Why Developers Are Losing Their Minds Over AI That Can't Write

Jeb is like an LLM that can only talk in Jeb is like an LLM that can only talk in multiple choice. It can read whatever multiple choice. It can read whatever multiple choice. It can read whatever you hand it, but it can't write even one you hand it, but it can't write even one you hand it, but it can't write even one word back. So, you supply all the word back. So, you supply all the word back. So, you supply all the possible answers in advance and it's possible answers in advance and it's possible answers in advance and it's going to only pick one. And I know what going to only pick one. And I know what going to only pick one. And I know what that sounds like to you. It sounds that sounds like to you. It sounds that sounds like to you. It sounds useless. Why would anybody want a model useless. Why would anybody want a model useless. Why would anybody want a model that cannot write in the age when we that cannot write in the age when we that cannot write in the age when we have LLMs that are so so good at have LLMs that are so so good at have LLMs that are so so good at writing? Look, Typesafe shipped Jev on writing? Look, Typesafe shipped Jev on writing? Look, Typesafe shipped Jev on September 15th. And I'll tell you how September 15th. And I'll tell you how September 15th. And I'll tell you how useful it is because within 24 hours, it useful it is because within 24 hours, it useful it is because within 24 hours, it was already the fastest adopted model in was already the fastest adopted model in was already the fastest adopted model in Verscell's AI gateway history. Faster Verscell's AI gateway history. Faster Verscell's AI gateway history. Faster than any LLM from any lab. You name the than any LLM from any lab. You name the than any LLM from any lab. You name the lab, faster than that. More than twice lab, faster than that. More than twice lab, faster than that. More than twice as many paid teams have adopted and paid as many paid teams have adopted and paid as many paid teams have adopted and paid for that product for Jev as for any for that product for Jev as for any for that product for Jev as for any model ever launched before in the first model ever launched before in the first model ever launched before in the first 24 hours. So either a lot of really 24 hours. So either a lot of really 24 hours. So either a lot of really really good developers have suddenly really good developers have suddenly really good developers have suddenly lost their minds at the same time or the lost their minds at the same time or the lost their minds at the same time or the thing that Jev is refusing to do right thing that Jev is refusing to do right thing that Jev is refusing to do right words is the entire point. In this video words is the entire point. In this video words is the entire point. In this video I'm going to give you four things. I'm I'm going to give you four things. I'm I'm going to give you four things. I'm going to name this kind of problem so going to name this kind of problem so going to name this kind of problem so you can spot a Jevshaped problem in your you can spot a Jevshaped problem in your you can spot a Jevshaped problem in your work. I'm going to tell you the ways work. I'm going to tell you the ways work. I'm going to tell you the ways people are using Jev in real systems.

people are using Jev in real systems. people are using Jev in real systems. I'm going to tell you what it costs to I'm going to tell you what it costs to I'm going to tell you what it costs to try, which is so much less than you try, which is so much less than you try, which is so much less than you think. And I'm going to tell you places think. And I'm going to tell you places think. And I'm going to tell you places where it falls over and isn't perfect, where it falls over and isn't perfect, where it falls over and isn't perfect, because nothing's perfect. I'm not here because nothing's perfect. I'm not here because nothing's perfect. I'm not here to tell you that Jev replaces LLMs. And to tell you that Jev replaces LLMs. And to tell you that Jev replaces LLMs. And that's exactly why it matters so much. that's exactly why it matters so much. that's exactly why it matters so much. Let's get into it. Fundamentally, we've Let's get into it. Fundamentally, we've Let's get into it. Fundamentally, we've been missing a basic building block for been missing a basic building block for been missing a basic building block for the AI age that we're in, and we didn't the AI age that we're in, and we didn't the AI age that we're in, and we didn't even know it. we've been missing a even know it. we've been missing a even know it. we've been missing a generalpurpose classifier that's really generalpurpose classifier that's really generalpurpose classifier that's really really fast. And now that it's here, it really fast. And now that it's here, it really fast. And now that it's here, it looks obvious. And by the way, a looks obvious. And by the way, a looks obvious. And by the way, a classifier, that's just a fancy word for classifier, that's just a fancy word for classifier, that's just a fancy word for picking multiple choice. So if you hear picking multiple choice. So if you hear picking multiple choice. So if you hear that, that's what that is. Like think of that, that's what that is. Like think of that, that's what that is. Like think of it as I have to pick an answer on my SAT it as I have to pick an answer on my SAT it as I have to pick an answer on my SAT test. We've been missing a generalurpose test. We've been missing a generalurpose test. We've been missing a generalurpose classifier that's fast enough, cheap classifier that's fast enough, cheap classifier that's fast enough, cheap enough, and capable enough to put enough, and capable enough to put enough, and capable enough to put judgment into all kinds of places where judgment into all kinds of places where judgment into all kinds of places where it was previously way too expensive to it was previously way too expensive to it was previously way too expensive to justify. But to understand what that justify. But to understand what that justify. But to understand what that means, we're going to go back in time. means, we're going to go back in time. means, we're going to go back in time. Let's start with something software has Let's start with something software has Let's start with something software has always been good at, something that always been good at, something that always been good at, something that they've been good at for decades. And they've been good at for decades. And they've been good at for decades. And you'll see where I'm going here. As an you'll see where I'm going here. As an you'll see where I'm going here. As an example, let's say an invoice is 30 days example, let's say an invoice is 30 days example, let's say an invoice is 30 days overdue. You need to flag it. If an overdue. You need to flag it. If an overdue. You need to flag it. If an order total exceeds a limit, you have a order total exceeds a limit, you have a order total exceeds a limit, you have a rule to flag that. You get the idea.

rule to flag that. You get the idea. rule to flag that. You get the idea. These are what we call deterministic These are what we call deterministic These are what we call deterministic rules. The same thing happens every rules. The same thing happens every rules. The same thing happens every time. You can write the condition down time. You can write the condition down time. You can write the condition down really precisely for that rule. And you really precisely for that rule. And you really precisely for that rule. And you can use ordinary code to make that can use ordinary code to make that can use ordinary code to make that happen. But now, let's add a complexity. happen. But now, let's add a complexity. happen. But now, let's add a complexity. Let's add something that code doesn't Let's add something that code doesn't Let's add something that code doesn't cover. Let's ask, does the customer cover. Let's ask, does the customer cover. Let's ask, does the customer sound like they're about to leave the sound like they're about to leave the sound like they're about to leave the contract in this email? Does this email contract in this email? Does this email contract in this email? Does this email contain a serious business opportunity contain a serious business opportunity contain a serious business opportunity or does it just have the words business or does it just have the words business or does it just have the words business opportunity in it? These are things that opportunity in it? These are things that opportunity in it? These are things that we as humans really really care about we as humans really really care about we as humans really really care about and they're things that I've seen AI or and they're things that I've seen AI or and they're things that I've seen AI or LLM used for today. But it turns out LLM used for today. But it turns out LLM used for today. But it turns out that they're classifier problems because that they're classifier problems because that they're classifier problems because what we're really doing is we're reading what we're really doing is we're reading what we're really doing is we're reading complicated text and we're outputting an complicated text and we're outputting an complicated text and we're outputting an action or a simple choice. And that action or a simple choice. And that action or a simple choice. And that turns out to be a huge number of turns out to be a huge number of turns out to be a huge number of problems in our world. These are problems in our world. These are problems in our world. These are problems that we want solved as people. problems that we want solved as people. problems that we want solved as people. And we've been using LLMs to solve them And we've been using LLMs to solve them And we've been using LLMs to solve them relatively expensively because they are relatively expensively because they are relatively expensively because they are so valuable. What happens if you can so valuable. What happens if you can so valuable. What happens if you can solve them cheaply? Jev gets a what you solve them cheaply? Jev gets a what you solve them cheaply? Jev gets a what you do when you have text that you want to do when you have text that you want to do when you have text that you want to apply complicated questions to, but you apply complicated questions to, but you apply complicated questions to, but you can get a simple set of output options.

