[{"data":1,"prerenderedAt":1691},["ShallowReactive",2],{"navigation_docs_en":3,"\u002Fen\u002Fai-engineering\u002Fintro\u002Fch012-foundation-model-use-cases":46,"\u002Fen\u002Fai-engineering\u002Fintro\u002Fch012-foundation-model-use-cases-surround":1686},[4],{"title":5,"icon":6,"path":7,"stem":8,"children":9,"page":45},"AI Engineering",null,"\u002Fen\u002Fai-engineering","en\u002F1.ai-engineering",[10],{"title":11,"icon":12,"path":13,"stem":14,"children":15,"page":45},"Introduction to Building AI Applications with Foundation Models","i-lucide-brain-circuit","\u002Fen\u002Fai-engineering\u002Fintro","en\u002F1.ai-engineering\u002F1.intro",[16,20,25,30,35,40],{"title":11,"path":17,"stem":18,"icon":19},"\u002Fen\u002Fai-engineering\u002Fintro\u002Fch01","en\u002F1.ai-engineering\u002F1.intro\u002Fch01","i-lucide-sparkles",{"title":21,"path":22,"stem":23,"icon":24},"The Rise of AI Engineering","\u002Fen\u002Fai-engineering\u002Fintro\u002Fch011-the-rise-of-ai-engineering","en\u002F1.ai-engineering\u002F1.intro\u002Fch011-the-rise-of-ai-engineering","i-lucide-history",{"title":26,"path":27,"stem":28,"icon":29},"Foundation Model Use Cases","\u002Fen\u002Fai-engineering\u002Fintro\u002Fch012-foundation-model-use-cases","en\u002F1.ai-engineering\u002F1.intro\u002Fch012-foundation-model-use-cases","i-lucide-layout-grid",{"title":31,"path":32,"stem":33,"icon":34},"Planning AI Applications","\u002Fen\u002Fai-engineering\u002Fintro\u002Fch013-planning-ai-applications","en\u002F1.ai-engineering\u002F1.intro\u002Fch013-planning-ai-applications","i-lucide-clipboard-list",{"title":36,"path":37,"stem":38,"icon":39},"The AI Engineering Stack","\u002Fen\u002Fai-engineering\u002Fintro\u002Fch014-the-ai-engineering-stack","en\u002F1.ai-engineering\u002F1.intro\u002Fch014-the-ai-engineering-stack","i-lucide-layers",{"title":41,"path":42,"stem":43,"icon":44},"Summary","\u002Fen\u002Fai-engineering\u002Fintro\u002Fch015-summary","en\u002F1.ai-engineering\u002F1.intro\u002Fch015-summary","i-lucide-flag",false,{"id":47,"title":26,"body":48,"description":1680,"extension":1681,"links":6,"meta":1682,"navigation":1683,"path":27,"seo":1684,"stem":28,"__hash__":1685},"docs_en\u002Fen\u002F1.ai-engineering\u002F1.intro\u002Fch012-foundation-model-use-cases.md",{"type":49,"value":50,"toc":1638},"minimark",[51,87,92,99,171,175,191,203,209,342,346,371,376,523,530,537,543,547,551,568,584,591,597,601,604,607,614,658,663,666,741,745,772,792,798,802,809,813,816,842,849,853,861,874,878,898,901,911,942,946,949,965,974,978,988,995,1024,1040,1043,1066,1070,1089,1106,1112,1116,1120,1143,1163,1172,1176,1179,1183,1207,1220,1224,1231,1235,1238,1291,1304,1308,1311,1323,1327,1333,1337,1344,1364,1370,1374,1377,1381,1403,1407,1418,1422,1425,1448,1461,1465,1468,1486,1493,1497,1500,1509,1516],[52,53,54,58],"u-page-hero",{},[55,56,26],"template",{"v-slot:title":57},"",[55,59,60,80],{"v-slot:description":57},[61,62,63,64,68,69],"p",{},"If you're not already building AI applications, now is a great time to start. The number of potential applications you could build with foundation models seems ",[65,66,67],"strong",{},"endless"," — whatever use case you think of, there's probably an AI for that.",[70,71,72],"sup",{},[73,74,79],"a",{"href":75,"ariaDescribedBy":76,"dataFootnoteRef":57,"id":78},"#user-content-fn-10",[77],"footnote-label","user-content-fnref-10","1",[61,81,82,83,86],{},"If you have an application in mind, you might want to jump to ",[84,85,31],"em",{},". If you're looking for inspiration, this section covers a wide range of industry-proven and promising use cases.",[88,89,91],"h2",{"id":90},"how-people-categorize-use-cases","How People Categorize Use Cases",[61,93,94,95,98],{},"It's impossible to list all potential use cases for AI. Even ",[84,96,97],{},"categorizing"," them is challenging, as different surveys use different lenses.",[100,101,102,126,140,153],"card-group",{},[103,104,107,113,114,117,118,121,122,125],"card",{"icon":105,"title":106},"i-lucide-users","AWS — Three Buckets",[73,108,112],{"href":109,"rel":110},"https:\u002F\u002Faws.amazon.com\u002Fai\u002Fgenerative-ai\u002Fuse-cases\u002F",[111],"nofollow","Amazon Web Services"," groups enterprise generative AI into ",[65,115,116],{},"customer experience",", ",[65,119,120],{},"employee productivity",", and ",[65,123,124],{},"process optimization",".",[103,127,130,131,136,137,125],{"icon":128,"title":129},"i-lucide-list","O'Reilly — Eight Categories","A ",[73,132,135],{"href":133,"rel":134},"https:\u002F\u002Fwww.oreilly.com\u002Fradar\u002Fgenerative-ai-in-the-enterprise\u002F",[111],"2024 O'Reilly survey"," split use cases into ",[65,138,139],{},"programming, data analysis, customer support, marketing copy, other copy, research, web design, and art",[103,141,144,149,150,125],{"icon":142,"title":143},"i-lucide-trending-up","Deloitte — Value Capture",[73,145,148],{"href":146,"rel":147},"https:\u002F\u002Fwww.deloitte.com\u002Fus\u002Fen\u002Fwhat-we-do\u002Fcapabilities\u002Fapplied-artificial-intelligence\u002Fcontent\u002Fai-use-cases.html",[111],"Deloitte"," categorizes by value: ",[65,151,152],{},"cost reduction, process efficiency, growth, and accelerating innovation",[103,154,157,162,163,166,167,170],{"icon":155,"title":156},"i-lucide-shield-alert","Gartner — Business Continuity",[73,158,161],{"href":159,"rel":160},"https:\u002F\u002Fwww.gartner.com\u002Fen\u002Ftopics\u002Fgenerative-ai",[111],"Gartner"," adds ",[84,164,165],{},"business continuity"," — adopt or go out of business. ",[65,168,169],{},"7%"," of 2,500 executives surveyed in 2023 cited this as their motivation.",[88,172,174],{"id":173},"which-occupations-are-most-exposed","Which Occupations Are Most Exposed?",[61,176,177,182,183,186,187,190],{},[73,178,181],{"href":179,"rel":180},"https:\u002F\u002Farxiv.org\u002Fabs\u002F2303.10130",[111],"Eloundou et al. (2023)"," has excellent research on how exposed different occupations are to AI. They defined a task as ",[84,184,185],{},"exposed"," if AI and AI-powered software can reduce the time needed to complete it by at least ",[65,188,189],{},"50%",". An occupation with 80% exposure means that 80% of its tasks are exposed.",[192,193,194,195,198,199,202],"tip",{},"Occupations with ",[65,196,197],{},"100% or close to 100% exposure"," include interpreters and translators, tax preparers, web designers, and writers. Occupations with ",[65,200,201],{},"no exposure"," include cooks, stonemasons, and athletes.",[204,205,206],"blockquote",{},[61,207,208],{},"Table 1-2. Occupations with the highest exposure to AI as annotated by humans. α refers to exposure to AI models directly, whereas β and ζ refer to exposures to AI-powered software. Table from Eloundou et al. (2023).",[210,211,212,228],"table",{},[213,214,215],"thead",{},[216,217,218,222,225],"tr",{},[219,220,221],"th",{},"Group",[219,223,224],{},"Occupations with highest exposure",[219,226,227],{},"% Exposure",[229,230,231,272,303],"tbody",{},[216,232,233,240,256],{},[234,235,236,237],"td",{},"Human ",[84,238,239],{},"α",[234,241,242,243,246,247,249,250,252,253,255],{},"Interpreters and translators",[244,245],"br",{},"Survey researchers",[244,248],{},"Poets, lyrics, and creative writers",[244,251],{},"Animal scientists",[244,254],{},"Public relations specialists",[234,257,258,259,261,262,264,265,267,268,267,270],{},"76.5",[244,260],{},"75.0",[244,263],{},"68.8",[244,266],{},"66.7",[244,269],{},[244,271],{},[216,273,274,279,288],{},[234,275,236,276],{},[84,277,278],{},"β",[234,280,246,281,283,284,255,286,252],{},[244,282],{},"Writers and authors",[244,285],{},[244,287],{},[234,289,290,291,293,294,296,297,299,300,302],{},"84.4",[244,292],{},"82.5",[244,295],{},"82.4",[244,298],{},"80.6",[244,301],{},"77.8",[216,304,305,310,331],{},[234,306,236,307],{},[84,308,309],{},"ζ",[234,311,312,313,315,316,318,319,283,321,323,324,326,329],{},"Mathematicians",[244,314],{},"Tax preparers",[244,317],{},"Financial quantitative analysts",[244,320],{},[244,322],{},"Web and digital interface designers",[244,325],{},[84,327,328],{},"Humans labeled 15 occupations as \"fully exposed\".",[244,330],{},[234,332,333,334,333,336,333,338,333,340,333],{},"100.0",[244,335],{},[244,337],{},[244,339],{},[244,341],{},[88,343,345],{"id":344},"eight-common-use-cases","Eight Common Use Cases",[347,348,349,350,353,354,357,358,361,362,370],"note",{},"To understand enterprise use cases, I interviewed ",[65,351,352],{},"50 companies"," and read ",[65,355,356],{},"over 100 case studies",". To understand consumer applications, I examined ",[65,359,360],{},"205 open source AI applications"," with at least 500 stars on GitHub.",[70,363,364],{},[73,365,369],{"href":366,"ariaDescribedBy":367,"dataFootnoteRef":57,"id":368},"#user-content-fn-11",[77],"user-content-fnref-11","2"," The list here is a reference — as you learn more about how to build foundation models in Chapter 2 and how to evaluate them in Chapter 3, you'll be able to form a better picture of what use cases foundation models can and should be used for.",[204,372,373],{},[61,374,375],{},"Table 1-3. Common generative AI use cases across consumer and enterprise applications.",[210,377,378,391],{},[213,379,380],{},[216,381,382,385,388],{},[219,383,384],{},"Category",[219,386,387],{},"Examples of consumer use cases",[219,389,390],{},"Examples of enterprise use cases",[229,392,393,402,416,433,450,467,483,504],{},[216,394,395,398,400],{},[234,396,397],{},"Coding",[234,399,397],{},[234,401,397],{},[216,403,404,407,410],{},[234,405,406],{},"Image and video",[234,408,409],{},"Photo and video editing",[234,411,412,413,415],{},"Presentation ",[244,414],{},"Ad generation",[216,417,418,421,427],{},[234,419,420],{},"Writing",[234,422,423,424,426],{},"Email",[244,425],{},"Social media and blog posts",[234,428,429,430,432],{},"Copywriting, search engine optimization (SEO)",[244,431],{},"Reports, memos, design docs",[216,434,435,438,444],{},[234,436,437],{},"Education",[234,439,440,441,443],{},"Tutoring",[244,442],{},"Essay grading",[234,445,446,447,449],{},"Employee onboarding",[244,448],{},"Employee upskill training",[216,451,452,455,461],{},[234,453,454],{},"Conversational bots",[234,456,457,458,460],{},"General chatbot",[244,459],{},"AI companion",[234,462,463,464,466],{},"Customer support",[244,465],{},"Product copilots",[216,468,469,472,478],{},[234,470,471],{},"Information aggregation",[234,473,474,475,477],{},"Summarization",[244,476],{},"Talk-to-your-docs",[234,479,474,480,482],{},[244,481],{},"Market research",[216,484,485,488,498],{},[234,486,487],{},"Data organization",[234,489,490,491,493],{},"Image search",[244,492],{},[73,494,497],{"href":495,"rel":496},"https:\u002F\u002Fen.wikipedia.org\u002Fwiki\u002FMemex",[111],"Memex",[234,499,500,501,503],{},"Knowledge management",[244,502],{},"Document processing",[216,505,506,509,515],{},[234,507,508],{},"Workflow automation",[234,510,511,512,514],{},"Travel planning",[244,513],{},"Event planning",[234,516,517,518,520,521],{},"Data extraction, entry, and annotation",[244,519],{},"Lead generation",[244,522],{},[347,524,525,526,529],{},"Because foundation models are general, applications built on top of them can solve many problems — so an application can belong to ",[65,527,528],{},"more than one category",". A bot can provide companionship and aggregate information. An application can extract structured data from a PDF and answer questions about that PDF.",[61,531,532,536],{},[73,533,535],{"href":534},".