What happens to programming languages when humans are no longer writing most of the code?
99 comments
---
- Correct by construction: the language makes invalid states or programs hard or impossible to express.
- Statically established: types, proofs, and static analysis establish properties before execution.
- Runtime-enforced: memory management, isolation, capability boundaries, and other runtime enforced properties.
- Empirically validated: program validation through tests, property-based testing, and fuzzing.
---
Along with being familiar, so it's easy to generate, is a huge part of why I'm building Zena: https://zena-lang.dev/
I don't have the AI-first rationale put into the public docs well just yet, but I mention some of it here: https://zena-lang.dev/guide/why-zena/#familiar-to-humans-and...
along with a doc in the repo on this topic: https://github.com/elematic/zena/blob/main/docs/design/ai-fi...
In short, the more deterministic, automated, checks the better. AI can deal with a pedantic language. I intend to add statically verified structured concurrency, units of measure, contracts, and eventually more and more formal methods into the language so it can be a familiar TYpeScript-like base with as many static guarantees as we can fit in.
I also think that fine-grained isolation, which Zena gets via Web Assembly, is critical for limiting the capabilities of generated code and the blast radius of bugs, vulnerabilities, and non-aligned behavior.
I do have an optimistic hope that a language also optimized for humans, readability and simple semantics especially, has value in the future, even when most code is generated. We'll see about that.
> integer literals become i32, float literals become f32
Don't use the shitty small sizes that have introduced countless bugs over the last decades, as a language default.
Especially if you care about correctness.
i32 is all over the examples.
Use 64 bits, like Python and Haskell; even JavaScript and thus TypeScript got floats right defaulting to f64.
My understanding is that i32 and f64 literals are pretty common among the more modern languages because f64 isn't much slower than f32 on modern CPUs and f32 costs precision, but i64 is slower than i32 (especially multiplies and divides or on 32 bit hardware) and i32 costs range, not precision, and most uses don't need the extra range.
One thing I would love to have as a feature is native compilation.
JS derivatives are that, a derivative to a scripting language.
One reason I haven't explored that is that I want to tailor the language for the more constrained environment of Wasm GC first.
Agents will naturally be bad at it due to a lack of examples.
Btw if folks have ideas of how to make Killswitch even harder for LLMs I’d appreciate them.
The one thing I do see sometimes is that agents sometimes don't take advantage of added features, but that's partially because the Zena code base doesn't use them as much yet. I'm working on skills and linter-based suggestions to use better patterns.
Even though this programming language is absolutely nowhere in any large language model’s training set, they have so far done extremely well at extrapolating from the small set of examples I’ve given it when I need an agent to generate some tests or whatever for me.
The future is cooked
LLMs that have to constantly fix their code and look up libraries or features are much slower, more expensive and more prone to create inefficient and slightly wrong code. Sure, an LLM can just create a new web framework from scratch, but you usually don't want to spend your tokens on that.
I also think that with LLMs, new languages will have more trouble building an ecosystem, because until now, the popular libraries had a "proof of work" that kinda also made sure that they were well-trodden and maintained. Now, anyone can just push their new generated library and we have no clue on how to compare (and tbh reading LLM-generated READMEs is also not pleasant).
"LLMs dont create anything new, if programmers stop reading the code technology will be forever frozen to 2022, no new programming languages, operating systems, concurrency primitives, databases, networking protocols, UI frameworks everything will be based on the training data and future generations will forget about all the primitives we now take for granted.
If someone creates a new programming language/ framework or new better way to do async or whatever, no one will use it because it is not in the training data and it wont take off because everyone is using LLMs. It will be like using the same Lego pieces over and over."
Esoteric languages like BrainFuck are esoteric and difficult precisely because they go out of their way to eschew conceptual links to existing languages or paradigms.
So if you invented a new type of esoteric language with arbitrary syntax and strange operators (not sure how you’d do that, exactly, iirc all fundamental binary operators are known) it might be impossible to use with an LLM even if the user manual was in context… but aside from that, languages and the underlying concepts are extremely generalizable.
Seems like a solvable problem though by generating synthetic data that’s guaranteed to be accurate through linters, compilers and tests.