can get a simple set of output options. can get a simple set of output options. So you might want a category, you might So you might want a category, you might So you might want a category, you might want a score, you might want a yes or a want a score, you might want a yes or a want a score, you might want a yes or a no to proceed. But to get there, you no to proceed. But to get there, you no to proceed. But to get there, you have to actually deeply understand all have to actually deeply understand all have to actually deeply understand all of the text that came before. Stuff that of the text that came before. Stuff that of the text that came before. Stuff that requires interpretation, that requires requires interpretation, that requires requires interpretation, that requires judgment, right? Customers express judgment, right? Customers express judgment, right? Customers express dissatisfaction in different ways. If we dissatisfaction in different ways. If we dissatisfaction in different ways. If we go back to our customer example, let's go back to our customer example, let's go back to our customer example, let's say a useful opportunity, it arrives in say a useful opportunity, it arrives in say a useful opportunity, it arrives in a badly written email. Can Jev find a badly written email. Can Jev find a badly written email. Can Jev find that? Certainly that's challenging to that? Certainly that's challenging to that? Certainly that's challenging to find with rules and it's somewhat find with rules and it's somewhat find with rules and it's somewhat expensive to find with AI and not always expensive to find with AI and not always expensive to find with AI and not always reliable. And so Jev gives us a tool to reliable. And so Jev gives us a tool to reliable. And so Jev gives us a tool to solve problems that we were solving not solve problems that we were solving not solve problems that we were solving not at all or expensively before. This is at all or expensively before. This is at all or expensively before. This is actually really relevant in the age of actually really relevant in the age of actually really relevant in the age of AI safety because right now agents AI safety because right now agents AI safety because right now agents typically have to ask either humans or typically have to ask either humans or typically have to ask either humans or LLMs if it's safe to proceed on an LLMs if it's safe to proceed on an LLMs if it's safe to proceed on an action. Jev gives us another tool action. Jev gives us another tool action. Jev gives us another tool because it can understand an incredibly because it can understand an incredibly because it can understand an incredibly complicated piece of agent text and code complicated piece of agent text and code complicated piece of agent text and code and then make a decision. Yes, you can and then make a decision. Yes, you can and then make a decision. Yes, you can proceed. No, you can't or you should ask proceed. No, you can't or you should ask proceed. No, you can't or you should ask a human. And Jev gives us a tool to do a human. And Jev gives us a tool to do a human. And Jev gives us a tool to do that very cheaply, very quickly, which that very cheaply, very quickly, which that very cheaply, very quickly, which means we can apply it over a much wider means we can apply it over a much wider means we can apply it over a much wider part of the agent run. The work I'm part of the agent run. The work I'm part of the agent run. The work I'm describing is what I am calling describing is what I am calling describing is what I am calling semideterministic.

semideterministic. semideterministic. It's a really big word, but I think It's a really big word, but I think It's a really big word, but I think you're going to start to get it by the you're going to start to get it by the you're going to start to get it by the end of this video. It's about the glue end of this video. It's about the glue end of this video. It's about the glue between language and choice. When we use between language and choice. When we use between language and choice. When we use a tool like Jeff, we or an agent can a tool like Jeff, we or an agent can a tool like Jeff, we or an agent can define a short list of available define a short list of available define a short list of available outcomes and what happens after each outcomes and what happens after each outcomes and what happens after each one. And then we can put an incredibly one. And then we can put an incredibly one. And then we can put an incredibly complicated input into Jev and ask Jev complicated input into Jev and ask Jev complicated input into Jev and ask Jev to sort. And so, yeah, judgment does to sort. And so, yeah, judgment does to sort. And so, yeah, judgment does remain probabilistic. We don't write the remain probabilistic. We don't write the remain probabilistic. We don't write the specific rules. The judgment remains in specific rules. The judgment remains in specific rules. The judgment remains in the realm. We've traditionally used LLMs the realm. We've traditionally used LLMs the realm. We've traditionally used LLMs to do tools like cloud or chat GPT and to do tools like cloud or chat GPT and to do tools like cloud or chat GPT and so judgment is still happening through a so judgment is still happening through a so judgment is still happening through a transformer architecture but because the transformer architecture but because the transformer architecture but because the output is so structured and because Jev output is so structured and because Jev output is so structured and because Jev was trained specifically for this kind was trained specifically for this kind was trained specifically for this kind of generalpurpose decisioning we get it of generalpurpose decisioning we get it of generalpurpose decisioning we get it really fast really really cheap. Now if really fast really really cheap. Now if really fast really really cheap. Now if you're a nerd like me this kind of you're a nerd like me this kind of you're a nerd like me this kind of problem is a very old kind of problem.

problem is a very old kind of problem. problem is a very old kind of problem. We've had machine learningdriven We've had machine learningdriven We've had machine learningdriven classifiers in computer science for a classifiers in computer science for a classifiers in computer science for a really, really long time. I've used them really, really long time. I've used them really, really long time. I've used them myself. I used them a ton when I was at myself. I used them a ton when I was at myself. I used them a ton when I was at Amazon. And they work like this. You Amazon. And they work like this. You Amazon. And they work like this. You collect examples, you label your data, collect examples, you label your data, collect examples, you label your data, you train or fine-tune a model, you you train or fine-tune a model, you you train or fine-tune a model, you evaluate it, and then you maintain your evaluate it, and then you maintain your evaluate it, and then you maintain your classifiers as the problem changes. And classifiers as the problem changes. And classifiers as the problem changes. And once you've built a really good one, it once you've built a really good one, it once you've built a really good one, it becomes extremely cheap to run that becomes extremely cheap to run that becomes extremely cheap to run that classifier. And the investment is often classifier. And the investment is often classifier. And the investment is often in getting the right classifier and the in getting the right classifier and the in getting the right classifier and the right data for the right problem. Until right data for the right problem. Until right data for the right problem. Until now, they've pretty much all been now, they've pretty much all been now, they've pretty much all been custom. Every new category you care custom. Every new category you care custom. Every new category you care about, if you wanted to do ML about, if you wanted to do ML about, if you wanted to do ML classifiers, you would have to do data classifiers, you would have to do data classifiers, you would have to do data work. You'd have to have specialized work. You'd have to have specialized work. You'd have to have specialized expertise. You'd have to have more expertise. You'd have to have more expertise. You'd have to have more evaluation. And all of that investment, evaluation. And all of that investment, evaluation. And all of that investment, of course, only makes sense when you're of course, only makes sense when you're of course, only makes sense when you're a giant stable platform and you have a giant stable platform and you have a giant stable platform and you have lots of investments to make and you have lots of investments to make and you have lots of investments to make and you have to scale and all of that. to scale and all of that. to scale and all of that. And you're not ever going to justify the And you're not ever going to justify the And you're not ever going to justify the millions and billions of tiny classifier millions and billions of tiny classifier millions and billions of tiny classifier type judgments that are scattered type judgments that are scattered type judgments that are scattered through all of our digital products, our through all of our digital products, our through all of our digital products, our digital lives, and our digital digital lives, and our digital digital lives, and our digital businesses. It's just too expensive.

businesses. It's just too expensive. businesses. It's just too expensive. You don't have enough data. It never is You don't have enough data. It never is You don't have enough data. It never is going to work. Now, LLMs are a part of going to work. Now, LLMs are a part of going to work. Now, LLMs are a part of how that story started to change. This how that story started to change. This how that story started to change. This is where we're at in 2026. LLM have is where we're at in 2026. LLM have is where we're at in 2026. LLM have given us a model that can reason with given us a model that can reason with given us a model that can reason with language. It has enough general language. It has enough general language. It has enough general knowledge to handle many classification knowledge to handle many classification knowledge to handle many classification tasks just from a description. And we do tasks just from a description. And we do tasks just from a description. And we do use them for that purpose today. We've use them for that purpose today. We've use them for that purpose today. We've started asking LLM to sort messages, to started asking LLM to sort messages, to started asking LLM to sort messages, to choose tools, to judge relevance, to choose tools, to judge relevance, to choose tools, to judge relevance, to screen documents, to check other models screen documents, to check other models screen documents, to check other models work. I know I've done that. I'm I bet work. I know I've done that. I'm I bet work. I know I've done that. I'm I bet you've done some of that, too. I've you've done some of that, too. I've you've done some of that, too. I've certainly seen examples of all of those, certainly seen examples of all of those, certainly seen examples of all of those, not just in my computer, but in not just in my computer, but in not just in my computer, but in production systems. We wrap up those production systems. We wrap up those production systems. We wrap up those answers in tool calls when we get fancy. answers in tool calls when we get fancy. answers in tool calls when we get fancy. We build agents around them. Harnesses We build agents around them. Harnesses We build agents around them. Harnesses are built around this kind of problem. are built around this kind of problem. are built around this kind of problem. Now ask yourself, how often have we been Now ask yourself, how often have we been Now ask yourself, how often have we been using an LLM, a text generating token using an LLM, a text generating token using an LLM, a text generating token architecture as a generalpurpose architecture as a generalpurpose architecture as a generalpurpose classifier? classifier? classifier? I've literally asked LLMs to look I've literally asked LLMs to look I've literally asked LLMs to look through my inbox and now I one of my through my inbox and now I one of my through my inbox and now I one of my first projects I am building a Jev tool first projects I am building a Jev tool first projects I am building a Jev tool that will classify my inbox for me. And that will classify my inbox for me. And that will classify my inbox for me. And that's not the only example. If we go that's not the only example. If we go that's not the only example. If we go back to the safety example, top of mind, back to the safety example, top of mind, back to the safety example, top of mind, can agents go rogue, etc. I can build can agents go rogue, etc. I can build can agents go rogue, etc. I can build something in Jev that reviews proposed something in Jev that reviews proposed something in Jev that reviews proposed commands to agents before they execute commands to agents before they execute commands to agents before they execute and just gives a thumbs up or a thumbs and just gives a thumbs up or a thumbs and just gives a thumbs up or a thumbs down. Or support systems. I can build down. Or support systems. I can build down. Or support systems. I can build something that asks a model which of my

