Make Tmux the OS
A long-time terminal user imagines an operating system UI built around tmux-like persistence and keyboard-centric window management, sparking debate over whether this vision just recreates tiling window managers and existing tools like i3, Niri, or Emacs. Commenters split between valuing highly configurable, power-user setups and arguing that mainstream adoption requires friendlier defaults and less manual configuration, perhaps aided by local LLMs. Underneath runs a broader argument about filesystems versus app-specific data silos and whether replacing traditional file-based organization with task- or app-centric models actually benefits users.
Tmux, multiplexers, and persistence
- Several comments focus on persistence across sessions and machines as tmux’s main value.
- Plugins like tmux-resurrect and tmux-continuum are suggested to restore layouts after restarts.
- Alternatives like zellij and other custom terminal multiplexers are mentioned; some note HTML email and rich content as friction points.
- A few report abandoning tmux entirely after moving to newer window managers or integrated terminal solutions.
Tiling window managers vs “tmux as OS”
- Many argue the article is really describing a tiling window manager, not an OS change.
- i3, sway, hyprland, Niri, EXWM, Ratpoison, various twms, and similar tools are cited as already solving most layout and navigation issues.
- On macOS and Windows, equivalents like yabai, Aerospace, OmniWM, Nehir, komorebi, and others are discussed, with mixed feedback on polish, stability, and complexity.
- Some say the author could solve their own problems with a bit of config; others stress config is a barrier for non‑technical users and wide adoption.
Vertical tabs, trees, and task flows
- Arc’s vertical tabs trigger debate; some resent that simple vertical lists became trendy while older tree-structured tab extensions offered richer task grouping.
- There’s enthusiasm for nested/tab-tree models that map to “task trees” (e.g., expanding/collapsing a whole HN-reading session), though support is now niche.
Filesystem vs app silos
- Strong pushback on the idea that filesystems are a “weak abstraction.”
- Pro-filesystem comments emphasize portability, neutrality, and the ability to switch apps without losing access to data.
- App-centric storage (Apple’s Photos/Notes/iCloud style) is criticized as opaque and hard to manage or diff; others reply that many app internals (caches, DBs) reasonably don’t expose raw files and that sandboxing and cloud sync push toward app databases.
- A broader “noun vs verb” framing appears: data‑centric filesystem model vs app‑centric model, with most here favoring the former.
Productivity, UX, and “normal users”
- Some insist UI is taste-driven “art,” and Unix-like mix‑and‑match plus config files (possibly generated by LLMs) is a long‑term win.
- Others argue basic competence in window and task management should be a professional expectation; chaotic desktops waste everyone’s time in collaboration.
- Counterarguments note that thinking, design, and code quality matter more than power‑user speed; minor window or typing inefficiencies are often irrelevant.
LLMs and configuration
- Multiple comments propose using local or bundled LLMs to generate window manager configs or let users describe their ideal desktop in natural language.
- Skeptics point to current LLM unreliability and caution against making them foundational to the desktop experience.