#7: The Last Programming Language
Some notes about the future of programming languages y un importante agradecimiento a Benito y su concierto en el Super Tazón.
Gracias Benito
Gracias Conejo malo. Cuando dijiste que todos somos América y listaste a tantos países, con las banderas, luego de tantas referencias tan mágicas que hacen que volvamos a ver el medio tiempo del Super Tazón, con una lágrima en el ojo, quiero agradecerte por tu arte y ese momento tan emotivo para la historia.
Por cierto, lo mejor de la semana, seguir reviviendo momentos, ver todos los remakes, ver a todos los canadienses que quieren ser latinos, ver a los haters de Bad Bunny que le dan mérito por hablar del continente americano y simplemente ver cómo secuestró el timeline de medio planeta por una semana. Genio.
Programming languages have always been born the same way. Someone decides we need a better way to talk to machines. A leader, a team, a company ships a new syntax, a new runtime, a new framework. JavaScript alone gave us an entire graveyard of frameworks in the last decade because of that cycle.
But what if that cycle is ending?
Agents are writing apps, fixing bugs, pushing commits to open source projects. Claude Code, Codex, and others are not just autocompleting lines of code. They’re building features, refactoring architectures, shipping pull requests. And here’s the thing: they don’t need a new language to do it. They work with what we already have.
The future evolution of Python and JavaScript might not be decided by a committee or a benevolent dictator. It might be decided by people’s agents. By Claude Code suggesting improvements, by Codex pushing optimizations. The next major version of your favorite language might be shaped more by AI contributions than human ones.
And if you zoom out further, the layer of abstraction between “I want an app like Uber for X” and running software might eventually just be machine code, without anyone caring what happened in between. The programming language becomes an implementation detail that only the agent sees.
Of course, some people will vibecode new languages into existence. Because XKCD already taught us that as a society we are terrible at standards, and I expect GenAI to continue that tradition. But I don’t see how a new Swift, a new Ruby on Rails, or a new Go will shine in the future. The agents will just improve what exists.
This is a risky bet and I might be proven totally wrong. We will see.
A few more things worth reading:
Claude Opus 4.6 and GPT-5.3-Codex both dropped the same day last week, minutes apart. I’ve been testing both since launch and Opus is doing a better job for me so far.
Anthropic closed a $30B Series G at a $380B valuation. Insane numbers.
Om Malik does a great comparison between OpenAI and Anthropic. Really great reading.
Capital One acquired Brex for $5.15B. Peaked at $12.3B in 2022. An amazing close for an amazing company. We should applaud this.
Speaking of new languages, the people behind Kotlin are already thinking of their next programming language. I could be proven wrong.
Spotify keeps growing in users and margins but the stock keeps tanking. This market is so freaking weird.
I’ve been navigating a lot into the OpenClaw ecosystem. If you’re super interested in this, not just following the news but actually working with the tool, reach out.
Btw, I’m writing this newsletter from San Francisco this week, and yes, I spent half my week talking about Bad Bunny and sharing reels. And I am proud of my procrastination spirit.
Programming languages might stop evolving through new syntaxes and instead get quietly shaped by AI agents improving the tools we already use