something that asks a model which of my something that asks a model which of my teams should receive a ticket based on a teams should receive a ticket based on a teams should receive a ticket based on a complex assessment of the actual risk to complex assessment of the actual risk to complex assessment of the actual risk to the customer. I can and have built these the customer. I can and have built these the customer. I can and have built these systems. Those are substantial parts of systems. Those are substantial parts of systems. Those are substantial parts of what we would now call AI workflows. And what we would now call AI workflows. And what we would now call AI workflows. And we've used LLMs for them in the past. we've used LLMs for them in the past. we've used LLMs for them in the past. And that's the tool that we had at the And that's the tool that we had at the And that's the tool that we had at the time, right? Jev is like a new Lego time, right? Jev is like a new Lego time, right? Jev is like a new Lego brick in building software. Jev brings brick in building software. Jev brings brick in building software. Jev brings that broad ability to interpret language that broad ability to interpret language that broad ability to interpret language that we love so much about LLMs and it that we love so much about LLMs and it that we love so much about LLMs and it brings it into a model designed brings it into a model designed brings it into a model designed specifically to put out really, really specifically to put out really, really specifically to put out really, really structured decisions. And when you do structured decisions. And when you do structured decisions. And when you do that, you get a tremendous advantage on that, you get a tremendous advantage on that, you get a tremendous advantage on cost and speed. Like I'm talking cost and speed. Like I'm talking cost and speed. Like I'm talking multiple orders of magnitude. So you can multiple orders of magnitude. So you can multiple orders of magnitude. So you can describe the question. You can describe describe the question. You can describe describe the question. You can describe all of the answers that are allowed. And all of the answers that are allowed. And all of the answers that are allowed. And it returns choices and scores and it returns choices and scores and it returns choices and scores and probabilities to the program. And again, probabilities to the program. And again, probabilities to the program. And again, I know that sounds silly, but that turns I know that sounds silly, but that turns I know that sounds silly, but that turns out to be incredibly powerful because we out to be incredibly powerful because we out to be incredibly powerful because we do it so much in our compute do it so much in our compute do it so much in our compute environments. And so if you're environments. And so if you're environments. And so if you're wondering, this really sounds wondering, this really sounds wondering, this really sounds complicated. I as a human, I'm not going complicated. I as a human, I'm not going complicated. I as a human, I'm not going to be able to describe these choices to be able to describe these choices to be able to describe these choices correctly. I got you. Guess what? It is correctly. I got you. Guess what? It is correctly. I got you. Guess what? It is built for agents to use for you. And built for agents to use for you. And built for agents to use for you. And we're going to get into that. And if we're going to get into that. And if we're going to get into that. And if you're wondering, can it evaluate you're wondering, can it evaluate you're wondering, can it evaluate multiple questions at once? The answer multiple questions at once? The answer multiple questions at once? The answer is yes. Several questions can be is yes. Several questions can be is yes. Several questions can be evaluated together. Jev can also be evaluated together. Jev can also be evaluated together. Jev can also be applied to really different kinds of applied to really different kinds of applied to really different kinds of classification problems without training classification problems without training classification problems without training a new one each time. That's part of why a new one each time. That's part of why a new one each time. That's part of why I call it a general purpose classifier.

I call it a general purpose classifier. I call it a general purpose classifier. I used to have to do specialized I used to have to do specialized I used to have to do specialized training when I was at Prime Video and training when I was at Prime Video and training when I was at Prime Video and other places. Others have had to do that other places. Others have had to do that other places. Others have had to do that as well. Now, if you're trying to as well. Now, if you're trying to as well. Now, if you're trying to understand where can I use Jev, and this understand where can I use Jev, and this understand where can I use Jev, and this is something we'll get into later in the is something we'll get into later in the is something we'll get into later in the video, you need to test it in a place video, you need to test it in a place video, you need to test it in a place where you plausibly have that kind of where you plausibly have that kind of where you plausibly have that kind of complicated text in and that simple complicated text in and that simple complicated text in and that simple choice out. Find those spots. And agents choice out. Find those spots. And agents choice out. Find those spots. And agents can help you find it, and I'll get to can help you find it, and I'll get to can help you find it, and I'll get to that later in the video. And then you that later in the video. And then you that later in the video. And then you need to see is it better than an LLM? Is need to see is it better than an LLM? Is need to see is it better than an LLM? Is it better than a machine learning it better than a machine learning it better than a machine learning classifier I already have going? Is it classifier I already have going? Is it classifier I already have going? Is it better than me choosing individually? better than me choosing individually? better than me choosing individually? And by the way, I'm going to give you a And by the way, I'm going to give you a And by the way, I'm going to give you a cheat code right now. It is almost cheat code right now. It is almost cheat code right now. It is almost certainly going to be cheaper than an certainly going to be cheaper than an certainly going to be cheaper than an LLM. It's going to be much cheaper than LLM. It's going to be much cheaper than LLM. It's going to be much cheaper than you. And so I I think in that sense, in you. And so I I think in that sense, in you. And so I I think in that sense, in most cases, you're going to want to use most cases, you're going to want to use most cases, you're going to want to use Jeff, but you'll have to test it and Jeff, but you'll have to test it and Jeff, but you'll have to test it and see. And we'll get into how you do that see. And we'll get into how you do that see. And we'll get into how you do that and where the weak spots are and all of and where the weak spots are and all of and where the weak spots are and all of that later in the video. I have looked that later in the video. I have looked that later in the video. I have looked across hundreds of use cases just been a across hundreds of use cases just been a across hundreds of use cases just been a few days. It's going to work out pretty few days. It's going to work out pretty few days. It's going to work out pretty well for you in most of those cases. well for you in most of those cases. well for you in most of those cases. They are ranging from biosciences to They are ranging from biosciences to They are ranging from biosciences to e-commerce, right? From from personal e-commerce, right? From from personal e-commerce, right? From from personal productivity all the way up to how we productivity all the way up to how we productivity all the way up to how we design generative interfaces. There are design generative interfaces. There are design generative interfaces. There are so many applications here and it just so many applications here and it just so many applications here and it just seems to work really well because it seems to work really well because it seems to work really well because it turns out in the computing world a lot turns out in the computing world a lot turns out in the computing world a lot of problems are shaped like complicated of problems are shaped like complicated of problems are shaped like complicated in simple out. And that combination of in simple out. And that combination of in simple out. And that combination of it understands language and it makes it understands language and it makes it understands language and it makes choices is why I think this new Lego choices is why I think this new Lego choices is why I think this new Lego brick is so important that we would call brick is so important that we would call brick is so important that we would call it a new primitive in AI architecture. a

it a new primitive in AI architecture. a it a new primitive in AI architecture. a primitive. It's just a fancy word for a primitive. It's just a fancy word for a primitive. It's just a fancy word for a building block that other people can building block that other people can building block that other people can compose into products. And it's compose into products. And it's compose into products. And it's technical speak, but it's not that hard technical speak, but it's not that hard technical speak, but it's not that hard to understand. So, we're talking about to understand. So, we're talking about to understand. So, we're talking about making the ability to understand making the ability to understand making the ability to understand language available at the level of an language available at the level of an language available at the level of an ordinary software function in a way that ordinary software function in a way that ordinary software function in a way that ties language into simple choice. And ties language into simple choice. And ties language into simple choice. And that's why it's such a big deal. that's why it's such a big deal. that's why it's such a big deal. Developers can now start putting Jev and Developers can now start putting Jev and Developers can now start putting Jev and classifiers all the way through software classifiers all the way through software classifiers all the way through software they already know how to build. And they they already know how to build. And they they already know how to build. And they make things so much faster and so much make things so much faster and so much make things so much faster and so much smarter because Jev itself is fast and smarter because Jev itself is fast and smarter because Jev itself is fast and it's smart and it's cheap. Really before it's smart and it's cheap. Really before it's smart and it's cheap. Really before now before the 15th of September when now before the 15th of September when now before the 15th of September when Jev launched you had two capabilities to Jev launched you had two capabilities to Jev launched you had two capabilities to build software with. You had build software with. You had build software with. You had deterministic code and you had LLMs. Now deterministic code and you had LLMs. Now deterministic code and you had LLMs. Now you have three. You have LLMs, you have three. You have LLMs, you have three. You have LLMs, generative models that can reason and generative models that can reason and generative models that can reason and develop plans that can use symbolic develop plans that can use symbolic develop plans that can use symbolic language that can write. You have a language that can write. You have a language that can write. You have a general purpose classifier that's Jev general purpose classifier that's Jev general purpose classifier that's Jev that can interpret any situation and that can interpret any situation and that can interpret any situation and choose very reliably among defined choose very reliably among defined choose very reliably among defined outcomes. You have deterministic tools outcomes. You have deterministic tools outcomes. You have deterministic tools or code that can calculate and retrieve or code that can calculate and retrieve or code that can calculate and retrieve records. We've been familiar with that records. We've been familiar with that records. We've been familiar with that one for 80 years, right? And so one for 80 years, right? And so one for 80 years, right? And so ultimately the history of software is ultimately the history of software is ultimately the history of software is the history of finding these new the history of finding these new the history of finding these new building blocks. When chat GPT launched building blocks. When chat GPT launched building blocks. When chat GPT launched we got that new building block, the we got that new building block, the we got that new building block, the first new building block in decades for first new building block in decades for first new building block in decades for software. Now we have a third one.