\u002Fmedia\u002Ffig-1-7.png","Figure 1-7"," shows the distribution of these use cases among the 205 open source applications. Note that the small percentage of education, data organization, and writing use cases doesn't mean these are unpopular — it just means those applications aren't open source. Builders may find them more suitable for enterprise use cases.",[61,538,539],{},[540,541],"img",{"alt":542,"src":534},"Figure 1-7. Distribution of use cases in the 205 open source repositories on GitHub.",[204,544,545],{},[61,546,542],{},[88,548,550],{"id":549},"enterprises-prefer-lower-risk-applications","Enterprises Prefer Lower-Risk Applications",[61,552,553,554,559,560,563,564,567],{},"The enterprise world generally prefers applications with lower risks. A ",[73,555,558],{"href":556,"rel":557},"https:\u002F\u002Fa16z.com\u002Fgenerative-ai-enterprise-2024\u002F",[111],"2024 a16z Growth report"," showed that companies are faster to deploy ",[65,561,562],{},"internal-facing applications"," (internal knowledge management) than ",[65,565,566],{},"external-facing applications"," (customer support chatbots).",[100,569,570,579],{},[103,571,574,575,578],{"icon":572,"title":573},"i-lucide-building-2","Internal-Facing","Helps companies develop AI engineering expertise while ",[65,576,577],{},"minimizing risks"," around data privacy, compliance, and potential catastrophic failures.",[103,580,583],{"icon":581,"title":582},"i-lucide-globe","External-Facing","Higher risk — customer-visible failures hit reputation, support load, and trust. Deployed more cautiously.",[61,585,586,587,590],{},"Similarly, while foundation models are open-ended and can be used for any task, many applications built on top of them are still close-ended, such as classification. ",[65,588,589],{},"Classification tasks are easier to evaluate",", which makes their risks easier to estimate.",[61,592,593],{},[540,594],{"alt":595,"src":596},"Figure 1-8. Companies are more willing to deploy internal-facing applications",".\u002Fmedia\u002Ffig-1-8.png",[204,598,599],{},[61,600,595],{},[192,602,603],{},"Even after seeing hundreds of AI applications, I still find new ones that surprise me every week. In the early days of the internet, few people foresaw that the dominating use case would one day be social media. As we learn to make the most out of AI, the use case that eventually dominates might surprise us. With luck, it'll be a good surprise.",[88,605,397],{"id":606},"coding",[61,608,609,610,613],{},"In multiple generative AI surveys, ",[65,611,612],{},"coding is hands down the most popular use case",". AI coding tools are popular both because AI is good at coding and because early AI engineers are coders who are more exposed to coding challenges.",[61,615,616,617,620,621,631,632,637,638,643,644,637,649,654,655,125],{},"One of the earliest successes of foundation models in production is ",[65,618,619],{},"GitHub Copilot",", whose ",[73,622,625,626,630],{"href":623,"rel":624},"https:\u002F\u002Fwww.theinformation.com\u002Fbriefings\u002Fmicrosoft-github-copilot-revenue-100-million-ARR-ai",[111],"annual recurring revenue crossed ",[627,628,629],"code",{},"$100"," million"," only two years after its launch. AI-powered coding startups have raised hundreds of millions of dollars — ",[73,633,636],{"href":634,"rel":635},"https:\u002F\u002Ftechcrunch.com\u002F2024\u002F08\u002F29\u002Fgenerative-ai-coding-startup-magic-lands-320m-investment-from-eric-schmidt-atlassian-and-others\u002F",[111],"Magic raised $320 million"," and ",[73,639,642],{"href":640,"rel":641},"https:\u002F\u002Fcursor.com\u002Fblog\u002Fseries-a",[111],"Anysphere raised $60 million",", both in August 2024. Open source tools like ",[73,645,648],{"href":646,"rel":647},"https:\u002F\u002Fgithub.com\u002Fgpt-engineer-org\u002Fgpt-engineer",[111],"gpt-engineer",[73,650,653],{"href":651,"rel":652},"https:\u002F\u002Fgithub.com\u002Fabi\u002Fscreenshot-to-code",[111],"screenshot-to-code"," both reached ",[65,656,657],{},"50,000 GitHub stars within a year",[659,660,662],"h3",{"id":661},"specialized-coding-tasks","Specialized Coding Tasks",[61,664,665],{},"Beyond general code completion, many tools specialize:",[667,668,669,679,697,705,718,726,733],"ul",{},[670,671,672,673,678],"li",{},"Extracting structured data from web pages and PDFs (",[73,674,677],{"href":675,"rel":676},"https:\u002F\u002Fgithub.com\u002Freworkd\u002FAgentGPT",[111],"AgentGPT",")",[670,680,681,682,117,687,117,692,678],{},"Converting English to code (",[73,683,686],{"href":684,"rel":685},"https:\u002F\u002Fgithub.com\u002Feosphoros-ai\u002FDB-GPT",[111],"DB-GPT",[73,688,691],{"href":689,"rel":690},"https:\u002F\u002Fgithub.com\u002Fsqlchat\u002Fsqlchat",[111],"SQL Chat",[73,693,696],{"href":694,"rel":695},"https:\u002F\u002Fgithub.com\u002FSinaptik-AI\u002Fpandas-ai",[111],"PandasAI",[670,698,699,700,678],{},"Generating a website from a design or screenshot (screenshot-to-code, ",[73,701,704],{"href":702,"rel":703},"https:\u002F\u002Fgithub.com\u002Fsawyerhood\u002Fdraw-a-ui",[111],"draw-a-ui",[670,706,707,708,117,713,678],{},"Translating between programming languages or frameworks (",[73,709,712],{"href":710,"rel":711},"https:\u002F\u002Fgithub.com\u002Fjoshpxyne\u002Fgpt-migrate",[111],"GPT-Migrate",[73,714,717],{"href":715,"rel":716},"https:\u002F\u002Fgithub.com\u002Fmckaywrigley\u002Fai-code-translator",[111],"AI Code Translator",[670,719,720,721,678],{},"Writing documentation (",[73,722,725],{"href":723,"rel":724},"https:\u002F\u002Fgithub.com\u002Fcontext-labs\u002Fautodoc",[111],"Autodoc",[670,727,728],{},[73,729,732],{"href":730,"rel":731},"https:\u002F\u002Fgithub.com\u002FGreyDGL\u002FPentestGPT",[111],"Creating tests (PentestGPT)",[670,734,735,736,678],{},"Generating commit messages (",[73,737,740],{"href":738,"rel":739},"https:\u002F\u002Fgithub.com\u002FNutlope\u002Faicommits",[111],"AI Commits",[659,742,744],{"id":743},"will-ai-replace-software-engineers","Will AI Replace Software Engineers?",[100,746,747,767],{},[103,748,751,756,757,762,763,766],{"icon":749,"title":750},"i-lucide-zap","The \"Yes\" Camp",[73,752,755],{"href":753,"rel":754},"https:\u002F\u002Fwww.tomshardware.com\u002Ftech-industry\u002Fartificial-intelligence\u002Fjensen-huang-advises-against-learning-to-code-leave-it-up-to-ai",[111],"Jensen Huang, CEO of NVIDIA",", predicts AI will replace human software engineers and that we should stop saying kids should learn to code. ",[73,758,761],{"href":759,"rel":760},"https:\u002F\u002Fwww.businessinsider.com\u002Faws-ceo-developers-stop-coding-ai-takes-over-2024-8",[111],"AWS CEO Matt Garman"," said in a leaked recording that most developers will stop coding in the near future — though he means their ",[65,764,765],{},"jobs will change",", not end.",[103,768,771],{"icon":769,"title":770},"i-lucide-shield","The \"No\" Camp","Many software engineers are convinced they will never be replaced by AI — both for technical and emotional reasons. (People don't like admitting that they can be replaced.)",[61,773,774,779,780,783,784,787,788,791],{},[73,775,778],{"href":776,"rel":777},"https:\u002F\u002Fwww.mckinsey.com\u002Fcapabilities\u002Ftech-and-ai\u002Four-insights\u002Funleashing-developer-productivity-with-generative-ai",[111],"McKinsey researchers"," found that AI can help developers be ",[65,781,782],{},"twice as productive"," for documentation and ",[65,785,786],{},"25–50% more productive"," for code generation and refactoring. Minimal productivity improvement was observed for highly complex tasks. In conversations with developers of AI coding tools, many told me that AI is much better at ",[65,789,790],{},"frontend than backend"," development.",[61,793,794],{},[540,795],{"alt":796,"src":797},"Figure 1-9. AI can help developers be significantly more productive, especially for simple tasks, but this applies less for highly complex tasks. Data by McKinsey.",".\u002Fmedia\u002Ffig-1-9.png",[204,799,800],{},[61,801,796],{},[347,803,804,805,808],{},"Regardless of whether AI replaces software engineers, it can certainly make them more productive — meaning companies can now accomplish more with ",[65,806,807],{},"fewer engineers",". AI may also disrupt the outsourcing industry, as outsourced tasks tend to be simpler ones outside of a company's core business.",[88,810,812],{"id":811},"image-and-video-production","Image and Video Production",[61,814,815],{},"Thanks to its probabilistic nature, AI is great for creative tasks. Some of the most successful AI startups are creative applications.",[100,817,818,832,837],{},[103,819,822,823,827,828,831],{"icon":820,"title":821},"i-lucide-image","Midjourney","Image generation. At one and a half years old (late 2023), ",[73,824,821],{"href":825,"rel":826},"https:\u002F\u002Fwww.theinformation.com\u002Fbriefings\u002Fai-startup-midjourney-expects-200-million-in-revenue",[111]," had already generated ",[65,829,830],{},"$200 million"," in annual recurring revenue.",[103,833,836],{"icon":834,"title":835},"i-lucide-wand-sparkles","Adobe Firefly","Photo editing built into the Adobe ecosystem.",[103,838,841],{"icon":839,"title":840},"i-lucide-film","Runway, Pika, Sora","Video generation tools setting a rapid pace for AI-driven video.",[61,843,844,845,848],{},"As of December 2023, among the ",[65,846,847],{},"top 10 free apps"," for Graphics & Design on the Apple App Store, half had \"AI\" in their names. Soon, graphics and design apps will incorporate AI by default and the word \"AI\" will disappear from product names. Chapter 2 discusses the probabilistic nature of AI in more detail.",[659,850,852],{"id":851},"ai-generated-profile-pictures","AI-Generated Profile Pictures",[61,854,855,856,125],{},"It's now common to use AI to generate profile pictures for social media, from LinkedIn to TikTok. Many candidates believe AI-generated headshots can ",[73,857,860],{"href":858,"rel":859},"https:\u002F\u002Fwww.bbc.com\u002Fnews\u002Fbusiness-67054382",[111],"increase their chances of landing a job",[347,862,863,864,869,870,873],{},"The perception of AI-generated profile pictures has changed dramatically. In 2019, ",[73,865,868],{"href":866,"rel":867},"https:\u002F\u002Fwww.wired.com\u002Fstory\u002Ffacebook-removes-accounts-ai-generated-photos\u002F",[111],"Facebook"," banned accounts using AI-generated profile photos for safety reasons. By 2023, many social media apps ",[65,871,872],{},"provide tools"," that let users generate profile photos with AI.",[659,875,877],{"id":876},"ads-and-marketing","Ads and Marketing",[61,879,880,881,889,890,893,894,897],{},"For enterprises, ads and marketing have been quick to incorporate AI.",[70,882,883],{},[73,884,888],{"href":885,"ariaDescribedBy":886,"dataFootnoteRef":57,"id":887},"#user-content-fn-12",[77],"user-content-fnref-12","3"," AI can generate promotional images and videos directly, brainstorm ideas, or produce first drafts for human experts to iterate on. You can use AI to generate multiple ads and test which one works best — or generate variations by ",[65,891,892],{},"season"," (changing leaf colors during fall) or ",[65,895,896],{},"location"," (adding snow during winter).",[88,899,420],{"id":900},"writing",[61,902,903,904,637,907,910],{},"AI has long been used to aid writing. If you use a smartphone, you're probably familiar with ",[65,905,906],{},"autocorrect",[65,908,909],{},"auto-completion",", both powered by AI. Writing is an ideal application: we do it a lot, it can be tedious, and we have a high tolerance for mistakes — if a model suggests something you don't like, you can just ignore it.",[192,912,913,916,917,922,923,926,927,637,930,933,934,937,938,941],{},[65,914,915],{},"An MIT study on ChatGPT and writing"," (",[73,918,921],{"href":919,"rel":920},"https:\u002F\u002Fwww.science.org\u002Fdoi\u002F10.1126\u002Fscience.adh2586",[111],"Noy and Zhang, 2023",") assigned occupation-specific writing tasks to ",[65,924,925],{},"453 college-educated professionals"," and randomly exposed half to ChatGPT. Among those exposed, ",[65,928,929],{},"average time decreased by 40%",[65,931,932],{},"output quality rose by 18%",". ChatGPT helped close the gap in output quality between workers — it's most helpful to those with less inclination for writing. Workers exposed during the experiment were ",[65,935,936],{},"2× as likely"," to report using it in their real job two weeks later, and ",[65,939,940],{},"1.6× as likely"," two months after.",[659,943,945],{"id":944},"consumer-use-cases","Consumer Use Cases",[61,947,948],{},"The use cases are obvious. Many use AI to communicate better — you can be angry in an email and ask AI to make it pleasant, or hand over bullet points and get back complete paragraphs. Several people claimed they no longer send an important email without asking AI to improve it first.",[61,950,951,952,960,961,964],{},"Students are using AI to write essays. Writers are