A Song of Types and Agents - Roberto Stagi, Ratel

Transcript

Hello, and thank you for being here. I'm Roberto, and today I'm going to tell you this song of Types and Agents. A song that speaks about languages that fight each other to conquer the throne in the AI realm, and how I think that TypeScript might actually be winning this war. But let's start from the beginning. A few years ago, whenever someone was building an AI, they were certainly using Python. There was no doubt. All the other languages were bowing to Python because of its dominion over the AI world. And then, in 2022, when Agipity was released, everybody started wondering and understanding that AI was becoming something more, was going outside of the bubble that it lived in for years, and started to become something more ambitious. And together with it, Python, which was the standard language for AI, became more ambitious as well. And that's how, in 2024, GitHub actually claimed the ladder and became the most popular language on GitHub. Everybody was happy. Python finally reached the top, but little did they know that there was another contender for that throne, another contender that was rising to challenge the claim that Python had on the throne. And this contender, as you may have guessed by now, was indeed TypeScript. But before talking about this, let me present myself. My name is Roberto. I'm the CTO and co-founder of Rato, a context layer for AI agents. And I'm also the EU ambassador for AI Socratic, a global community of AI builders meeting once per month to discuss the latest news in AI. I'm also a long-time JavaScript-then-turned-TypeScript developer. And that's why I am talking about TypeScript today. So let's begin. As we said, AI started moving up the stack. It was moving from the infrastructure layer of these models, machine learning, and all the related ecosystem toward the application layer. This means that AI stopped being something that you train, and it started being something that you ship inside your application. Applications started featuring AI, started having features powered by AI, which means that we started having applications that think. And the application layer was not Python's. The application layer has been TypeScript for a pretty long time now. Don't get me wrong. I still think that Python has its own application. I still think that the brain of the agent and all the AI world is actually still owned by Python. All the training, the research, the GPU serving is all Python. It has been Python all along, and it's going to be Python for a long time. And what's changing is actually the application layer. A few years ago, if you wanted to build something with AI built into the application, you had to use Python. But today, that's not the case anymore. And that's what the shift is about. TypeScript doesn't just own the UI or the backend. It started owning also the agentic layer of our application. And that's why, in August 2025, TypeScript actually passed Python as the most used language on GitHub. And the funny thing is that the reason the GitHub report gave was the same. In 2024, it said AI leads Python to top language, while in 2025, it said AI leads TypeScript as the first language. And in both cases, as you can see, the number of global developers was surging. In 2025, we even have one new developer joining GitHub every second. So, what actually changed in this year? We were flooded with new developers. In 2024, these new developers were reaching for Python. Or maybe existing developers were reaching for Python. And in 2025, they reached for TypeScript instead. What changed between 2024 and 2025 was actually coding agents. The coding agents grew up. We saw the players establishing themselves: Lavapog, CloudCode, Courser, Codex. They became the default way to build applications. And the default way in which these coding agents actually build the applications was TypeScript. And since every new app, pretty much every new app, is an agent today, because they ship these AI and agentic capabilities, they are hungry to embed AI inside themselves, the demand to have more AI integrations, more and more AI integrations, doesn't fall onto Python. It falls on TypeScript. And pretty much all the tools that we use to build AI today already run on TypeScript. We even saw an AI lab acquiring a JavaScript runtime. Last December, Anthropic acquired a ban. But still, okay, everybody is using TypeScript because of the coding agents, and we are having more and more demand to embed AI inside TypeScript applications. But does this mean that we should do it? This is a fair question. It's an honest question, and it's a question worth answering. And the answer, in my opinion, can be yes, for several reasons. The first one is that, since TypeScript is the default language for coding agents today, we can expect that they will become better and better in TypeScript, because we are having more and more applications in TypeScript, which are going to build the training of next coding agents. And then, we are having deeper integrations and more native integrations from these coding agents toward TypeScript. And we can expect that the quality of the output in TypeScript is going to be better and better from these coding agents. Since we are building applications, and we want to have the highest quality of these applications, it might make sense to build agents, which are the new kind of applications, in TypeScript. And also, if you use TypeScript, you are actually tapping into what is probably the richest package manager out there. So here, npm comes with everything, pretty much everything: authentication, payments, UI, infra. It's the deepest app-layer tail that there is. So since, again, AI is coming toward the application layer, we need to integrate with all this right now. And tapping into npm is a very convenient way to do that. Also, by building in TypeScript, you can have one