software. Now we have a third one. software. Now we have a third one. That's how big this is. One of the That's how big this is. One of the That's how big this is. One of the biggest events in the history of biggest events in the history of biggest events in the history of computer science in my view, and I'm computer science in my view, and I'm computer science in my view, and I'm just going to stand on that. I'm just just going to stand on that. I'm just just going to stand on that. I'm just going to say it. The question that all going to say it. The question that all going to say it. The question that all of us have to ask now is where does this of us have to ask now is where does this of us have to ask now is where does this belong in the software we use every day? belong in the software we use every day? belong in the software we use every day? The examples that people are building The examples that people are building The examples that people are building suggest several different arrangements suggest several different arrangements suggest several different arrangements that you can use in your own product. that you can use in your own product. that you can use in your own product. And what I want to do is I want to go And what I want to do is I want to go And what I want to do is I want to go through them quickly here because I need through them quickly here because I need through them quickly here because I need you to understand and have some examples you to understand and have some examples you to understand and have some examples of how the Legos fit together so that we of how the Legos fit together so that we of how the Legos fit together so that we together get a sense of what we can together get a sense of what we can together get a sense of what we can expect and how to use this new tool in expect and how to use this new tool in expect and how to use this new tool in our arsenal. So let's jump into how I our arsenal. So let's jump into how I our arsenal. So let's jump into how I see us using Jev. The first see us using Jev. The first see us using Jev. The first architectural arrangement that I'm architectural arrangement that I'm architectural arrangement that I'm seeing puts a classifier like Jev in seeing puts a classifier like Jev in seeing puts a classifier like Jev in between messy incoming information and between messy incoming information and between messy incoming information and software that knows what to do next. The software that knows what to do next. The software that knows what to do next. The classifier's job is to determine what classifier's job is to determine what classifier's job is to determine what this complicated input means with this this complicated input means with this this complicated input means with this limited information set. So your code limited information set. So your code limited information set. So your code can then use that result to choose an can then use that result to choose an can then use that result to choose an existing process. Let me give you just existing process. Let me give you just existing process. Let me give you just three examples that make that three examples that make that three examples that make that relationship super easy to understand.

relationship super easy to understand. relationship super easy to understand. Let's say a support ticket comes in. Jev Let's say a support ticket comes in. Jev Let's say a support ticket comes in. Jev is going to select a category for that is going to select a category for that is going to select a category for that ticket. It chooses billing. It judges ticket. It chooses billing. It judges ticket. It chooses billing. It judges that it should be answered today and it that it should be answered today and it that it should be answered today and it flags the risk of losing a customer. It flags the risk of losing a customer. It flags the risk of losing a customer. It happens very fast. Jev handles that kind happens very fast. Jev handles that kind happens very fast. Jev handles that kind of in between job, that that of in between job, that that of in between job, that that categorization job. And the decisions categorization job. And the decisions categorization job. And the decisions Jev makes then determine who gets the Jev makes then determine who gets the Jev makes then determine who gets the ticket on which team, how quickly it ticket on which team, how quickly it ticket on which team, how quickly it gets attention, and maybe it kicks off gets attention, and maybe it kicks off gets attention, and maybe it kicks off like an actually write the email or like an actually write the email or like an actually write the email or write the support ticket response, etc. write the support ticket response, etc. write the support ticket response, etc. And this works for tickets, it works for And this works for tickets, it works for And this works for tickets, it works for email. For example, if Jeff reads an email. For example, if Jeff reads an email. For example, if Jeff reads an email, it can identify the inquiry is a email, it can identify the inquiry is a email, it can identify the inquiry is a marketing opportunity. It can then marketing opportunity. It can then marketing opportunity. It can then categorize that as worth a reply and it categorize that as worth a reply and it categorize that as worth a reply and it can estimate the size of the can estimate the size of the can estimate the size of the opportunity. A model can then draft the opportunity. A model can then draft the opportunity. A model can then draft the response. But the decision to put response. But the decision to put response. But the decision to put attention there to to make the LLM do attention there to to make the LLM do attention there to to make the LLM do that that already got decided by Jev. that that already got decided by Jev. that that already got decided by Jev. Okay, let's hop into the software world Okay, let's hop into the software world Okay, let's hop into the software world for just a second. Let's say you're for just a second. Let's say you're for just a second. Let's say you're using an agent for code. An agent might using an agent for code. An agent might using an agent for code. An agent might propose deleting a build folder and propose deleting a build folder and propose deleting a build folder and force pushing code. These are relatively force pushing code. These are relatively force pushing code. These are relatively dangerous actions. Those are dramatic.

dangerous actions. Those are dramatic. dangerous actions. Those are dramatic. Jev can recommend asking the user before Jev can recommend asking the user before Jev can recommend asking the user before proceeding. It can do so really proceeding. It can do so really proceeding. It can do so really efficiently and really quickly. To be efficiently and really quickly. To be efficiently and really quickly. To be extremely clear, I am not saying that extremely clear, I am not saying that extremely clear, I am not saying that LLMs are done. Quite the opposite. But I LLMs are done. Quite the opposite. But I LLMs are done. Quite the opposite. But I want you to understand first what Jev want you to understand first what Jev want you to understand first what Jev does so you can see where LLMs now have does so you can see where LLMs now have does so you can see where LLMs now have a growing role. And to understand the a growing role. And to understand the a growing role. And to understand the impact of swapping out Jev for LLMs on impact of swapping out Jev for LLMs on impact of swapping out Jev for LLMs on the same job, I'm going to give you an the same job, I'm going to give you an the same job, I'm going to give you an example from X. Uh developer Nakshhatra example from X. Uh developer Nakshhatra example from X. Uh developer Nakshhatra Sakena described using an existing Sakena described using an existing Sakena described using an existing pipeline for a business that already pipeline for a business that already pipeline for a business that already ingests thousands of tax documents and ingests thousands of tax documents and ingests thousands of tax documents and has been doing so via AI, via LLM. As has been doing so via AI, via LLM. As has been doing so via AI, via LLM. As soon as Jev came out, he built a Jev soon as Jev came out, he built a Jev soon as Jev came out, he built a Jev classifier to do that same work. And he classifier to do that same work. And he classifier to do that same work. And he published a comparison. What's the cost? published a comparison. What's the cost? published a comparison. What's the cost? What's the speed? like this is what we What's the speed? like this is what we What's the speed? like this is what we do in a business. He is reporting 34x do in a business. He is reporting 34x do in a business. He is reporting 34x less cost to do the job and six times less cost to do the job and six times less cost to do the job and six times faster. That's a massive gain. You faster. That's a massive gain. You faster. That's a massive gain. You already have the business process. You already have the business process. You already have the business process. You have to classify your tax docs. Look at have to classify your tax docs. Look at have to classify your tax docs. Look at that. In other words, Jev is cheaply, that. In other words, Jev is cheaply, that. In other words, Jev is cheaply, efficiently, and quickly working out efficiently, and quickly working out efficiently, and quickly working out what questions are worth asking. One what questions are worth asking. One what questions are worth asking. One more example I'm going to give here is more example I'm going to give here is more example I'm going to give here is from Zack Gone. Zach reported sorting from Zack Gone. Zach reported sorting from Zack Gone. Zach reported sorting 20,000 of his own emails, his own 20,000 of his own emails, his own 20,000 of his own emails, his own Slacks, his own transcripts about like Slacks, his own transcripts about like Slacks, his own transcripts about like complaints, upsells, missed follow-ups.