If it’s ~9 figures to train a frontier model, it seems like training on a new language could be a rounding error if it was a priority.
I could see the appeal of a new agentic-friendly language that’s focused on minimizing tokens. The standard library could be massive with no concern of making the language easily readable or learnable.
This makes sense and is very insightful, thank you. But with that solution it seems that only LLM companies will be in the position to create new languages.
My personal experience is that LLMs are quite good at one-shotting complex solutions in both Rust and Go, and tend to be idiomatic.
newdsl describe some_feature
DSLs climb the ladder of abstraction and constrain the solution space at the language level, which IMO provides tighter feedback to both humans and machines.
Implement whatever abstractions you think LLMs should work in terms of in whatever language is handy, and have your LLM use those abstractions.
LLMs are generally more tolerant of tedium than humans, but they make mistakes more often on repetitive mechanical tasks. I've had Claude write PTX directly once, and Claude just wrote majority of it and commented something along the line of "repeat this block 7 more time with these minor changes" instead of writing them out, so the code didn't work.
So, no, replacing compilers with LLMs is probably a worse option than having them code a compiler/programming language.
Plugging my own thing to use as example:
https://github.com/yuechen-li-dev/oct
Oct is the first programming language that I made with Codex. It started out as "Octave Modern" and was never intended to be a language for LLMs in the first place, but rather a teaching language that I've been thinking about for a decade because of my frustration with academic code and specifically reproducibility, with Python and Matlab in particular.
But it turns out the same design choices that was made to prevent bad patterns from academic code also made it pretty good for LLM coding: statically typed, immutable by default, GC'd with fast compile and runtime because it compiles to Go, along with features designed for scientific compute like SI units and builtin graphing etc.
But now it just took a life of its own, so it has extra features like templates/concepts, iterators, async/await, database query, build system for C/C++, SystemVerilog/WASM(WIP) codegen, LaTeX/pdf generation etc. None of them were features that were developed in isolation of "what an LLM agent might want to write" but to address a specific problem that I had encountered or to address specific failure modes that Codex/Claude actually had.
That's why I think Oct is probably one of the better languages for AI to write/generate, not because it was designed to be AI friendly in abstract, but that it's developed against how AI actually writes code, even though again, it is still very much a work in progress.
I'm with you there, I think I starred Oct when I came across it. I've been on that quest since 2014, we should collaborate! I'll be presenting this work at IROS tomorrow, I'd love to hear what you think: https://mech-lang.org/iros-r4r-2026/index.html
So here is my experiment with Kalman filters here.
https://github.com/yuechen-li-dev/oct/tree/main/Experiments/...
It's been a while, but I think the findings there are mostly that Kalman filters are relatively heavy and fairly narrow in application in that it's good at filter Gaussian noise out but not much else, and it's kind of branchy so it doesn't run on the GPU very well, you can try to see if a simpler feedforward like Smith predictor can work as well.
Also, something to try out: the explicit state machine stacks/pushdown automata are only half of the equation, the other （and imo more important） half is argmax/utility AI based transition instead of traditional state machine graph.
But yeah, if your target is embedded/bare-metal application for robotics, since Oct really isn't designed for it, maybe you would like to check out what I'm currently working on, the Concept programming language?
Matching a desired output is a global goal, but even that sometimes works now. Someone sent me a LLM-generated JPEG 2000 decoder. They got Fable to generate a decoder that uses a GPU to get the same answer as the reference implementation gets on the GPU.
Net result faster (10-70x) and massively lower memory footprint. Now we’ve got a service that can do something essentially instantly that most people wouldn’t have even bothered to try before.
Read the full thread on Hacker News →
Related stories
- The Verge · 0 points · 2 days ago
- Can you forget how you feel about Meta?theverge.comThe Verge · 0 points · 9 days ago
- The Verge · 0 points · 4 days ago
- Can John Ternus find Apple’s next big thing?theverge.comThe Verge · 0 points · 9 days ago
- Hacker News · 73 points · 9 days ago
- Evolving programming languages in the AI eratwitter.comHacker News · 5 points · 5 days ago