WEBVTT NOTE Sentence-level transcript of https://www.youtube.com/watch?v=UlFB6efYN5Q NOTE One cue per sentence. Cue ids are the line anchors on /transcripts/UlFB6efYN5Q.html. A cue ends where the next begins, or 2 s after its last word. s1 00:00:02.000 --> 00:00:04.600 Hello, and thank you for being here. s2 00:00:04.600 --> 00:00:11.200 I'm Roberto, and today I'm going to tell you this song of types and agents. s3 00:00:11.200 --> 00:00:21.400 Uh basically a song that speaks about languages that fight each other to conquer what's the throne in the AI realm. s4 00:00:21.400 --> 00:00:27.960 And how I think that TypeScript might actually be winning this war. s5 00:00:27.960 --> 00:00:30.520 But let's start from the beginning. s6 00:00:30.520 --> 00:00:39.400 A few years ago, whenever someone was building uh in AI, they were certainly using Python. s7 00:00:39.400 --> 00:00:41.600 Like there was no doubt. s8 00:00:41.600 --> 00:00:51.680 All the other languages were bowing to Python because um because of its dominion uh over the AI world. s9 00:00:51.680 --> 00:01:05.040 And then, when in 2022 ChatGPT was released, everybody started wondering and starting understanding that AI was becoming something more. s10 00:01:05.040 --> 00:01:09.760 Was going outside of the bubble that lived for years. s11 00:01:09.760 --> 00:01:14.120 And started to becoming something more ambitious. s12 00:01:14.120 --> 00:01:22.440 And together with it, Python, which was the standard language for AI, became more ambitious as well. s13 00:01:22.440 --> 00:01:34.360 And that's how in 2024 uh GitHub actually claim claimed the ladder and became the most popular language on GitHub. s14 00:01:34.360 --> 00:01:37.160 So, you know, everybody was happy. s15 00:01:37.160 --> 00:01:40.680 Python finally reached the top. s16 00:01:40.680 --> 00:01:48.720 But little did they know that there was another contender for the throne. s17 00:01:48.720 --> 00:01:54.440 Another contender that was rising to challenge the claim that Python had on the throne. s18 00:01:54.440 --> 00:02:01.560 And this contender, as you may have guessed by now, was indeed TypeScript. s19 00:02:01.840 --> 00:02:05.560 But, before talking about this, let me present myself. s20 00:02:05.560 --> 00:02:07.560 My name is Roberto. s21 00:02:07.560 --> 00:02:12.959 I'm the CTO and co-founder of Reto, a context layer for AI agents. s22 00:02:12.959 --> 00:02:23.600 And I'm also the EU ambassador for AI's Pratic, a global community of AI builders meeting once per month to discuss the latest news in AI. s23 00:02:23.600 --> 00:02:28.920 I'm also a long-time JavaScript then turned TypeScript developer. s24 00:02:28.920 --> 00:02:35.360 And that's basically why I am talking about TypeScript today. s25 00:02:35.360 --> 00:02:38.680 So, let's begin. s26 00:02:38.680 --> 00:02:43.080 As we said, AI started moving up to the stack. s27 00:02:43.080 --> 00:02:54.800 It was moving from the infrastructure layer of these models, machine learning, and all related ecosystem towards the application layer. s28 00:02:54.800 --> 00:03:06.840 This means that AI stopped being something that you train, and it started being something that you ship inside your application. s29 00:03:06.840 --> 00:03:13.480 Applications started featuring AI, starting having features powered by AI. s30 00:03:13.480 --> 00:03:19.480 Which basically means that we started having applications that think. s31 00:03:19.480 --> 00:03:24.600 And the application layer was not Python's. s32 00:03:24.600 --> 00:03:32.200 The application layer has been TypeScript's for pretty long time now. s33 00:03:32.720 --> 00:03:39.239 Don't get me wrong, like I still think that Python has its own application. s34 00:03:39.239 --> 00:03:50.000 Like I still think that the brain of the of the agent and all the of all the AI world is actually still owned by Python. s35 00:03:50.000 --> 00:03:54.680 All the training, the research, the GPU serving is all Python's. s36 00:03:54.680 --> 00:04:00.120 Uh Has been Python's all along and it's going to be Python's for long time yet. s37 00:04:00.120 --> 00:04:05.640 And um what's changing is actually the application layer. s38 00:04:05.640 --> 00:04:14.040 Like few years ago, if you wanted to build something with AI built in the application, you had to use Python. s39 00:04:14.040 --> 00:04:16.959 But today, that's not the case anymore. s40 00:04:16.959 --> 00:04:20.239 And that's all uh the shift is about. s41 00:04:20.239 --> 00:04:24.000 TypeScript doesn't just own, you know, the UI or the back end. s42 00:04:24.000 --> 00:04:29.919 Started owning also the agentic layer of our application. s43 00:04:31.000 --> 00:04:41.320 And that's why