July 8, 2026
Signal, noise, and comment chaos
NoiseLang: Where N = 5 is a Dirac delta
A 9-year passion project returns—and the comments instantly turned into a nerd showdown
TLDR: NoiseLang is a revived side project that turns messy probability questions into something a computer can explore automatically, and its creator says AI helped finally make it happen. Commenters were intrigued, but quickly turned the thread into a debate over whether it’s truly new or just joining a crowded family of existing tools.
A programmer dusted off a nine-year-old dream and came back with NoiseLang, a strange-and-fascinating tool where every number is treated like a cloud of possible outcomes. In plain English: instead of writing endless homework-style math by hand, you can ask questions like “what are the odds 23 people share a birthday?” and let the computer do the heavy lifting. The creator says modern AI finally helped turn the idea from “cute weekend fantasy” into something real, complete with a fast engine, browser support, and flashy demos.
But the real party started in the comments, where the community immediately went full “cool project, but have you seen these five other probability languages?” One person casually dropped monad-bayes, another shot back with Stan, and then someone else basically said, “you should probably mention the entire probabilistic programming universe,” name-checking webppl, anglican, and pymc like they were assembling the Avengers. The vibe was less “wow, amazing” and more friendly academic cage match: is this a fresh invention, a reinvention, or a stylish remix?
Then came the unexpectedly funny side quest: one commenter couldn’t even open a linked site because it was on a blocklist, leading to a hilariously deadpan non-drama—“I’m not expressing an opinion on the site”—before posting the official NoiseLang link instead. So yes, the project is ambitious, but the comments turned it into what the internet loves most: a mix of admiration, nitpicking, receipts, and extremely online humor.
Key Points
- •NoiseLang is a probabilistic language project revived after nine years and recently advanced with the help of AI tools.
- •The language represents every value as a probability distribution, treating constants as Dirac delta distributions and random variables as the same kind of object.
- •Variable names refer to fixed nodes in the computation graph, while independent random variables are created through separate draws using `~` or `~[N]`.
- •NoiseLang performs no computation until a probability or expectation is requested, at which point it runs large-scale Monte Carlo simulation and reports estimates with standard error.
- •The current implementation uses a compiler architecture with a Cranelift-based JIT, a WASM backend, numerical runtime code, and an internal append-only DAG called `RvGraph`.