complaints, upsells, missed follow-ups. complaints, upsells, missed follow-ups. He's an entrepreneur, right? And he said He's an entrepreneur, right? And he said He's an entrepreneur, right? And he said it took 7 minutes to go through all of it took 7 minutes to go through all of it took 7 minutes to go through all of those 20,000 plus pieces of information. those 20,000 plus pieces of information. those 20,000 plus pieces of information. It cost him a total of, wait for it, $1. It cost him a total of, wait for it, $1. It cost him a total of, wait for it, $1. And I'm giving you the exact dollar And I'm giving you the exact dollar And I'm giving you the exact dollar comparisons here because that's what comparisons here because that's what comparisons here because that's what made my jaw hit. And the cheapness opens made my jaw hit. And the cheapness opens made my jaw hit. And the cheapness opens up new possibilities. Okay, one more up new possibilities. Okay, one more up new possibilities. Okay, one more example that illustrates. We've talked example that illustrates. We've talked example that illustrates. We've talked about speed. We've talked about about speed. We've talked about about speed. We've talked about cheapness of Jev. Let's talk about cheapness of Jev. Let's talk about cheapness of Jev. Let's talk about scale. Durya Unutz is a bioscientist and scale. Durya Unutz is a bioscientist and scale. Durya Unutz is a bioscientist and he's deep on AI. And when Jev came out, he's deep on AI. And when Jev came out, he's deep on AI. And when Jev came out, Durya took the scientific research Durya took the scientific research Durya took the scientific research problem he cares about, which is in the problem he cares about, which is in the problem he cares about, which is in the realm of immunology and he set up an realm of immunology and he set up an realm of immunology and he set up an experiment for Jev. What he asked is he experiment for Jev. What he asked is he experiment for Jev. What he asked is he asked Jev to select the top 100 most asked Jev to select the top 100 most asked Jev to select the top 100 most important immunology questions from important immunology questions from important immunology questions from 10,000 10,000 10,000 different literature grounded different literature grounded different literature grounded candidates. And in other words, he he candidates. And in other words, he he candidates. And in other words, he he had it grounded in the literature had it grounded in the literature had it grounded in the literature because he's a scientist and these were because he's a scientist and these were because he's a scientist and these were all the immunology questions he cares all the immunology questions he cares all the immunology questions he cares about. He had a list of 10,000. He said, about. He had a list of 10,000. He said, about. He had a list of 10,000. He said, "Jeb, please pick the right top 100."

"Jeb, please pick the right top 100." "Jeb, please pick the right top 100." Now, immunology questions are super Now, immunology questions are super Now, immunology questions are super different from customer service different from customer service different from customer service questions like we've been discussing. It questions like we've been discussing. It questions like we've been discussing. It illustrates Jev's range here. Jev was illustrates Jev's range here. Jev was illustrates Jev's range here. Jev was able to reliably come up with what he able to reliably come up with what he able to reliably come up with what he views as about the right top 100 from views as about the right top 100 from views as about the right top 100 from that 10,000 and again did it quickly and that 10,000 and again did it quickly and that 10,000 and again did it quickly and cheaply. And 10,000 immunology cheaply. And 10,000 immunology cheaply. And 10,000 immunology questions, it's a big corpus. It's a lot questions, it's a big corpus. It's a lot questions, it's a big corpus. It's a lot of stuff. These are not short questions. of stuff. These are not short questions. of stuff. These are not short questions. Now, I'm not saying that Jev can look at Now, I'm not saying that Jev can look at Now, I'm not saying that Jev can look at the immunology questions and immediately the immunology questions and immediately the immunology questions and immediately solve them. That's an LLMshaped problem. solve them. That's an LLMshaped problem. solve them. That's an LLMshaped problem. I am saying that if you want to figure I am saying that if you want to figure I am saying that if you want to figure out where to apply your attention, I out where to apply your attention, I out where to apply your attention, I think it's a good illustration and as think it's a good illustration and as think it's a good illustration and as information grows and grows and grows information grows and grows and grows information grows and grows and grows faster and faster in the age of AI, faster and faster in the age of AI, faster and faster in the age of AI, getting attention right is incredibly getting attention right is incredibly getting attention right is incredibly valuable. So, we've walked through two valuable. So, we've walked through two valuable. So, we've walked through two big patterns for Jev, right? Jev is a big patterns for Jev, right? Jev is a big patterns for Jev, right? Jev is a shim where you can see information calls shim where you can see information calls shim where you can see information calls out. And we've also looked at Jev as out. And we've also looked at Jev as out. And we've also looked at Jev as kind of a a figure outer of big problem kind of a a figure outer of big problem kind of a a figure outer of big problem spaces like the immunology example.

spaces like the immunology example. spaces like the immunology example. Something that helps us pick out what Something that helps us pick out what Something that helps us pick out what matters from this giant pile. The third matters from this giant pile. The third matters from this giant pile. The third architectural pattern that I am seeing architectural pattern that I am seeing architectural pattern that I am seeing actually reverses who is driving work in actually reverses who is driving work in actually reverses who is driving work in AI systems. So in this case, we're AI systems. So in this case, we're AI systems. So in this case, we're seeing patterns where Jev is going to seeing patterns where Jev is going to seeing patterns where Jev is going to choose the next step instead of an AI. choose the next step instead of an AI. choose the next step instead of an AI. So James Ward described this as putting So James Ward described this as putting So James Ward described this as putting classifiers into the outer loop of a classifiers into the outer loop of a classifiers into the outer loop of a harness. And I think that's a good way harness. And I think that's a good way harness. And I think that's a good way to talk about it because then generative to talk about it because then generative to talk about it because then generative models are inside that loop as tools. So models are inside that loop as tools. So models are inside that loop as tools. So in a situation like this where a whole in a situation like this where a whole in a situation like this where a whole AI system is running and it's it's doing AI system is running and it's it's doing AI system is running and it's it's doing like a a workflow. it's going through like a a workflow. it's going through like a a workflow. it's going through it's passing uh a document through it's passing uh a document through it's passing uh a document through through various iterations in a business through various iterations in a business through various iterations in a business something like that Jev can then be the something like that Jev can then be the something like that Jev can then be the chooser Jev can then say in this chooser Jev can then say in this chooser Jev can then say in this situation with this state of this situation with this state of this situation with this state of this document is the next step a tool is the document is the next step a tool is the document is the next step a tool is the next step calling an LLM to produce text next step calling an LLM to produce text next step calling an LLM to produce text or do I call an LLM that's a frontier or do I call an LLM that's a frontier or do I call an LLM that's a frontier model to reason because it's a hard model to reason because it's a hard model to reason because it's a hard problem and it's an exception or do I problem and it's an exception or do I problem and it's an exception or do I call a human and the result then gets call a human and the result then gets call a human and the result then gets kicked off in the rest of the loop and kicked off in the rest of the loop and kicked off in the rest of the loop and you might think, okay, that's for you might think, okay, that's for you might think, okay, that's for orchestration for business process. How orchestration for business process. How orchestration for business process. How does that matter to you and me? Well, it does that matter to you and me? Well, it does that matter to you and me? Well, it turns out that using a browser is the turns out that using a browser is the turns out that using a browser is the same kind of problem as that workflow if same kind of problem as that workflow if same kind of problem as that workflow if you look at it from Jev's perspective.

you look at it from Jev's perspective. you look at it from Jev's perspective. if you put on your Jev glasses. In fact, if you put on your Jev glasses. In fact, if you put on your Jev glasses. In fact, there's an open- source agent out there there's an open- source agent out there there's an open- source agent out there that browser use has released where Jev that browser use has released where Jev that browser use has released where Jev can choose an operation and an element can choose an operation and an element can choose an operation and an element because what's happening with internet because what's happening with internet because what's happening with internet navigation is you have a limited number navigation is you have a limited number navigation is you have a limited number of choices or buttons or links and Jev of choices or buttons or links and Jev of choices or buttons or links and Jev just needs to choose between them and just needs to choose between them and just needs to choose between them and the information it has. And I've seen the information it has. And I've seen the information it has. And I've seen setups where Jev is looking at written setups where Jev is looking at written setups where Jev is looking at written information to navigate. I've seen information to navigate. I've seen information to navigate. I've seen situations where Jev is composing an situations where Jev is composing an situations where Jev is composing an interface, putting new buttons and new interface, putting new buttons and new interface, putting new buttons and new text in place live as you speak because text in place live as you speak because text in place live as you speak because Jev understands how to compose blocks Jev understands how to compose blocks Jev understands how to compose blocks and select options from a limited set. and select options from a limited set. and select options from a limited set. So you can follow a design system and So you can follow a design system and So you can follow a design system and live compose an interface on the fly. live compose an interface on the fly. live compose an interface on the fly. And in both that browser session which And in both that browser session which And in both that browser session which is really visceral, you can like see it is really visceral, you can like see it is really visceral, you can like see it happen live and in the workflow system happen live and in the workflow system happen live and in the workflow system which is harder to understand, Jev is which is harder to understand, Jev is which is harder to understand, Jev is doing the same job. Jev is basically doing the same job. Jev is basically doing the same job. Jev is basically deciding what goes where. Now, the last deciding what goes where. Now, the last deciding what goes where. Now, the last arrangement or architectural pattern I'm arrangement or architectural pattern I'm arrangement or architectural pattern I'm going to describe for you is one that I going to describe for you is one that I going to describe for you is one that I think is the most exciting. It's think is the most exciting. It's think is the most exciting. It's something that gives you a sense of how something that gives you a sense of how something that gives you a sense of how big a deal it is when intelligence is big a deal it is when intelligence is big a deal it is when intelligence is absolutely everywhere. In this case, the absolutely everywhere. In this case, the absolutely everywhere. In this case, the code may supply all of the available code may supply all of the available code may supply all of the available components and actions. There's not components and actions. There's not components and actions. There's not necessarily an LLM in play. And all Jev necessarily an LLM in play. And all Jev necessarily an LLM in play. And all Jev is doing is interpreting what the user is doing is interpreting what the user is doing is interpreting what the user is trying to do live. and choosing the is trying to do live. and choosing the is trying to do live. and choosing the appropriate response really fast. So, an appropriate response really fast. So, an appropriate response really fast. So, an example would help here. Nater Dobit example would help here. Nater Dobit example would help here. Nater Dobit showed a spreadsheet where all I as a showed a spreadsheet where all I as a showed a spreadsheet where all I as a user had to do is type the word urgency