in August 2025, TypeScript TypeScript actually passed Python as the most used language on GitHub. s44 00:04:41.320 --> 00:04:47.960 And the funny thing is that the reason that the GitHub reports gave was the same. s45 00:04:47.960 --> 00:04:53.960 Like in 2024, it said AI leads Python to the top language. s46 00:04:53.960 --> 00:05:01.280 While in 2025, it said AI leads TypeScript as the first language. s47 00:05:01.280 --> 00:05:10.200 And in both cases, as you can see, the global developers, the number of global developers were surging. s48 00:05:10.200 --> 00:05:18.480 In 2020 2025, we even have one new developer joining GitHub every second. s49 00:05:18.560 --> 00:05:24.120 So, what actually changed in this year? s50 00:05:24.120 --> 00:05:29.040 Like yeah, we were flooded from like new developers. s51 00:05:29.040 --> 00:05:37.160 In 2024, these newcomers were reaching for Python, or even maybe existing developers were reaching for Python. s52 00:05:37.160 --> 00:05:42.080 And in 2025, they reached for TypeScript instead. s53 00:05:42.400 --> 00:05:48.640 What changed between 2024 and 2025 was actually coding agents. s54 00:05:48.640 --> 00:06:00.680 The coding agents grew up like we saw established um we saw the players establishing themselves like Lava Cloud Code, Cursor, Codex. s55 00:06:00.680 --> 00:06:04.280 They became the default way to build applications. s56 00:06:04.280 --> 00:06:14.920 And the default way to uh to which these coding agents actually built the applications was TypeScript. s57 00:06:14.920 --> 00:06:30.480 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 s58 00:06:30.480 --> 00:06:32.440 inside themselves. s59 00:06:32.440 --> 00:06:41.519 The demand to have more AI integrations, more and more AI integrations doesn't fall on through Python. s60 00:06:41.519 --> 00:06:41.640 [clears throat] s61 00:06:41.640 --> 00:06:45.600 It falls on TypeScript. s62 00:06:47.400 --> 00:06:54.880 And pretty much all the tools that we use to build AI today already run on TypeScript. s63 00:06:54.880 --> 00:07:05.680 We even saw an AI lab acquiring a JavaScript runtime like last December Anthropic acquired Bun. s64 00:07:05.919 --> 00:07:21.000 But still, you know, okay, everybody use is using TypeScript because of the coding agents and we are having more and more demand to build uh AI to embed AI inside TypeScript s65 00:07:21.000 --> 00:07:22.680 application. s66 00:07:22.680 --> 00:07:25.040 But does this mean that we should do it? s67 00:07:25.040 --> 00:07:32.360 Like this is a fair question, like it's an honest question and it's a question worth answering. s68 00:07:32.360 --> 00:07:38.480 And the answer in my opinion can be yes like for several reasons. s69 00:07:38.480 --> 00:07:55.040 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 in TypeScript because we are having more and more application in TypeScript, which are going to field s70 00:07:55.040 --> 00:07:58.240 the training of next coding agents. s71 00:07:58.240 --> 00:08:10.120 And then we are having uh deeper integrations and more native integrations from these coding agents towards TypeScript, and we can expect that the quality of the output in TypeScript s72 00:08:10.120 --> 00:08:14.120 is going to be better and better from these coding agents. s73 00:08:14.120 --> 00:08:28.040 Since we're building applications, and uh we want to have like the highest quality of these applications, might make sense to build agents, which are the new kind of applications in TypeScript. s74 00:08:28.040 --> 00:08:38.960 And also, if you use TypeScript, you are actually tapping into what is probably the richest package manager out there. s75 00:08:38.960 --> 00:08:48.040 NPM comes with everything, uh pretty much everything, like authentication, payments, UI, infra. s76 00:08:48.040 --> 00:08:52.600 Like it's uh the deepest up layer tail that there is. s77 00:08:52.600 --> 00:09:01.960 So, since again, AI is coming towards the application layer, we need to integrate with all these right now. s78 00:09:01.960 --> 00:09:07.880 And tapping inside NPM is a very convenient way to do that. s79 00:09:09.200 --> 00:09:18.560 Also, you have by building in TypeScript, you can have one single language throughout all your code base. s80 00:09:18.560 --> 00:09:22.560 You can have one single code base for the whole application. s81 00:09:22.560 --> 00:09:28.640 Because you can use TypeScript for your agent loop, for the tools, for the back end service, for the UI. s82 00:09:28.640 --> 00:09:35.800 While if you use Python, you probably have to split uh split it at least into two services. s83 00:09:35.800 --> 00:09:41.320 Which means, you know, one service with FastAPI, Pydantic