></ol><p class="my-4 text-sm font-normal italic text-gray-500 dark:text-gray-400">Sources for the timeline: <a class=text-blue-400 href=https://timelines.issarice.com/wiki/Timeline_of_ChatGPT>timelines.issarice.com</a>, <a class=text-blue-400 href=https://medium.com/@researchgraph/brief-introduction-to-the-history-of-large-language-models-llms-3c2efa517112>Research Graph</a>, <a class=text-blue-400 href=https://synthedia.substack.com/p/a-timeline-of-large-language-model>synthedia.substack.com'),y=d('<li class="mb-10 ms-4"><div class="absolute w-3 h-3 bg-gray-200 rounded-full mt-1.5 -start-1.5 border border-white dark:border-gray-900 dark:bg-gray-700"></div><time class="mb-1 text-sm font-normal leading-none text-gray-400 dark:text-gray-500"></time><h3 class="text-lg font-semibold text-gray-900 dark:text-white"></h3><p class="mb-4 text-base font-normal text-gray-500 dark:text-gray-400">'),x=d('<a target=_blank class="inline-flex items-center px-4 py-2 text-sm font-medium text-gray-900 bg-white border border-gray-200 rounded-lg hover:bg-gray-100 hover:text-blue-700 focus:z-10 focus:ring-4 focus:outline-none focus:ring-gray-100 focus:text-blue-700 dark:bg-gray-800 dark:text-gray-400 dark:border-gray-600 dark:hover:text-white dark:hover:bg-gray-700 dark:focus:ring-gray-700"><svg class="w-3 h-3 ms-2 rtl:rotate-180"aria-hidden=true xmlns=http://www.w3.org/2000/svg fill=none viewBox="0 0 14 10"><path stroke=currentColor stroke-linecap=round stroke-linejoin=round stroke-width=2 d="M1 5h12m0 0L9 1m4 4L9 9">');function w(){return(()=>{var r=f(),a=r.firstChild,l=a.firstChild,o=l.nextSibling,t=o.nextSibling;return e(t,i(n,{title:"ELIZA",time:"1966",description:"Created by Joseph Weizenbaum at MIT, ELIZA is considered to be the first chatbot ever built by humans.",buttonText:"Try it",buttonLink:"https://web.njit.edu/~ronkowit/eliza.html"}),null),e(t,i(n,{title:"SmarterChild: pioneering AI chatbot",time:"2001",description:"Early chatbot SmarterChild gains popularity among preteens for its witty responses on AIM and MSN. It is a precursor to modern AI bots like Siri and Alexa, even inspiring investors to fund Siri."}),null),e(t,i(n,{title:"Creation of OpenAI",time:"December 2015",description:"OpenAI is introduced as a non-profit organization focused on advancing digital intelligence in a way that benefits humanity as a whole."}),null),e(t,i(n,{title:"The invention of Transformers Architecture",time:"June 2017",description:"Transformers came out in 2017 with the paper âAttention is all you needâ from Vaswani and colleagues of a Google Team.",buttonText:"Read the paper",buttonLink:"https://arxiv.org/abs/1706.03762"}),null),e(t,i(n,{title:"GPT-1 launch",time:"June 2018",description:"OpenAI unveils GPT-1, its initial version of GPT."}),null),e(t,i(n,{title:"Google's BERT",time:"October 2018",description:"Pre-trained on vast amounts of text, BERT was the first proper foundational language model that could be fine-tuned for specific tasks, setting new performance standards across various benchmarks.",buttonText:"Read the paper",buttonLink:"https://arxiv.org/abs/1810.04805"}),null),e(t,i(n,{title:"Launch of GPT-2",time:"February 2019",description:"The second version of GPT (Generative Pre-trained Transformer), known as GPT-2, makes its debut."}),null),e(t,i(n,{title:"Partnership between Microsoft and OpenAI",time:"July 2019",description:"The partnership between Microsoft and OpenAI is disclosed, marking the commencement of their collaborative efforts in advancing AI technology."}
1),null),e(t,i(n,{title:"GPT-3 launch",time:"June 2020",description:"OpenAI introduces GPT-3, which has over a thousand times the number of parameters compared to its initial predecessor."}),null),e(t,i(n,{title:"OpenAI launches GPT-3.5 featuring ChatGPT",time:"November 2022",description:"OpenAI introduces GPT-3, which has over a thousand times the number of parameters compared to its initial predecessor."}),null),e(t,i(n,{title:"ChatGPT achieves unprecedented growth",time:"February 2023",description:"It is reported that ChatGPT reached an estimated 100 million monthly active users in January, just two months after its launch, making it the fastest-growing consumer application in history."}),null),e(t,i(n,{title:"Google Research publishes a paper on LLM ads",time:"October 2023",description:"The research exlores auction mechanisms for AI-generated content.",buttonText:"Read the paper",buttonLink:"https://arxiv.org/abs/2310.10826"}),null),e(t,i(n,{title:"Google redesigns search and starts rolling out AI Overviews",time:"May 2024",description:"Google users will see an AI-generated summary at the top of many of their search results."}),null),e(t,i(n,{title:"Google starts experimenting with ads in AI Overview",time:"May 2024",description:"The ads are contextual, but they are not blended in the response."}),null),e(t,i(n,{title:"Introducing Bing generative search",time:"June 2024",description:"LLM answers are now a part of Bing search results."}),null),e(t,i(n,{title:"Copilot responses with ads",time:"October 2024",description:"Copilot will now contain ads, but for now they are shown separately."}),null),e(t,i(n,{title:"LLM ads blended in responses",time:"Soon",description:"We are about to see the raise of ads blended in LLM responses.",buttonText:"Check out the demo",buttonLink:"/demo.html"}),null),r})()}function n(r){return(()=>{var a=y(),l=a.firstChild,o=l.nextSibling,t=o.nextSibling,c=t.nextSibling;return e(o,()=>r.time),e(t,()=>r.title),e(c,()=>r.description),e(a,(()=>{var u=g(()=>!!(r.buttonText&&r.buttonLink));return()=>u()&&(()=>{var s=x(),h=s.firstChild;return e(s,()=>r.buttonText,h),b(()=>p(s,"href",r.buttonLink)),s})()})(),null),a})()}document.addEventListener("DOMContentLoaded",()=>{m(()=>i(w,{}),document.getElementById("timeline"))});
Line numbers count LF bytes from the start of the resource, as the search results do. Vendor segments are library code the classifier recognised; they are stored but not indexed. Bytes are shown as Latin1 characters, one per byte.