single language throughout all your codebase. You can have one single codebase for the whole application, because you can use TypeScript for your agent loop, for the tools, for the backend service, for the UI. While, if you use Python, you probably have to split it at least into two services, which means one service with FastAPI, PyDAMP, TKI, and whatever, and then another separate React application that you need to sync between these two with a contract, which you have to maintain and synchronize. And speaking of contracts, with TypeScript, you can have one single consistent typing across all your application. While if you use Python instead, you, at some point, will stop at a boundary, because you will have your agent, maybe your backend, etc., with one consistent typing, and then you will have your React application, or Vue, or whatever, with another set of typing, which you need to synchronize between the two. So, if you use TypeScript, you can use Zod as a single schema throughout all your application, which is very convenient. You can define the type once. You can use this type in the backend and in the model. And you can use the same type in your UI. One type, checked end to end. Also, it makes sense to build in TypeScript today, also in the AI ecosystem, because we are seeing a surge in the AI ecosystem as well. Take the Versoilize AI SDK, for example. You can see that in just one year, it went from 1.6 million to 15.1 million downloads per week, which is between 9x and 10x in just one year. So finally, to put everything together, in my opinion, yes, it makes sense to build AI agents in TypeScript, because you can leverage the de facto default language for coding agents, you can have one single language for your whole application, your whole codebase, you can tap into a fast-growing AI ecosystem, you can have consistent typing across all your application, and you can tap into the richest package manager that there is, npm. So, you might ask, was all this unpredictable? And the answer is actually no. Someone predicted this many years ago, almost 20 years ago. Jeff Atwood said any application that can be written in JavaScript will eventually be written in JavaScript. And, as you probably know, in the last few years, we have a corollary of this: an application that could be written in JavaScript will eventually be written in TypeScript. And so, we can say that any application, even the agentic ones, will be written in TypeScript. Jeff Atwood said that that abstracts that abstracts that abstracts that abstracts that abstracts that abstracts that abstracts that abstracts that abstracts that abstracts that abstracts that abstracts that abstracts that abstracts that abstracts that abstracts that abstracts that abstracts that abstracts that abstracts that abstracts that abstracts that abstracts that abstracts that abstracts that abstracts that abstracts that abstracts that abstracts that abstracts that abstracts that abstracts that abstracts that abstracts that abstracts that abstracts that abstracts that abstracts that abstracts that abstracts that abstracts that abstracts that abstracts that abstracts that abstracts that abstracts that abstracts that abstracts that abstracts that abstracts that abstracts that abstracts that abstracts that abstracts that abstracts that abstracts that abstracts the agents, which is the application layer today, so the agent that called the models will probably ship on npm. So everything on the inference layer is going to be Python, but everything else, probably all TypeScript. Let me leave you with one recommendation, then. Keep training in Python. As I said, I don't see that one going away soon, but please consider building the agents and the applications in TypeScript, because if you don't do that now, if you overlook TypeScript, you are probably going to fall behind. That was all on my side today. I thank you all for your listening, and please scan the QR code for the slides. Reach out to me if you agree or if you disagree, if you have any feedback, and let's get in touch. Thank you. Bye bye. TypeScript doesn't just own, you know, the UI or the backend. It started owning also the agentic layer of our application. And that's why, in August 2025, TypeScript actually passed Python as the most used language on GitHub. And the funny thing is that the reason that the GitHub report gave was the same. Like, in 2024, it said AI leads Python to top language. While in 2025, it said AI leads TypeScript as the first language. And in both cases, as you can see, the global developers, the number of global developers were surging. In 2025, we even have one new developer joining GitHub every second. So, what actually changed in this year? Like, yeah, we were flooded from, like, new developers. In 2024, these new cameras were reaching for Python. Or even maybe existing developers were reaching for Python. And in 2025, they reached for TypeScript instead. What changed between 2024 and 2025 was actually coding agents. The coding agents grew up. Like, we saw established... We saw the players establishing themselves. Like, Lavapog, CloudCode, Courser, Codex. They became the default way to build applications. And the default way to which these coding agents actually build the applications was TypeScript. And, you know, since every new app, pretty much every new app is an agent today, because they ship these AI and agentic capabilities, they are hungry to embed AI inside themselves, the demand to have more AI integrations, more and more AI integrations, doesn't fall onto Python. And, it falls on TypeScript. And, pretty much all the tools that we use to build AI today already run on TypeScript. We even saw an AI lab acquiring a JavaScript runtime. Like, last December, Anthropic acquired a ban. But, still, you know, okay, everybody is using TypeScript because of the coding agents. And, we are having more and more demand to build... To embed