user had to do is type the word urgency user had to do is type the word urgency into a column header because we have into a column header because we have into a column header because we have those things like is it urgent, is it those things like is it urgent, is it those things like is it urgent, is it not urgent? It's a business problem not urgent? It's a business problem not urgent? It's a business problem classifier. And as soon as you type that classifier. And as soon as you type that classifier. And as soon as you type that in, Jev reads that input, understands in, Jev reads that input, understands in, Jev reads that input, understands that that column needs to be classified that that column needs to be classified that that column needs to be classified and instantaneously classifies the and instantaneously classifies the and instantaneously classifies the entire column row by row. Jev generated entire column row by row. Jev generated entire column row by row. Jev generated judgment of urgency. Now imagine extend judgment of urgency. Now imagine extend judgment of urgency. Now imagine extend that spreadsheet pattern to other ways that spreadsheet pattern to other ways that spreadsheet pattern to other ways we use software. So one column could we use software. So one column could we use software. So one column could evaluate urgency. Another could flag evaluate urgency. Another could flag evaluate urgency. Another could flag missing information. Another could missing information. Another could missing information. Another could identify the relevant team. Ordinary identify the relevant team. Ordinary identify the relevant team. Ordinary formulas can combine all those judgments formulas can combine all those judgments formulas can combine all those judgments with dates and amounts. And yes, you can with dates and amounts. And yes, you can with dates and amounts. And yes, you can get to an LLM at the end with an get to an LLM at the end with an get to an LLM at the end with an eventual response if you need to. But eventual response if you need to. But eventual response if you need to. But you're not asking the LLM to do that you're not asking the LLM to do that you're not asking the LLM to do that kind of urgency classification. I think kind of urgency classification. I think kind of urgency classification. I think the spreadsheet is so interesting the spreadsheet is so interesting the spreadsheet is so interesting because it illustrates what happens when because it illustrates what happens when because it illustrates what happens when Jev brings down the cost of intelligence Jev brings down the cost of intelligence Jev brings down the cost of intelligence and intelligence goes everywhere because and intelligence goes everywhere because and intelligence goes everywhere because the spreadsheet then becomes a place the spreadsheet then becomes a place the spreadsheet then becomes a place where calculations and language where calculations and language where calculations and language judgments are working together judgments are working together judgments are working together seamlessly. I think it's a nice example seamlessly. I think it's a nice example seamlessly. I think it's a nice example of how big a change we are on the cusp of how big a change we are on the cusp of how big a change we are on the cusp of today. Look, this is just scratching of today. Look, this is just scratching of today. Look, this is just scratching the surface. We are so early that I the surface. We are so early that I the surface. We are so early that I would expect to discover a ton of new would expect to discover a ton of new would expect to discover a ton of new surprising use cases in the days ahead.

surprising use cases in the days ahead. surprising use cases in the days ahead. You may discover some of those use cases You may discover some of those use cases You may discover some of those use cases as you listen to this video. I encourage as you listen to this video. I encourage as you listen to this video. I encourage you. Now, is Jev perfect? Of course not. you. Now, is Jev perfect? Of course not. you. Now, is Jev perfect? Of course not. It's an excellent generalpurpose It's an excellent generalpurpose It's an excellent generalpurpose classifier. classifier. classifier. Like anything in computer science, Like anything in computer science, Like anything in computer science, there's no perfect tool for everything. there's no perfect tool for everything. there's no perfect tool for everything. It still makes some mistakes. It's It still makes some mistakes. It's It still makes some mistakes. It's absolutely something you should test on absolutely something you should test on absolutely something you should test on your particular classification problems. your particular classification problems. your particular classification problems. But the thing I want you to pay But the thing I want you to pay But the thing I want you to pay attention to in this video as you attention to in this video as you attention to in this video as you understand these patterns of where Jev understand these patterns of where Jev understand these patterns of where Jev is appearing is I want you to realize is appearing is I want you to realize is appearing is I want you to realize capability can unlock a wider range of capability can unlock a wider range of capability can unlock a wider range of applications than a bunch of applications than a bunch of applications than a bunch of hyperspecific capabilities. And so part hyperspecific capabilities. And so part hyperspecific capabilities. And so part of what makes Jev powerful is that it of what makes Jev powerful is that it of what makes Jev powerful is that it applies fairly consistently across such applies fairly consistently across such applies fairly consistently across such a wide range of examples as I've given a wide range of examples as I've given a wide range of examples as I've given like the email example. Jev works in all like the email example. Jev works in all like the email example. Jev works in all of these and that's part of why it's so of these and that's part of why it's so of these and that's part of why it's so powerful. My focus in this video has powerful. My focus in this video has powerful. My focus in this video has been expanding our imaginations been expanding our imaginations been expanding our imaginations together. And this is where I want to together. And this is where I want to together. And this is where I want to take it home and give you like action take it home and give you like action take it home and give you like action steps. How you go from, okay, I get an steps. How you go from, okay, I get an steps. How you go from, okay, I get an idea of what Jevshaped problems are. idea of what Jevshaped problems are. idea of what Jevshaped problems are. Now, how do I try Jev? And the good Now, how do I try Jev? And the good Now, how do I try Jev? And the good thing is Jev is so cheap to try that you thing is Jev is so cheap to try that you thing is Jev is so cheap to try that you can just ask AI to help you come up with can just ask AI to help you come up with can just ask AI to help you come up with an idea if you're out of ideas and it's an idea if you're out of ideas and it's an idea if you're out of ideas and it's going to be fine. Getting started with going to be fine. Getting started with going to be fine. Getting started with Jev is as simple as literally copying Jev is as simple as literally copying Jev is as simple as literally copying the type safe setup prompt for agents the type safe setup prompt for agents the type safe setup prompt for agents that is on their website into cloud code that is on their website into cloud code that is on their website into cloud code or codeex or any other agent that you or codeex or any other agent that you or codeex or any other agent that you have. Chat GPT works great too and it's have. Chat GPT works great too and it's have. Chat GPT works great too and it's going to tell your agent to install the going to tell your agent to install the going to tell your agent to install the type safe skill and use it on your type safe skill and use it on your type safe skill and use it on your project. Typesafe is the company that