AI, and whatever. s84 00:09:41.320 --> 00:09:55.040 And then another um separate React application that you need to sync between between these two with a uh with a contract, which you have to maintain and synchronize. s85 00:09:55.480 --> 00:10:07.200 And speaking of contracts, with TypeScript, you can have one single consistent typing across all your all your application. s86 00:10:07.200 --> 00:10:18.440 While if you use Python instead, you at some point will stop at a boundary, cuz you will have your agent, maybe your back end, etc. with one consistent typing, s87 00:10:18.440 --> 00:10:30.720 and then you will have your React application or Vue or whatever with um another set of typing at which you need to synchronize between the two. s88 00:10:30.920 --> 00:10:38.160 So, if you use TypeScript, you can use Zod as a single schema throughout all your application, which is very convenient. s89 00:10:38.160 --> 00:10:40.200 You can define a type once. s90 00:10:40.200 --> 00:10:48.360 You can use this type in the back end and uh in the model, and you can use the same type in your UI. s91 00:10:48.360 --> 00:10:52.240 One type, checked and went. s92 00:10:53.280 --> 00:11:05.240 Also, like it makes sense to build in TypeScript today uh also in the AI ecosystem because we are seeing a very surge in in the AI ecosystem as well. s93 00:11:05.240 --> 00:11:18.880 Like take the Versatile AI SDK, for example, you can see that in just 1 year, it went from 1.6 million to 15.1 million downloads per week, which is between 9 and 10x s94 00:11:18.880 --> 00:11:21.839 in just 1 year. s95 00:11:22.520 --> 00:11:24.920 So, finally, I'd to put everything together. s96 00:11:24.920 --> 00:11:37.440 In my opinion, yes, it makes sense to build AI agents in TypeScript because you have like a uh you can leverage the the de facto default language for coding agent. s97 00:11:37.440 --> 00:11:42.520 You can have one single language for your whole application and your whole code base. s98 00:11:42.520 --> 00:11:46.720 You can tap into uh fast-growing AI ecosystem. s99 00:11:46.720 --> 00:11:51.160 You can have consistent typing uh across all your application. s100 00:11:51.160 --> 00:11:57.960 And you can tap into the richest package manager that there is, NPM. s101 00:11:59.680 --> 00:12:05.880 So, um you might ask was all this unpredictable? s102 00:12:05.880 --> 00:12:12.760 And the answer is actually no. Someone predicted this uh many years ago, almost 20 years ago. s103 00:12:12.760 --> 00:12:20.960 Jeff Atwood said, "Any application that can be written in JavaScript will eventually be written in JavaScript." s104 00:12:20.960 --> 00:12:32.960 And you know, as you as you probably know in the last few years, we have a corollary of this that any application that could be written in JavaScript will eventually be written in TypeScript. s105 00:12:32.960 --> 00:12:42.120 And so basically, we can say that any application, even the gigantic ones, will be written in TypeScript. s106 00:12:43.800 --> 00:12:47.839 And be mindful that what I showed you today is just the beginning. s107 00:12:47.839 --> 00:12:50.839 Like, we are just getting started. s108 00:12:50.839 --> 00:13:02.240 You can project this in a few years and you can see that on the application layer, the difference between TypeScript and Python is actually going to widen from here. s109 00:13:02.600 --> 00:13:09.680 So, um as I said, the model can still run on pip. s110 00:13:09.680 --> 00:13:20.120 But the agents, which is the application layer today, so the agent that called the models will probably ship on NPM. s111 00:13:20.120 --> 00:13:25.280 So, everything on the inference layer, you know, it's going to be Python. s112 00:13:25.280 --> 00:13:30.720 But everything but else, probably all TypeScript. s113 00:13:31.200 --> 00:13:34.240 Let me leave you with one recommendation then. s114 00:13:34.240 --> 00:13:36.640 Um keep training in Python. s115 00:13:36.640 --> 00:13:41.560 As I said, I don't see that's one going away soon. s116 00:13:41.560 --> 00:13:47.200 But please consider building the agents and the applications in TypeScript. s117 00:13:47.200 --> 00:13:55.800 Because if you don't do that now, if you overlook TypeScript, you are probably going to fall behind. s118 00:13:56.000 --> 00:13:58.400 That was all on my side today. s119 00:13:58.400 --> 00:14:01.520 I thank you all for your listening. s120 00:14:01.520 --> 00:14:05.520 And please scan the QR code for the slides. s121 00:14:05.520 --> 00:14:11.680 Reach out to me if you agree or if you disagree, if you have any feedback, and let's get in touch. s122 00:14:11.680 --> 00:14:13.200 Thank you. s123 00:14:13.200 --> 00:14:15.200 Bye-bye.