using AI to write books.",[70,953,954],{},[73,955,959],{"href":956,"ariaDescribedBy":957,"dataFootnoteRef":57,"id":958},"#user-content-fn-13",[77],"user-content-fnref-13","4"," Many startups already use AI to generate children's, fan fiction, romance, and fantasy books. Unlike traditional books, AI-generated books can be ",[65,962,963],{},"interactive"," — a plot can change depending on the reader's preference. A children's reading app identifies the words a child has trouble with and generates stories centered around them.",[61,966,967,968,973],{},"Note-taking and email apps like Google Docs, Notion, and Gmail all use AI to help users improve their writing. ",[73,969,972],{"href":970,"rel":971},"https:\u002F\u002Farxiv.org\u002Fabs\u002F2305.09857",[111],"Grammarly"," finetunes a model to make users' writing more fluent, coherent, and clear.",[659,975,977],{"id":976},"enterprise-use-cases","Enterprise Use Cases",[61,979,980,981,637,984,987],{},"For enterprises, AI writing is common in sales, marketing, and general team communication. Many managers told me they use AI to help write performance reports. AI helps craft cold outreach emails, ad copywriting, and product descriptions. CRM apps like ",[65,982,983],{},"HubSpot",[65,985,986],{},"Salesforce"," also have tools for enterprise users to generate web content and outreach emails.",[61,989,990,991,994],{},"AI seems particularly good with ",[65,992,993],{},"SEO"," — perhaps because many models are trained with data from the internet, which is populated with SEO-optimized text.",[996,997,998,1001,1002,1007,1008,1011,1012,1015,1016],"warning",{},[65,999,1000],{},"SEO content farms"," — AI is so good at SEO it has enabled a new generation of content farms. These set up junk websites filled with AI-generated content to rank high on Google, then sell ad spots through exchanges. In June 2023, ",[73,1003,1006],{"href":1004,"rel":1005},"https:\u002F\u002Fwww.newsguardtech.com\u002Fmisinformation-monitor\u002Fjune-2023\u002F",[111],"NewsGuard"," identified almost ",[65,1009,1010],{},"400 ads from 141 popular brands"," on junk AI-generated websites. One junk website produced ",[65,1013,1014],{},"1,200 articles a day",". Unless something is done to curtail this, the future of internet content will be AI-generated, and it'll be pretty bleak.",[70,1017,1018],{},[73,1019,1023],{"href":1020,"ariaDescribedBy":1021,"dataFootnoteRef":57,"id":1022},"#user-content-fn-14",[77],"user-content-fnref-14","5",[1025,1026,1027,1030,1031,1036,1037,125],"caution",{},[65,1028,1029],{},"The 2023 Amazon scandal"," — the ",[73,1032,1035],{"href":1033,"rel":1034},"https:\u002F\u002Fwww.nytimes.com\u002F2023\u002F08\u002F05\u002Ftravel\u002Famazon-guidebooks-artificial-intelligence.html",[111],"New York Times"," reported that Amazon was flooded with shoddy AI-generated travel guidebooks, each outfitted with an author bio, a website, and rave reviews — ",[65,1038,1039],{},"all AI-generated",[88,1041,437],{"id":1042},"education",[61,1044,1045,1046,1049,1050,1053,1054,1059,1060,1065],{},"Whenever ChatGPT is down, OpenAI's Discord server is flooded with students complaining about being unable to complete their homework. Several education boards, including the ",[65,1047,1048],{},"New York City Public Schools"," and the ",[65,1051,1052],{},"Los Angeles Unified School District",", were quick to ",[73,1055,1058],{"href":1056,"rel":1057},"https:\u002F\u002Fwww.forbes.com\u002Fsites\u002Fariannajohnson\u002F2023\u002F01\u002F18\u002Fchatgpt-in-schools-heres-where-its-banned-and-how-it-could-potentially-help-students\u002F?sh=3fdc0b656e2c",[111],"ban ChatGPT"," for fear of cheating, but ",[73,1061,1064],{"href":1062,"rel":1063},"https:\u002F\u002Fwww.nytimes.com\u002F2023\u002F08\u002F24\u002Fbusiness\u002Fschools-chatgpt-chatbot-bans.html",[111],"reversed their decisions"," just months later.",[659,1067,1069],{"id":1068},"personalized-learning","Personalized Learning",[192,1071,1072,1078],{},[61,1073,1074,1077],{},[65,1075,1076],{},"It's strange that ads are personalized but education is not."," AI can adapt materials to the format best suited for each student:",[667,1079,1080,1083,1086],{},[670,1081,1082],{},"Auditory learners can ask AI to read materials out loud",[670,1084,1085],{},"Students who love animals can ask AI to feature more animals in visualizations",[670,1087,1088],{},"Those who find code easier than math can ask AI to translate equations into code",[61,1090,1091,1092,1095,1096,1101,1102,1105],{},"AI is especially helpful for ",[65,1093,1094],{},"language learning"," — you can ask AI to roleplay practice scenarios. ",[73,1097,1100],{"href":1098,"rel":1099},"https:\u002F\u002Fblog.duolingo.com\u002Fhow-duolingo-experts-work-with-ai\u002F",[111],"Pajak and Bicknell (Duolingo, 2022)"," found that out of four stages of course creation, ",[65,1103,1104],{},"lesson personalization"," is the stage that benefits most from AI.",[61,1107,1108],{},[540,1109],{"alt":1110,"src":1111},"Figure 1-10. AI can be used throughout all four stages of course creation at Duolingo, but it's the most helpful in the personalization stage. Image from Pajak and Bicknell (Duolingo, 2022).",".