junehwi

Rather than compressing a judgment about reality into a single boolean, it seems more natural to break it down into smaller observations and evidence, then derive conclusions from them. Users wouldn't need to touch the ontology directly; they'd just answer small, concrete questions like "Is there a sidewalk here?"

Thanks for answering. That's really interesting. Maybe agents flip this around. Instead of humans maintaining executable docs, the actual work generates the document. It could become something like a PR-style review layer for agent work. You don't necessarily need to understand the underlying code or tooling, but you can inspect what changed, why it changed, and approve or reject it. Do you think that would address any of the scaling problems you saw?

>non-AI customer base are all huge advantages if not moats.

This is the interesting part to me. People talk about a “SaaSpocalypse” because AI makes SaaS features cheap to copy, yet deeply embedded SaaS still accumulates integrations, data, and switching costs. Gemini is a good example: Google can put AI directly into Gmail, Docs, Drive, Search, etc., where people already work. Meanwhile, frontier-model performance leads often seem to disappear within months. Could model quality itself actually be a less durable moat than workflow and distribution? Curious where people who’ve worked in ML for a long time see the moat actually compounding.

I'm curious about this too. Org-babel, Jupyter, Atuin Desktop, etc. have explored similar ideas, but none became a mainstream way of working.

For people who used these tools long-term: what held them back? UX/accessibility, or do people simply prefer keeping documents separate from the actual work?