project. Typesafe is the company that project. Typesafe is the company that produced Jevs. So then you create an produced Jevs. So then you create an produced Jevs. So then you create an account, you generate an API key, which account, you generate an API key, which account, you generate an API key, which is as simple as a button click, and then is as simple as a button click, and then is as simple as a button click, and then you have the agent walk through you have the agent walk through you have the agent walk through connecting it up, and the agents really connecting it up, and the agents really connecting it up, and the agents really can do that now. And I've actually done can do that now. And I've actually done can do that now. And I've actually done it, right? I use this to build my email it, right? I use this to build my email it, right? I use this to build my email classifier. And if you're not sure what classifier. And if you're not sure what classifier. And if you're not sure what that job is, try something like this. that job is, try something like this. that job is, try something like this. Tell the LLM, use the type safe skill to Tell the LLM, use the type safe skill to Tell the LLM, use the type safe skill to find a place in the project we're find a place in the project we're find a place in the project we're working on together where asking an LLM working on together where asking an LLM working on together where asking an LLM to choose among defined outcomes is to choose among defined outcomes is to choose among defined outcomes is suboptimal. Build a Jev version instead. suboptimal. Build a Jev version instead. suboptimal. Build a Jev version instead. And then I want you to compare the And then I want you to compare the And then I want you to compare the results, compare the speed, compare the results, compare the speed, compare the results, compare the speed, compare the cost, and report back. And when I'm cost, and report back. And when I'm cost, and report back. And when I'm saying Jev is dirt cheap, I really mean saying Jev is dirt cheap, I really mean saying Jev is dirt cheap, I really mean dirt cheap. Look at the published price dirt cheap. Look at the published price dirt cheap. Look at the published price points. 4.2 2 cents per million input points. 4.2 2 cents per million input points. 4.2 2 cents per million input tokens with zero charge for output. tokens with zero charge for output. tokens with zero charge for output. Zero. The outputs are the expensive part Zero. The outputs are the expensive part Zero. The outputs are the expensive part with LLMs. So at a th000 input tokens with LLMs. So at a th000 input tokens with LLMs. So at a th000 input tokens per request, a,000 calls to Jev cost you per request, a,000 calls to Jev cost you per request, a,000 calls to Jev cost you 4 cents, which means 10,000 calls to Jev 4 cents, which means 10,000 calls to Jev 4 cents, which means 10,000 calls to Jev cost you 42 cents. The answer to life, cost you 42 cents. The answer to life, cost you 42 cents. The answer to life, the universe, and everything. That's how the universe, and everything. That's how the universe, and everything. That's how cheap Jev is to operate. and you can cheap Jev is to operate. and you can cheap Jev is to operate. and you can therefore afford to experiment with therefore afford to experiment with therefore afford to experiment with really big data or a substantial number really big data or a substantial number really big data or a substantial number of examples and you're going to be just of examples and you're going to be just of examples and you're going to be just fine on cost. I'm posting a whole guide.

fine on cost. I'm posting a whole guide. fine on cost. I'm posting a whole guide. It's linked in the description below. It's linked in the description below. It's linked in the description below. It'll show you how to get Jev installed It'll show you how to get Jev installed It'll show you how to get Jev installed in like 5 minutes or less. A prompt you in like 5 minutes or less. A prompt you in like 5 minutes or less. A prompt you can point at the work you already have can point at the work you already have can point at the work you already have and it will come back and tell you which and it will come back and tell you which and it will come back and tell you which part of the work you're already working part of the work you're already working part of the work you're already working on that's shaped like Jev. And then on on that's shaped like Jev. And then on on that's shaped like Jev. And then on top of that, I have recipe cards for top of that, I have recipe cards for top of that, I have recipe cards for each of the patterns I just walked you each of the patterns I just walked you each of the patterns I just walked you through, written so you can hand them through, written so you can hand them through, written so you can hand them straight to your coding agent and it straight to your coding agent and it straight to your coding agent and it builds the architecture for you. This is builds the architecture for you. This is builds the architecture for you. This is the beauty of Jev. Part of why Jev works the beauty of Jev. Part of why Jev works the beauty of Jev. Part of why Jev works is agents can help you use Jeff. You is agents can help you use Jeff. You is agents can help you use Jeff. You don't have to invent the shape. You can don't have to invent the shape. You can don't have to invent the shape. You can pick the pattern from the guide that pick the pattern from the guide that pick the pattern from the guide that matches your problem and you can start matches your problem and you can start matches your problem and you can start from that working recipe card and get from that working recipe card and get from that working recipe card and get right into using Jev to save money. Let right into using Jev to save money. Let right into using Jev to save money. Let me take it just a little bit further. me take it just a little bit further. me take it just a little bit further. Again, the pricing is killing me. It's Again, the pricing is killing me. It's Again, the pricing is killing me. It's part of why this is so exciting. A part of why this is so exciting. A part of why this is so exciting. A million requests for jug each a thousand million requests for jug each a thousand million requests for jug each a thousand input tokens. It's still just 42 bucks. input tokens. It's still just 42 bucks. input tokens. It's still just 42 bucks. I keep emphasizing the cheapness and I I keep emphasizing the cheapness and I I keep emphasizing the cheapness and I keep emphasizing how fast it is. I gave keep emphasizing how fast it is. I gave keep emphasizing how fast it is. I gave you seconds and like how fast it is you seconds and like how fast it is you seconds and like how fast it is earlier in this video because I think earlier in this video because I think earlier in this video because I think that is what is unlocking the real that is what is unlocking the real that is what is unlocking the real business value here. Think how many business value here. Think how many business value here. Think how many business judgments you could start business judgments you could start business judgments you could start making if you could send a thousand making if you could send a thousand making if you could send a thousand tokens a million different times. Hypes tokens a million different times. Hypes tokens a million different times. Hypes safe's own launch evaluation reported safe's own launch evaluation reported safe's own launch evaluation reported gains approaching two orders of gains approaching two orders of gains approaching two orders of magnitude or 100x faster on speed and magnitude or 100x faster on speed and magnitude or 100x faster on speed and more than two orders of magnitude or more than two orders of magnitude or more than two orders of magnitude or more than 100x cheaper in cost compared more than 100x cheaper in cost compared more than 100x cheaper in cost compared to an LLM. And of course you think well to an LLM. And of course you think well to an LLM. And of course you think well a launch announcement is made to look a launch announcement is made to look a launch announcement is made to look good Nate why are you buying that? That good Nate why are you buying that? That good Nate why are you buying that? That is actually stuff I've seen this is actually stuff I've seen this is actually stuff I've seen this release. I've used it and so many other release. I've used it and so many other release. I've used it and so many other developers have seen already. And this developers have seen already. And this developers have seen already. And this is where we come back to the Jev and the

is where we come back to the Jev and the is where we come back to the Jev and the strategic impact of Jieven's paradox. strategic impact of Jieven's paradox. strategic impact of Jieven's paradox. Ultimately, the strategic effect of a Ultimately, the strategic effect of a Ultimately, the strategic effect of a large cost reduction is that it changes large cost reduction is that it changes large cost reduction is that it changes which questions are worth asking. It which questions are worth asking. It which questions are worth asking. It increases the possible uses of increases the possible uses of increases the possible uses of intelligence. If classification is 10% intelligence. If classification is 10% intelligence. If classification is 10% of your existing AI bill today, imagine of your existing AI bill today, imagine of your existing AI bill today, imagine making it 1%. It saves roughly your making it 1%. It saves roughly your making it 1%. It saves roughly your entire slice. Your current bill is only entire slice. Your current bill is only entire slice. Your current bill is only the beginning of the story, though, the beginning of the story, though, the beginning of the story, though, because Jieven's paradox reminds us that because Jieven's paradox reminds us that because Jieven's paradox reminds us that the work you decided was worth paying the work you decided was worth paying the work you decided was worth paying for when it was expensive is not all of for when it was expensive is not all of for when it was expensive is not all of the work you would want to do if it was the work you would want to do if it was the work you would want to do if it was cheap. You have so much more work you cheap. You have so much more work you cheap. You have so much more work you can do if intelligence is this cheap. can do if intelligence is this cheap. can do if intelligence is this cheap. That's what we're talking about here. That's what we're talking about here. That's what we're talking about here. These abandoned possibilities are where These abandoned possibilities are where These abandoned possibilities are where the future now lies. Now that we have the future now lies. Now that we have the future now lies. Now that we have this third primitive in the computing this third primitive in the computing this third primitive in the computing space, a company that samples customer space, a company that samples customer space, a company that samples customer calls can examine every single call, calls can examine every single call, calls can examine every single call, every single time for multiple different every single time for multiple different every single time for multiple different kinds of problems and evolve that kinds of problems and evolve that kinds of problems and evolve that classifier on the fly. A team that tags classifier on the fly. A team that tags classifier on the fly. A team that tags documents once can rapidly reclassify documents once can rapidly reclassify documents once can rapidly reclassify based on customer input every single day based on customer input every single day based on customer input every single day its whole history when a new opportunity its whole history when a new opportunity its whole history when a new opportunity appears. It costs almost nothing. An appears. It costs almost nothing. An appears. It costs almost nothing. An interface that asks you to stop and interface that asks you to stop and interface that asks you to stop and explain what you want can evolve and explain what you want can evolve and explain what you want can evolve and make useful suggestions as you work.