\u002Fmedia\u002Ffig-1-10.png",[204,1113,1114],{},[61,1115,1110],{},[659,1117,1119],{"id":1118},"quizzes-debate-and-teaching-assistants","Quizzes, Debate, and Teaching Assistants",[61,1121,1122,1123,1126,1127,1132,1133,1138,1139,1142],{},"AI can generate quizzes — both multiple-choice and open-ended — and evaluate the answers. AI can become a ",[65,1124,1125],{},"debate partner"," as it's much better at presenting different views on the same topic than the average human. ",[73,1128,1131],{"href":1129,"rel":1130},"https:\u002F\u002Fblog.khanacademy.org\u002Fharnessing-ai-so-that-all-students-benefit-a-nonprofit-approach-for-equal-access\u002F",[111],"Khan Academy"," offers ",[73,1134,1137],{"href":1135,"rel":1136},"https:\u002F\u002Fwww.khanmigo.ai\u002F",[111],"AI-powered"," teaching assistants to students and course assistants to teachers. One innovative method: teachers assign ",[65,1140,1141],{},"AI-generated essays"," for students to find and correct mistakes.",[996,1144,1145,1148,1149,1152,1153,1156,1157,1162],{},[65,1146,1147],{},"Disruption is real."," Chegg, a company that helps students with homework, saw its share price plummet from ",[65,1150,1151],{},"$28"," when ChatGPT launched in November 2022 to ",[65,1154,1155],{},"$2"," in September 2024 — ",[73,1158,1161],{"href":1159,"rel":1160},"https:\u002F\u002Fwww.wired.com\u002Fstory\u002Fchegg-embraced-ai-chatgpt-ate-its-lunch-anyway\u002F",[111],"students have been turning to AI for help"," instead.",[192,1164,1165,1168,1169,125],{},[65,1166,1167],{},"The opportunity:"," if the risk is that AI can replace many skills, the opportunity is that AI can be used as a tutor to learn any skill. For many skills, AI can help someone get up to speed quickly and then continue learning on their own — to become ",[65,1170,1171],{},"better than AI",[88,1173,1175],{"id":1174},"conversational-bots","Conversational Bots",[61,1177,1178],{},"Conversational bots are versatile. They help us find information, explain concepts, and brainstorm ideas. AI can be your companion and therapist. It can emulate personalities, letting you talk to a digital copy of anyone you like.",[659,1180,1182],{"id":1181},"companions","Companions",[61,1184,1185,1186,637,1191,1195,1196,1201,1202,125],{},"Digital girlfriends and boyfriends have become weirdly popular in an incredibly short amount of time. Many are already spending more time talking to bots than to humans (see discussions ",[73,1187,1190],{"href":1188,"rel":1189},"https:\u002F\u002Fwww.reddit.com\u002Fr\u002FCharacterAI\u002Fcomments\u002F11r4gwi\u002Fspent_more_time_on_charecter_ai_than_speaking_to\u002F",[111],"here",[73,1192,1190],{"href":1193,"rel":1194},"https:\u002F\u002Fwww.myaiobsession.com\u002Fis-it-so-wrong-to-spend-two-hours?hide_intro_popup=true",[111],"). Some are worried that AI will ",[73,1197,1200],{"href":1198,"rel":1199},"https:\u002F\u002Fthehill.com\u002Fopinion\u002Ftechnology\u002F4218666-ai-girlfriends-are-ruining-an-entire-generation-of-men\u002F",[111],"ruin"," ",[73,1203,1206],{"href":1204,"rel":1205},"https:\u002F\u002Fedition.cnn.com\u002Fvideos\u002Fbusiness\u002F2023\u002F10\u002F01\u002Fai-girlfriends-ruining-generation-of-men-smerconish-vpx.cnn",[111],"dating",[61,1208,1209,1210,1213,1214,1219],{},"In research, people have used groups of conversational bots to ",[65,1211,1212],{},"simulate a society",", enabling studies on social dynamics (",[73,1215,1218],{"href":1216,"rel":1217},"https:\u002F\u002Farxiv.org\u002Fabs\u002F2304.03442",[111],"Park et al., 2023",").",[659,1221,1223],{"id":1222},"enterprise-bots","Enterprise Bots",[61,1225,1226,1227,1230],{},"For enterprises, the most popular bots are customer support bots. They help companies save costs while improving customer experience by responding sooner than human agents. AI can also be ",[65,1228,1229],{},"product copilots"," that guide customers through painful and confusing tasks — filing insurance claims, doing taxes, looking up corporate policies.",[659,1232,1234],{"id":1233},"beyond-text","Beyond Text",[61,1236,1237],{},"The success of ChatGPT prompted a wave of text-based conversational bots, but text isn't the only interface.",[100,1239,1240,1256,1268],{},[103,1241,1244,1247,1248],{"icon":1242,"title":1243},"i-lucide-mic","Voice Assistants",[65,1245,1246],{},"Google Assistant, Siri, Alexa"," — around for years.",[70,1249,1250],{},[73,1251,1255],{"href":1252,"ariaDescribedBy":1253,"dataFootnoteRef":57,"id":1254},"#user-content-fn-15",[77],"user-content-fnref-15","6",[103,1257,1260,1261,1264,1265,125],{"icon":1258,"title":1259},"i-lucide-box","3D Conversational Bots","Common in ",[65,1262,1263],{},"games",", gaining traction in ",[65,1266,1267],{},"retail and marketing",[103,1269,1272,1273,637,1278,1219,1283],{"icon":1270,"title":1271},"i-lucide-gamepad-2","Smart NPCs","Non-player characters powered by AI (NVIDIA's demos of ",[73,1274,1277],{"href":1275,"rel":1276},"https:\u002F\u002Fwww.youtube.com\u002Fwatch?v=uryeFhnNzEs",[111],"Inworld",[73,1279,1282],{"href":1280,"rel":1281},"https:\u002F\u002Fwww.youtube.com\u002Fwatch?v=psrXGPh80UM",[111],"Convai",[70,1284,1285],{},[73,1286,1290],{"href":1287,"ariaDescribedBy":1288,"dataFootnoteRef":57,"id":1289},"#user-content-fn-16",[77],"user-content-fnref-16","7",[61,1292,1293,1294,637,1297,1300,1301,125],{},"NPCs are essential for advancing the storyline of many games. Without AI, they're typically scripted to do simple actions with a limited range of dialogues. AI can make these NPCs much smarter, changing the dynamics of existing games like ",[84,1295,1296],{},"The Sims",[84,1298,1299],{},"Skyrim"," — and enabling ",[65,1302,1303],{},"new games never possible before",[88,1305,1307],{"id":1306},"information-aggregation","Information Aggregation",[61,1309,1310],{},"Many people believe success depends on the ability to filter and digest useful information. But keeping up with emails, Slack messages, and news can be overwhelming.",[192,1312,1313,1314,117,1319,1322],{},"According to ",[73,1315,1318],{"href":1316,"rel":1317},"https:\u002F\u002Fwww.salesforce.com\u002Fnews\u002Fpress-releases\u002F2023\u002F09\u002F07\u002Fai-usage-research\u002F",[111],"Salesforce's 2023 Generative AI Snapshot Research",[65,1320,1321],{},"74% of generative AI users"," use it to distill complex ideas and summarize information.",[659,1324,1326],{"id":1325},"talk-to-your-docs","Talk-to-Your-Docs",[61,1328,1329,1330,1332],{},"For consumers, many applications process your documents — contracts, disclosures, papers — and let you retrieve information conversationally. This use case is also called ",[84,1331,1325],{},". AI can summarize websites, research, and create reports on topics of your choice.",[659,1334,1336],{"id":1335},"enterprise-aggregation","Enterprise Aggregation",[61,1338,1339,1340,1343],{},"Information aggregation and distillation are essential for enterprise operations. More efficient information flow can help an organization become ",[65,1341,1342],{},"leaner",", reducing