Ask HN: How do systems (or people) detect when a text is written by an LLM

There is a massive demand for reliable, local-first LLM detection tools that don't rely on opaque, unreliable commercial APIs. Building an open-source library that aggregates various linguistic analysis methods like perplexity tracking and token distribution analysis would be highly valuable.

Suggested repo: detectGPT-lite

"Stop relying on black-box APIs; build your own lightweight, transparent LLM content classifier."

Estimated effort: 40h