make useful suggestions as you work. make useful suggestions as you work. Each time intelligence becomes that much Each time intelligence becomes that much Each time intelligence becomes that much cheaper, you can afford to do that much cheaper, you can afford to do that much cheaper, you can afford to do that much more. That's what Jievens means. So more. That's what Jievens means. So more. That's what Jievens means. So tasks that were done once per document tasks that were done once per document tasks that were done once per document become daily per section or per customer become daily per section or per customer become daily per section or per customer or per interaction. A decision at the or per interaction. A decision at the or per interaction. A decision at the beginning of an agent task becomes a beginning of an agent task becomes a beginning of an agent task becomes a safety decision you can make at every safety decision you can make at every safety decision you can make at every relevant step. And that's why the name relevant step. And that's why the name relevant step. And that's why the name Jev is so apt. William Stanley Jivens of Jev is so apt. William Stanley Jivens of Jev is so apt. William Stanley Jivens of course observed that greater efficiency course observed that greater efficiency course observed that greater efficiency could expand the uses of a resource could expand the uses of a resource could expand the uses of a resource enough to increase overall consumption. enough to increase overall consumption. enough to increase overall consumption. That was back in the 19th century. Types That was back in the 19th century. Types That was back in the 19th century. Types safe named the model for that idea safe named the model for that idea safe named the model for that idea because cheaper intelligence can create because cheaper intelligence can create because cheaper intelligence can create massively more demand for intelligence massively more demand for intelligence massively more demand for intelligence in places where there was previously no in places where there was previously no in places where there was previously no model that we would really be able to model that we would really be able to model that we would really be able to afford to use. I think the idea of afford to use. I think the idea of afford to use. I think the idea of intelligence too cheap to meter captures intelligence too cheap to meter captures intelligence too cheap to meter captures what I'm so excited about here. The what I'm so excited about here. The what I'm so excited about here. The reason this is such a big deal is that reason this is such a big deal is that reason this is such a big deal is that we are for the first time so cheap, so we are for the first time so cheap, so we are for the first time so cheap, so efficient and so fast that intelligence efficient and so fast that intelligence efficient and so fast that intelligence can literally disappear into ordinary can literally disappear into ordinary can literally disappear into ordinary features into everything we look at.

features into everything we look at. features into everything we look at. Interfaces can evolve intelligently. Interfaces can evolve intelligently. Interfaces can evolve intelligently. I've seen those demos with Jev already. I've seen those demos with Jev already. I've seen those demos with Jev already. It can mean that our most capable It can mean that our most capable It can mean that our most capable generative models end up becoming even generative models end up becoming even generative models end up becoming even more valuable. Remember when I said LLMs more valuable. Remember when I said LLMs more valuable. Remember when I said LLMs weren't dead? This is why more weren't dead? This is why more weren't dead? This is why more classification can uncover a whole classification can uncover a whole classification can uncover a whole universe of new opportunities worth universe of new opportunities worth universe of new opportunities worth pursuing, more questions worth pursuing, more questions worth pursuing, more questions worth investigating with high-grade investigating with high-grade investigating with high-grade intelligence, and more exceptions that intelligence, and more exceptions that intelligence, and more exceptions that require LLMs to think and ultimately require LLMs to think and ultimately require LLMs to think and ultimately humans to think. Each of those can humans to think. Each of those can humans to think. Each of those can create work for a model that reasons and create work for a model that reasons and create work for a model that reasons and writes with language. So the total pie writes with language. So the total pie writes with language. So the total pie of useful AI work, including LLM work, of useful AI work, including LLM work, of useful AI work, including LLM work, is growing really, really rapidly. And is growing really, really rapidly. And is growing really, really rapidly. And when I look at the adoption curve that I when I look at the adoption curve that I when I look at the adoption curve that I shared at the top of this video, I see shared at the top of this video, I see shared at the top of this video, I see developers recognizing a capability that developers recognizing a capability that developers recognizing a capability that they've repeatedly had to improvise and they've repeatedly had to improvise and they've repeatedly had to improvise and being able to finally do it correctly. being able to finally do it correctly. being able to finally do it correctly. We've used rules where interpretation We've used rules where interpretation We've used rules where interpretation was needed. We've built dedicated was needed. We've built dedicated was needed. We've built dedicated classifiers when the investment made classifiers when the investment made classifiers when the investment made sense. We've put LLMs into countless sense. We've put LLMs into countless sense. We've put LLMs into countless places where the immediate output was a places where the immediate output was a places where the immediate output was a decision. Now we all have a new building decision. Now we all have a new building decision. Now we all have a new building block. And I love that on the TypeSafe block. And I love that on the TypeSafe block. And I love that on the TypeSafe homepage, it's so easy to use. They know homepage, it's so easy to use. They know homepage, it's so easy to use. They know how we use this stuff. It's literally how we use this stuff. It's literally how we use this stuff. It's literally just hit copy, give this to your agent.

just hit copy, give this to your agent. just hit copy, give this to your agent. And so we can all have it, not just And so we can all have it, not just And so we can all have it, not just developers. And we're finding uses for developers. And we're finding uses for developers. And we're finding uses for it across all of these systems. My it across all of these systems. My it across all of these systems. My challenge to you, look at the software challenge to you, look at the software challenge to you, look at the software you use every day through the lens of you use every day through the lens of you use every day through the lens of Jev. Heatedly interpret a piece of Jev. Heatedly interpret a piece of Jev. Heatedly interpret a piece of information. Just choose among a few information. Just choose among a few information. Just choose among a few choices or outcomes. Which of those choices or outcomes. Which of those choices or outcomes. Which of those judgments is buried inside a really judgments is buried inside a really judgments is buried inside a really expensive AI call today? Where have you expensive AI call today? Where have you expensive AI call today? Where have you simply gone without the judgment? simply gone without the judgment? simply gone without the judgment? Because putting intelligence there Because putting intelligence there Because putting intelligence there seemed like it was too expensive or too seemed like it was too expensive or too seemed like it was too expensive or too impractical. And that last question, of impractical. And that last question, of impractical. And that last question, of course, is the most exciting one. What course, is the most exciting one. What course, is the most exciting one. What if every element could be intelligent? if every element could be intelligent? if every element could be intelligent? Every row could respond to what you Every row could respond to what you Every row could respond to what you mean. Every document could be evaluated mean. Every document could be evaluated mean. Every document could be evaluated against what matters to you now. Every against what matters to you now. Every against what matters to you now. Every proposed action can get a relevant check proposed action can get a relevant check proposed action can get a relevant check on it. From an agent safety perspective, on it. From an agent safety perspective, on it. From an agent safety perspective, you can change your priorities and have you can change your priorities and have you can change your priorities and have the software reconsider the information the software reconsider the information the software reconsider the information around you. Now, we're absolutely going around you. Now, we're absolutely going around you. Now, we're absolutely going to need systems that reason and write to need systems that reason and write to need systems that reason and write and invent and explain. Those are LLM and invent and explain. Those are LLM and invent and explain. Those are LLM shaped problems. But we can also give shaped problems. But we can also give shaped problems. But we can also give the rest of software a much richer the rest of software a much richer the rest of software a much richer ability to make useful decisions. Jev ability to make useful decisions. Jev ability to make useful decisions. Jev makes that possibility concrete enough makes that possibility concrete enough makes that possibility concrete enough to build with. And the range of what to build with. And the range of what to build with. And the range of what people are building tells me that this people are building tells me that this people are building tells me that this is a massive deal. And that's why I had is a massive deal. And that's why I had is a massive deal. And that's why I had to share it with you today. Sam Alvin to share it with you today. Sam Alvin to share it with you today. Sam Alvin said something a couple of months ago said something a couple of months ago said something a couple of months ago that stuck with me. We live in an that stuck with me. We live in an that stuck with me. We live in an industry where one person working in industry where one person working in industry where one person working in monklike isolation can come out of monklike isolation can come out of monklike isolation can come out of nowhere with a surprising insight that nowhere with a surprising insight that nowhere with a surprising insight that moves the whole industry forward. This moves the whole industry forward. This moves the whole industry forward. This feels like one of those moments. Jev is

feels like one of those moments. Jev is feels like one of those moments. Jev is one of those moments. A tiny team built one of those moments. A tiny team built one of those moments. A tiny team built this. So if you're that one person or this. So if you're that one person or this. So if you're that one person or you're that tiny team and you're tapping you're that tiny team and you're tapping you're that tiny team and you're tapping away at this strange idea nobody asked away at this strange idea nobody asked away at this strange idea nobody asked for, this is your encouragement to keep for, this is your encouragement to keep for, this is your encouragement to keep going. You might have the next Jev. And going. You might have the next Jev. And going. You might have the next Jev. And if you've tried Jev, I want to hear if you've tried Jev, I want to hear if you've tried Jev, I want to hear about it in the comments. And of course, about it in the comments. And of course, about it in the comments. And of course, if you haven't, I really encourage you if you haven't, I really encourage you if you haven't, I really encourage you to give it a shot. You want something to give it a shot. You want something to give it a shot. You want something 100x faster, 100x cheaper. See what it 100x faster, 100x cheaper. See what it 100x faster, 100x cheaper. See what it can do for you. Understand how it's can do for you. Understand how it's can do for you. Understand how it's different from something we've seen different from something we've seen different from something we've seen before. Throw it at a problem that feels before. Throw it at a problem that feels before. Throw it at a problem that feels like it has a lot of mess going in and a like it has a lot of mess going in and a like it has a lot of mess going in and a few choices going out. See what you get.