[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"$f2wjzvs788ith5":3},{"_id":4,"slug":5,"title":6,"subtitle":7,"kind":8,"cards":9,"tags":57,"categories":59,"source":61,"lang":64,"author":65,"audioState":68,"stats":69,"publishedAt":72,"renderer":73},"6abac42cca21c797c7e9a433","what-the-hell-is-jev-and-why-does-it-matter-in-2027-0e651fce","What the Hell Is JEV? And Why Does It Matter in 2027?","Remember when we first started using LLMs in development?","news",[10,12,17,22,27,32,37,42,47,52],{"headline":6,"body":7,"imageUrl":11,"sourceImageUrl":11},"https:\u002F\u002Fmedia2.dev.to\u002Fdynamic\u002Fimage\u002Fwidth=1200,height=627,fit=cover,gravity=auto,format=auto\u002Fhttps%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2F7si8rsohjkfg8ijsvx2c.png",{"headline":13,"body":14,"imageUrl":15,"images":16},"A few years ago, getting an AI model","A few years ago, getting an AI model to understand a piece of text and generate a useful response already felt like magic. Then the models kept getting better. They started writing code, debugging applications, summarizing documents, reasoning through problems, calling tools, working with images, and eventually acting more like agents. AI slowly went from something we experimented with to something we could actually build products around.","\u002Fapi\u002Fmedia\u002Fposts\u002Fwhat-the-hell-is-jev-and-why-does-it-matter-in-2027-0e651fce\u002F1.webp",{"local":15},{"headline":18,"body":19,"imageUrl":20,"images":21},"And now we're at an interesting point. We've","And now we're at an interesting point. We've made these models incredibly capable at generating things. But here's the question I've started thinking about: Do we really need all that capability for every AI task inside our applications?","\u002Fapi\u002Fmedia\u002Fposts\u002Fwhat-the-hell-is-jev-and-why-does-it-matter-in-2027-0e651fce\u002F2.webp",{"local":20},{"headline":23,"body":24,"imageUrl":25,"images":26},"Because sometimes, our application doesn't need to return","Because sometimes, our application doesn't need to return essay every time. Sometimes we need an answer in just yes or no. We've Made LLMs Really, Really Good When large language models first became popular, the idea was pretty simple: Over the years, that changed dramatically.","\u002Fapi\u002Fmedia\u002Fposts\u002Fwhat-the-hell-is-jev-and-why-does-it-matter-in-2027-0e651fce\u002F3.webp",{"local":25},{"headline":28,"body":29,"imageUrl":30,"images":31},"Models got better at understanding context, following instructions","Models got better at understanding context, following instructions, writing code, reasoning through problems, using tools, working with images and eventually acting more like agents.","\u002Fapi\u002Fmedia\u002Fposts\u002Fwhat-the-hell-is-jev-and-why-does-it-matter-in-2027-0e651fce\u002F4.webp",{"local":30},{"headline":33,"body":34,"imageUrl":35,"images":36},"\"Analyze this document, search the web, write some","\"Analyze this document, search the web, write some code, call an API and figure out why my production deployment is broken.\"","\u002Fapi\u002Fmedia\u002Fposts\u002Fwhat-the-hell-is-jev-and-why-does-it-matter-in-2027-0e651fce\u002F5.webp",{"local":35},{"headline":38,"body":39,"imageUrl":40,"images":41},"It is pretty impressive but here's where things","It is pretty impressive but here's where things get interesting. Most applications don't need an AI model to solve a PhD-level problem every time they call one. Sometimes, we just need a small answer. For example: Does this transaction look suspicious? Does this transaction look suspicious? Should this user receive this offer? Should this user receive this offer? Which category does this product belong to? Which category does this product belong to? Does this customer look like they're going to churn? Does this customer look like they're going to churn? These aren't necessarily conversations. They're decisions. And that brings us to a question. Do We Really Need a Huge LLM For This? Imagine your application needs to answer: A traditional LLM can obviously do this.","\u002Fapi\u002Fmedia\u002Fposts\u002Fwhat-the-hell-is-jev-and-why-does-it-matter-in-2027-0e651fce\u002F6.webp",{"local":40},{"headline":43,"body":44,"imageUrl":45,"images":46},"But the model is still fundamentally a generative","But the model is still fundamentally a generative model. It is generating a response. Your application then has to deal with that response. Maybe it returns:","\u002Fapi\u002Fmedia\u002Fposts\u002Fwhat-the-hell-is-jev-and-why-does-it-matter-in-2027-0e651fce\u002F7.webp",{"local":45},{"headline":48,"body":49,"imageUrl":50,"images":51},"You can obviously solve a lot of this","You can obviously solve a lot of this with structured outputs, schemas and validation but the underlying question remains: Why are we using a system designed to generate language when our application only needs a predefined decision? What if the model was designed around the decision instead? That's where System One Models come into the picture. Think about how we normally use AI in an application.","\u002Fapi\u002Fmedia\u002Fposts\u002Fwhat-the-hell-is-jev-and-why-does-it-matter-in-2027-0e651fce\u002F8.webp",{"local":50},{"headline":53,"body":54,"imageUrl":55,"images":56},"We send some input, ask the model something","We send some input, ask the model something, and get a generated response back. But what if we already know the question our application needs to answer? That's the problem TypeSafe AI is trying to approach differently with System One Models.","\u002Fapi\u002Fmedia\u002Fposts\u002Fwhat-the-hell-is-jev-and-why-does-it-matter-in-2027-0e651fce\u002F9.webp",{"local":55},[58],"dev",[60],"Technology",{"name":62,"url":63},"Dev.to","https:\u002F\u002Fdev.to\u002Fsolitrix02\u002Fwhat-the-hell-is-jev-and-why-does-it-matter-in-2027-583a","en",{"handle":66,"displayName":67},"spots","Spots","queued",{"views":70,"likes":71,"saves":71,"shares":71,"completions":71,"opens":71,"skips":71,"depthSum":71},3,0,"2026-09-28T19:46:52.171Z","local"]