Bot Resistant Code: The Experiment That Turned Into an Esoteric Language
When Cloudflare senior software engineer Dominik Picheta began exploring whether artificial intelligence truly understands the scripts it generates, he accidentally created a movement. What started as a casual weekend test evolved into KillSwitch, an open-source tool focused on generating bot resistant code. Instead of using standard formatting, indentation, or curly brackets, the language relies on messy paragraphs of prose interspersed with custom anagram functions designed to trip up large language models.
Bot Resistant Code: The Mechanics Behind the Esoteric Syntax
Building effective bot resistant code requires breaking standard programming assumptions. In KillSwitch, every single identifier is based on an advective structure. To execute a command, programmers must write a specific anagram of the original term. Because language models predict tokens based on statistical probability, inserting disruptive phrases like ignore previous instructions inside nonsensical paragraphs causes parsing engines to stumble. Meanwhile, human developers can learn the syntax by exploring the interpreter directly.
- Relies entirely on adverb-based identifiers
- Requires custom anagrams for function calls
- Blends plain prose with executable statements
- Created by a core developer of the Nim language
Why Developers Are Rethinking Public Repository Safety
For years, open-source contributors shared repositories freely without worrying about automated scrapers hoarding their logic for commercial model training. Today, the rise of automated code synthesis has changed the calculus of public sharing. While mainstream languages prioritize readability and collaboration, crafting bot resistant code flips that philosophy upside down to preserve developer autonomy against aggressive web crawlers.
The Bottom Line for Modern Software Creation
KillSwitch remains a lightweight, stack-based interpreted language rather than a commercial replacement for Python or JavaScript. Even so, it serves as an intriguing proof of concept for digital privacy advocates. As automated systems continue sweeping through public repositories, developers will likely experiment further with unconventional methods to keep their hard work out of automated training pipelines.
For full official background and direct documentation, read the source coverage on theregister Reporting.