the burden on middle management.",[347,1345,1346,1349,1350,1355,1356,1359,1360,1363],{},[65,1347,1348],{},"Instacart's Fast Breakdown."," When ",[73,1351,1354],{"href":1352,"rel":1353},"https:\u002F\u002Ftech.instacart.com\u002Fscaling-productivity-with-ava-instacarts-internal-ai-assistant-ed7f02558d84",[111],"Instacart"," launched an internal prompt marketplace, one of the most popular templates was ",[84,1357,1358],{},"Fast Breakdown",". It asks AI to summarize meeting notes, emails, and Slack conversations into ",[65,1361,1362],{},"facts, open questions, and action items"," — which can then be automatically inserted into a project tracking tool and assigned to the right owners.",[61,1365,1366,1367,125],{},"AI can help you surface critical information about potential customers and run analyses on competitors. The more information you gather, the more important it is to organize it — information aggregation goes hand in hand with ",[65,1368,1369],{},"data organization",[88,1371,1373],{"id":1372},"data-organization","Data Organization",[61,1375,1376],{},"One thing certain about the future is that we'll continue producing more and more data. Smartphone users will keep taking photos and videos. Companies will keep logging everything about their products, employees, and customers. Billions of contracts are created each year. Photos, videos, logs, and PDFs are all unstructured or semi-structured data — and it's essential to organize it in a way that can be searched later.",[659,1378,1380],{"id":1379},"visual-search","Visual Search",[61,1382,1383,1384,1387,1388,1201,1396,1399,1400,125],{},"AI can automatically generate text descriptions about images and videos, or match text queries with visuals. ",[65,1385,1386],{},"Google Photos"," uses AI to surface images matching search queries.",[70,1389,1390],{},[73,1391,1395],{"href":1392,"ariaDescribedBy":1393,"dataFootnoteRef":57,"id":1394},"#user-content-fn-17",[77],"user-content-fnref-17","8",[65,1397,1398],{},"Google Image Search"," goes a step further: if there's no existing image matching users' needs, it can ",[65,1401,1402],{},"generate one",[659,1404,1406],{"id":1405},"data-analysis","Data Analysis",[61,1408,1409,1410],{},"AI is very good with data analysis. It can write programs to generate data visualizations, identify outliers, and make predictions like revenue forecasts.",[70,1411,1412],{},[73,1413,1417],{"href":1414,"ariaDescribedBy":1415,"dataFootnoteRef":57,"id":1416},"#user-content-fn-18",[77],"user-content-fnref-18","9",[659,1419,1421],{"id":1420},"intelligent-document-processing","Intelligent Document Processing",[61,1423,1424],{},"Enterprises can use AI to extract structured information from unstructured data, which can be used to organize and search.",[100,1426,1427,1436],{},[103,1428,1431,1432,1435],{"icon":1429,"title":1430},"i-lucide-credit-card","Simple Extraction","Credit cards, driver's licenses, ",[65,1433,1434],{},"receipts",", tickets, contact information from email footers.",[103,1437,1440,1443,1444,1447],{"icon":1438,"title":1439},"i-lucide-file-text","Complex Extraction",[65,1441,1442],{},"Contracts",", reports, ",[65,1445,1446],{},"charts",", and more — turning long documents into queryable data.",[192,1449,1450,1451,1454,1455,1460],{},"The ",[65,1452,1453],{},"IDP (Intelligent Document Processing)"," industry is estimated to reach ",[73,1456,1459],{"href":1457,"rel":1458},"https:\u002F\u002Fwww.globenewswire.com\u002Fnews-release\u002F2023\u002F11\u002F21\u002F2783889\u002F0\u002Fen\u002FIntelligent-Document-Processing-Market-Size-to-Surpass-USD-12-81-billion-by-2030-exhibiting-a-CAGR-of-32-9.html",[111],"$12.81 billion by 2030, growing 32.9%"," per year.",[88,1462,1464],{"id":1463},"workflow-automation","Workflow Automation",[61,1466,1467],{},"Ultimately, AI should automate as much as possible.",[100,1469,1470,1478],{},[103,1471,1474,1475,125],{"icon":1472,"title":1473},"i-lucide-user","For End Users","Booking restaurants, requesting refunds, planning trips, filling out forms — the ",[65,1476,1477],{},"boring daily tasks",[103,1479,1482,1483,125],{"icon":1480,"title":1481},"i-lucide-briefcase","For Enterprises","Lead management, invoicing, reimbursements, customer requests, data entry — the ",[65,1484,1485],{},"repetitive tasks",[61,1487,1488,1489,1492],{},"One especially exciting use case is using AI models to ",[65,1490,1491],{},"synthesize data",", which can then be used to improve the models themselves. You can use AI to create labels for your data, looping in humans to improve the labels. We discuss data synthesis in Chapter 8.",[659,1494,1496],{"id":1495},"agents","Agents",[61,1498,1499],{},"Access to external tools is required to accomplish many tasks. To book a restaurant, an application might need permission to open a search engine to look up the restaurant's number, use your phone to make calls, and add appointments to your calendar.",[347,1501,1502,1503,1505,1506],{},"AIs that can plan and use tools are called ",[65,1504,1495],{},". The level of interest around agents borders on obsession — but it's not entirely unwarranted. AI agents have the potential to make every person vastly more productive and generate vastly more economic value. ",[65,1507,1508],{},"Agents are a central topic in Chapter 6.",[192,1510,1511,1512,1515],{},"It's been a lot of fun looking into different AI applications. One of my favorite things to daydream about is the different applications I can build. However, ",[65,1513,1514],{},"not all applications should be built"," — the next section discusses what to consider before building one.",[1517,1518,1521,1526],"section",{"className":1519,"dataFootnotes":57},[1520],"footnotes",[88,1522,1525],{"className":1523,"id":77},[1524],"sr-only","Footnotes",[1527,1528,1529,1546,1565,1584,1593,1602,1611,1620,1629],"ol",{},[670,1530,1532,1533,1538,1539],{"id":1531},"user-content-fn-10","Fun fact: as of September 16, 2024, the website ",[73,1534,1537],{"href":1535,"rel":1536},"https:\u002F\u002Ftheresanaiforthat.com\u002F",[111],"theresanaiforthat.com"," lists 16,814 AIs for 