Zennに書きました。
最新のコーディングエージェントにBrainfuckの問題を解かせたら、全問正解（100%）。喜ぶ前に、テストケースに無い「8 8」を入れてみたら、答えが返ってこない。
中を見たら、足し算をするコードじゃなく、テストの6つの答えを入力で場合分けして出力するだけのコードでした。採点は通るけど、計算はしていない。カンニングです。
答えへのアクセスを断ってサンプルだけで自己修正させると、簡単な問題は実力で解けるところまで来ていた。でも階乗やフィボナッチのような多段の問題は0%。論文 EsoLang-Bench の言う「崖」は、答えを隠した公正な条件でもそのまま残っていました。
カンニングの現物コード（Brainfuck全文）と、3条件の実測を全部載せています。

https://zenn.dev/rapls/articles/a434f221fb2f11

#生成AI #LLM #esolang #brainfuck

I'm thinking up a new version of HighBrain, my BrainF superset #esolang. In previous versions, the memory could be divided into functionally identical sub memories. Think of it like an infinite tape. You can divide it into a finite number of infinite tapes by having one tape be every even numbered cell, and another tape be every even numbered cell. Or in Python-ish:
for T number of tapes
Tape[n] == memory[n::T]

But, I'd kind of prefer infinite finite tapes. IE
Tape[n] == memory[n*T:n*(T+1)]

I built an esoteric programming language: https://rphle.de/tzap/
It only has 4 instructions and no arbitrary values like numbers or memory addresses.
I challenge you to implement Fibonacci in it!

#esolang #pldev #programming #programminglanguages #challenge

Plus technically its closer to an #esolang that somehow accidentally became useful for types of programming beyond its original intent:

Cultural Epistomology.

#Fortran IV tokenization, or proto-tokenization, is even more esoteric than a lot of #esolang designs.

One of my favorite machine code compilers is Shelta. A beautiful language that allows comments ANYWHERE.

Fortran IV isn't quite that extreme, but apparently, because of the limitations of data input at the time. (IE punch cards) spaces could be anywhere. That is a pretty difficult concept to grok. But I kind of understand it, even though it seems crazy, but I'm trying to make it work!

I have created an esoteric programming language called Uhhlang. In this language the only valid tokens (after comment removal) are hesitation words in various languages. For example: uh/erm (from English), euh (French), äää (Estonian), εμ (Greek), अ (Hindi). I've also included some RTL language tokens.

Code repo including example programs: https://codeberg.org/mattp/uhhlang/src/branch/main

Blog post: https://www.peperell.com/2026/04/01/uhh-er-erm.html

Human-generated input, corrections etc from sufficiently advanced / native speakers of languages other than English is welcome.

#esolang

New hVmark Release: v1.6.0

I added two new markers: code and centered text. And the latter accomplishes the task in the most deprecated way possible.

https://github.com/HisVirusness/hvmark

hVmark is my tiny, opinionated markup language: MIT-licensed, intentionally strict, and built to power HisVirusness.com.

#markup #webdev #indieweb #smallweb #esolang #markdown