AI inside TypeScript application. But, does this mean that we should do it? Like, this is a fair question. Like, it's an honest question. And, it's a question worth answering. And, the answer, in my opinion, can be yes. Like, for several reasons. The first one is that, since TypeScript is the default language for coding agents today, we can expect that they will become better and better in TypeScript, because we are having more and more applications in TypeScript, which are going to build the training of next coding agents. And, then, we are having deeper integrations and more native integrations from these coding agents towards TypeScript. And, we can expect that the quality of the output in TypeScript is going to be better and better from these coding agents. Since we are building applications, and we want to have, like, the highest quality of these applications, it might make sense to build agents, which are the new kind of applications in TypeScript. And, also, if you use TypeScript, you are actually tapping into what is probably the richest package manager out there. So, here, NPM comes with everything, pretty much everything, like authentication, payments, UI, infra, like it's the deepest up layer tail that there is. So, since, again, AI is coming towards the application layer, we need to integrate with all this right now. And tapping inside NPM is a very convenient way to do that. So, you can also, by building in TypeScript, you can have one single language throughout all your codebase. You can have one single codebase for the whole application, because you can use TypeScript for your agent loop, for the tools, for the backend service, for the UI. While, if you use Python, you probably have to split it at least into two services, which means, you know, one service with FastAPI, PyDAMP, TKI, and whatever. And then, another separate React application that you need to sync between these two with a contract, which you have to maintain and synchronize. And speaking of contracts, with TypeScript, you can have one single consistent typing across all your application. While if you use Python instead, you, at some point, will stop at a boundary, because you will have your agent, maybe your backend, etc., with one consistent typing. And then, you will have your React application, or Vue, or whatever, with another set of typing, which you need to synchronize between the two. So, if you use TypeScript, you can use Zod as a single schema throughout all your application, which is very convenient. You can define the type once. You can use this type in the backend and in the model. And you can use the same type in your UI. One type checked end-to-end. Also, like, it makes sense to build in TypeScript today, also in the AI ecosystem, because we are seeing a very surge in the AI ecosystem as well. Like, take the Versoilize AI SDK, for example. You can see that in just one year, it went from 1.6 million to 15.1 million downloads per week, which is between 9 and 10x in just one year. So, finally, like, to put everything together, in my opinion, yes, it makes sense to build the AI agents in TypeScript, because you have, like, you can leverage the de facto defaults language for coding agent, you can have one single language for your whole application, your whole code base. You can tap into a fast-growing AI ecosystem. You can have consistent typing across all your application. And you can tap into the richest package manager that there is, NPM. So, you might ask, was all this unpredictable? And the answer is actually no. Someone predicted this many years ago, almost 20 years ago. Jeff Atwood said any application that can be written in JavaScript will eventually be written in JavaScript. Jeff Atwood said any application that can be written in JavaScript. And, you know, as you probably know, in the last few years, we have a corollary of this, that an application that could be written in JavaScript will eventually be written in TypeScript. And so, basically, we can say that any application, even the agentic ones, will be written in TypeScript. Jeff Atwood said that in that abstracts that that abstracts that abstracts that abstracts that abstracts that abstracts that abstracts that abstracts that abstracts that abstracts that abstracts that abstracts that abstracts that abstracts that abstracts that abstracts that abstracts that abstracts that abstracts that abstracts that abstracts that abstracts that abstracts that abstracts that abstracts that abstracts that abstracts that abstracts that abstracts that abstracts that abstracts that abstracts that abstracts that abstracts that abstracts that abstracts that abstracts that abstracts that abstracts that abstracts that abstracts that abstracts that abstracts that abstracts that abstracts that abstracts that abstracts that abstracts that abstracts that abstracts that abstracts that abstracts that abstracts that abstracts that abstracts that abstracts that abstracts that abstracts that abstracts that abstracts that abstracts that abstracts that abstracts that abstracts that abstracts that abstracts that abstracts that abstracts that abstracts that abstracts that abstracts that the agents, which is the application layer today, so the agent that called the models will probably ship on npm. So everything on the inference layer, you know, it's going to be Python, but everything else, probably all TypeScript. Let me leave you with one recommendation then. Keep training in Python. As I said, I don't see that one going away soon, but please consider building the agents and the applications in TypeScript, because if you don't do that now, if you overlook TypeScript, you are probably going to fall behind. That was all on my side today. I thank you all for your listening, and please scan the QR code for the slides. Reach out to me if you agree or if you disagree, if you have any feedback, and let's get in touch. Thank you. Bye bye.