Ctatedev introduced Zero, a new systems programming language for AI agents. The language is positioned as a faster and smaller alternative to existing options. Core features include explicit capabilities, JSON-based diagnostics, and typed safe fixes. These elements target reduced friction when agents generate, interpret, and repair code. The launch was announced through direct posts and reposts referencing ctatedev.
Guillermo Rauch#400
Introducing Zero
The programming language for agents.
I wanted a systems language that was faster, smaller, and easier for agents to use and repair.
Explicit capabilities. JSON diagnostics. Typed safe fixes.
Made for agents on day zero.
Positive users praise Zero as an amazing new systems language for AI agents and congratulate the developer, while negative users dismiss the project with insults calling it dumb or unnecessary.
No Digg Deeper questions have been answered for this story yet.
Star it, fork it, join the experiment
https://github.com/vercel-labs/zero
@ctatedev so agents can't write any of the currently existing languages? 🤔
why would a brand new language be easier for agents vs languages that have billions and billions of existing, functioning lines of code throughout history that have been extensively used in training sets?
@yo_swif @ctatedev Tenés un punto. Pero:
Los lenguajes de programación que conocemos y practicamente su historia misma reafirman que siempre fueron hechos para humanos, en nuestro idioma, en nuestra comprensión.
Un agente de IA no es un humano, los lenguajes no están optimizados para él.
I did not try it yet, but it does quite a few of the things that I wrote about recently! https://lucumr.pocoo.org/2026/2/9/a-language-for-agents/
Introducing Zero
The programming language for agents.
I wanted a systems language that was faster, smaller, and easier for agents to use and repair.
Explicit capabilities. JSON diagnostics. Typed safe fixes.
Made for agents on day zero.
@ctatedev what makes this the programming language for agents?
@ctatedev It feels OK to be "AI first" as we also are. But shouldn't it also be human readable?
Also, why is "hello world" 16.2 KiB if is it a "systems language".
Koru is under 2 KiB for a static ELF binary.
@ctatedev In the year of our lord, 2026, there are only 3 languages worth to think about: Python, Rust and C++
@ctatedev @matiNotFound very nice, similar concept that led me to build fzy https://github.com/saint0x/fzy
@ctatedev I've been thinking about this for awhile. Languages were written (more or less) for humans to grok, but AI is a different animal. Only problem is, it's not trained on a zillion examples of our new fangled language. So not sure there is a win here.
@ctatedev I'd love to talk about the decision surface you worked from. I'm just about to launch ultraviolet into alpha. DM me. http://ultraviolet-lang.org
@cramforce 98.7% there
@ctatedev WEN self-hosting?
@ctatedev i'm assuming the aarch64 backend intentionally does nothing, but what's the point of the project in general? is there a design doc available to read anywhere? many of the decisions make absolutely 0 sense to me
@ctatedev Don’t we already have Whitespace for that? https://en.wikipedia.org/wiki/Whitespace_(programming_language)
@ctatedev feels like typescript and rust had a baby
looks lovely
@yo_swif @ctatedev Porque es demasiado rígido. Y podría funcionar en términos de eficiencia pero perderías todo el resto. Lo ideal sería buscar un punto intermedio donde no tengas problemas de portabilidad. Lo innegable hoy es que los lenguajes actuales no están optimizados para agentes.
@mitsuhiko i am absolutely 100% unsurprised.
@ctatedev LLMs generally suck at what they have not seen before. Why do you think they will be better at this?
@ctatedev pattern matching wen? https://zerolang.ai/reference#control-flow
@ctatedev https://github.com/cezarc1/websocket-stt-bench