14,688 tasks and 4,803 jobs. ",[73,1540,1545],{"href":1541,"ariaLabel":1542,"className":1543,"dataFootnoteBackref":57},"#user-content-fnref-10","Back to reference 1",[1544],"data-footnote-backref","↩",[670,1547,1549,1550,916,1555,1559,1560],{"id":1548},"user-content-fn-11","Exploring different AI applications is perhaxps one of my favorite things about writing this book. It's a lot of fun seeing what people are building. You can ",[73,1551,1554],{"href":1552,"rel":1553},"https:\u002F\u002Fhuyenchip.com\u002Fllama-police",[111],"find the list of open source AI applications",[73,1556,1557],{"href":1557,"rel":1558},"https:\u002F\u002Fgoodailist.com\u002Frepos",[111],") that I track. The list is updated every 12 hours. ",[73,1561,1545],{"href":1562,"ariaLabel":1563,"className":1564,"dataFootnoteBackref":57},"#user-content-fnref-11","Back to reference 2",[1544],[670,1566,1568,1569,1574,1575,1578,1579],{"id":1567},"user-content-fn-12","Because enterprises usually spend a lot of money on ads and marketing, automation there can lead to huge savings. On average, 11% of a company's budget is spent on marketing. See ",[73,1570,1573],{"href":1571,"rel":1572},"https:\u002F\u002Fdeloitte.wsj.com\u002Fcmo\u002Fwho-has-the-biggest-marketing-budgets-1485234137",[111],"\"Marketing Budgets Vary by Industry\""," (Christine Moorman, ",[84,1576,1577],{},"WSJ",", 2017). ",[73,1580,1545],{"href":1581,"ariaLabel":1582,"className":1583,"dataFootnoteBackref":57},"#user-content-fnref-12","Back to reference 3",[1544],[670,1585,1587,1588],{"id":1586},"user-content-fn-13","I have found AI very helpful in the process of writing this book, and I can see that AI will be able to automate many parts of the writing process. When writing fiction, I often ask AI to brainstorm ideas on what it thinks will happen next or how a character might react to a situation. I'm still evaluating what kind of writing can be automated and what kind can't be. ",[73,1589,1545],{"href":1590,"ariaLabel":1591,"className":1592,"dataFootnoteBackref":57},"#user-content-fnref-13","Back to reference 4",[1544],[670,1594,1596,1597],{"id":1595},"user-content-fn-14","My hypothesis is that we'll become so distrustful of content on the internet that we'll only read content generated by people or brands we trust. ",[73,1598,1545],{"href":1599,"ariaLabel":1600,"className":1601,"dataFootnoteBackref":57},"#user-content-fnref-14","Back to reference 5",[1544],[670,1603,1605,1606],{"id":1604},"user-content-fn-15","It surprises me how long it takes Apple and Amazon to incorporate generative AI advances into Siri and Alexa. A friend thinks it's because these companies might have higher bars for quality and compliance, and it takes longer to develop voice interfaces than chat interfaces. ",[73,1607,1545],{"href":1608,"ariaLabel":1609,"className":1610,"dataFootnoteBackref":57},"#user-content-fnref-15","Back to reference 6",[1544],[670,1612,1614,1615],{"id":1613},"user-content-fn-16","Disclaimer: I'm an advisor of Convai. ",[73,1616,1545],{"href":1617,"ariaLabel":1618,"className":1619,"dataFootnoteBackref":57},"#user-content-fnref-16","Back to reference 7",[1544],[670,1621,1623,1624],{"id":1622},"user-content-fn-17","I currently have over 40,000 photos and videos in my Google Photos. Without AI, it'd be near impossible for me to search for the photos I want, when I want them. ",[73,1625,1545],{"href":1626,"ariaLabel":1627,"className":1628,"dataFootnoteBackref":57},"#user-content-fnref-17","Back to reference 8",[1544],[670,1630,1632,1633],{"id":1631},"user-content-fn-18","Personally, I also find AI good at explaining data and graphs. When encountering a confusing graph with too much information, I ask ChatGPT to break it down for me. ",[73,1634,1545],{"href":1635,"ariaLabel":1636,"className":1637,"dataFootnoteBackref":57},"#user-content-fnref-18","Back to reference 9",[1544],{"title":57,"searchDepth":1639,"depth":1639,"links":1640},2,[1641,1642,1643,1644,1645,1650,1654,1658,1662,1667,1671,1676,1679],{"id":90,"depth":1639,"text":91},{"id":173,"depth":1639,"text":174},{"id":344,"depth":1639,"text":345},{"id":549,"depth":1639,"text":550},{"id":606,"depth":1639,"text":397,"children":1646},[1647,1649],{"id":661,"depth":1648,"text":662},3,{"id":743,"depth":1648,"text":744},{"id":811,"depth":1639,"text":812,"children":1651},[1652,1653],{"id":851,"depth":1648,"text":852},{"id":876,"depth":1648,"text":877},{"id":900,"depth":1639,"text":420,"children":1655},[1656,1657],{"id":944,"depth":1648,"text":945},{"id":976,"depth":1648,"text":977},{"id":1042,"depth":1639,"text":437,"children":1659},[1660,1661],{"id":1068,"depth":1648,"text":1069},{"id":1118,"depth":1648,"text":1119},{"id":1174,"depth":1639,"text":1175,"children":1663},[1664,1665,1666],{"id":1181,"depth":1648,"text":1182},{"id":1222,"depth":1648,"text":1223},{"id":1233,"depth":1648,"text":1234},{"id":1306,"depth":1639,"text":1307,"children":1668},[1669,1670],{"id":1325,"depth":1648,"text":1326},{"id":1335,"depth":1648,"text":1336},{"id":1372,"depth":1639,"text":1373,"children":1672},[1673,1674,1675],{"id":1379,"depth":1648,"text":1380},{"id":1405,"depth":1648,"text":1406},{"id":1420,"depth":1648,"text":1421},{"id":1463,"depth":1639,"text":1464,"children":1677},[1678],{"id":1495,"depth":1648,"text":1496},{"id":77,"depth":1639,"text":1525},"A tour of industry-proven and promising use cases for foundation models — from coding and creative work to writing, education, chatbots, information aggregation, data organization, and workflow automation.","md",{},{"icon":29},{"title":26,"description":1680},"OcEmGUwIq1qd0_6nIaKSrUMAkJZAqnnHILoOGwN7bWk",[1687,1689],{"title":21,"path":22,"stem":23,"description":1688,"icon":24,"children":-1},"Trace how decades of advances in language models, self-supervision, and multimodality produced foundation models — and turned AI engineering into a discipline of its own.",{"title":31,"path":32,"stem":33,"description":1690,"icon":34,"children":-1},"How to evaluate use cases, build vs buy, set success metrics, plan milestones, and maintain AI products in a fast-moving landscape.",1778484800915]