A local Google Docs-style review loop logs every note to JSONL
basedjensen · x · 2026-07-28
The post describes a lightweight local review loop that mimics Google Docs for annotation and iteration.
A tiny server serves a tutorial page with an injected comment overlay: you can highlight text, attach notes, and have each note written directly into a JSONL file, while leaving the real index.html untouched. The screenshot shows the workflow in action, including scratchpad edits and inline comments, suggesting a practical pattern for review-driven editing and feedback collection.
More from Research
- Princeton Research: Automating News Knowledge Surveys with LLMs — manoelribeiro · 2026-07-29
- Princeton field study finds LLM templates raised feedback rates by 9.4 points — manoelribeiro · 2026-07-29
- WikiVault edits made up 27% of Korean Wikipedia’s byte growth with just 0.1% editor adoption — manoelribeiro · 2026-07-29
- Audit of 11 agentic AI systems finds they still fail at scientific evidence synthesis — manoelribeiro · 2026-07-29
- A survey of 221 people finds news knowledge and news anxiety are not tightly linked — manoelribeiro · 2026-07-29
- LLM adoption in Wikipedia coursework reached about 20% by spring 2025 — manoelribeiro · 2026-07-29