Mindwtr

I'm currently using the Mindwtr app to manage tasks, it's great and cross platform and all. However I wanted to propose a feature and I looked at the issues and got a bit weirded out by the tone, and it's clear to me they are using an AI to write their responses, but even more intesting it seems also fully automated. Things like saying to someone having the contributor tag on GitHub (and who also seems writing using an LLM) "Great breakdown — you've clearly looked at the codebase." after they provided input on a feature You can see the thread here and many more examples in the repo: https://github.com/dongdongbh/Mindwtr/issues/361 Right now I'm a bit torn, the app works well, it's open source, but at the same time I don't feel like it's okay to do this, I see it a bit like abandoning a project to the whims of the LLM. So, would you keep

Hello HN, I built Mindwtr because I wanted the speed and polish of apps like Things 3, but with the data ownership of Emacs Org-mode. It is an open-source (MIT/AGPL), local-first implementation of the Getting Things Done (GTD) methodology. The Stack: - Desktop: Tauri v2 (Rust) + React. It runs heavily on Arch Linux (my daily driver). - Mobile: React Native (Expo) on Android. - Sync: It treats the file system as the source of truth. Data is stored in JSON/TOML, allowing you to sync via Syncthing, Dropbox, or Git. Key Features: - "Copilot" Capture: Uses local regex and logic to auto-tag contexts/projects as you type. - Privacy: No cloud servers. Your data never leaves your device unless you sync it yourself. - AI Hooks: Optional integration (BYO Key) to break down vague tasks or review stale lists, but completely functional offline without