The author's central thesis is simple: 'tmux is the OS,' but made usable for normal people. Tmux is a terminal multiplexer that provides scrollable, session-persisted, detachable, and task-oriented window management. The author argues this model solves fundamental problems with contemporary desktop interfaces that have not kept pace with evolving hardware and usage patterns. # Problems with Current Window Management The author identifies several issues with how operating systems manage windows in 2026. Screen real estate constantly changes as users move between laptop and desktop setups—typically about six times daily. Different monitor configurations require different approaches: a single large monitor offers a different experience than multiple monitors, yet operating systems don't recognize which screen serves as the primary work display versus secondary reference displays. When undocking a laptop, windows collapse onto one screen, requiring manual resizing that wastes time and disrupts workflow. Beyond hardware changes, user behaviors vary significantly. The author describes three distinct groups: window maximizers who expand one window to fill most of the screen; near maximizers who use one large window plus smaller utility windows for status or chat; and those who carefully coordinate all windows on their screen. Additionally, nearly everyone maintains a distinction between 'public windows' (acceptable for others to see) and 'private windows' containing sensitive information that should remain hidden—a concept validated by research from 2004 that remains unaddressed by mainstream operating systems. # Browsers as Mini Operating Systems and the Weak Filesystem A second major shift is that most software now runs within browser tabs. The author observes: 'In 2026 everybody quietly agrees that most of your software runs in a browser tab. These web applications will cover a wide variety of use-cases that have historically been owned by local applications.' Yet operating systems treat browsers as normal applications rather than recognizing them as virtualized operating systems hosting multiple complex applications within tabs. Compounding this issue, the traditional filesystem concept is weakening. Application data increasingly lives in proprietary silos: notes in Apple Notes belong to Apple Notes rather than the filesystem, texts reside in iMessage's database, and Teams/Slack content remains locked within those applications unless manually exported. Even when Firefox displays a PDF, the file doesn't integrate with the host operating system without explicit user action. As the author notes, this creates broken chains where 'something writes a file, the user opens an app on it, the app writes it back, the file gets sent to someone else'—a workflow that assumed files lived where the operating system could see them, an assumption no longer valid. # Historical Context and Previous Attempts The desire for better window management is not new. The author references a 1986 paper describing 'Rooms: the use of multiple virtual workspaces to reduce space contention in a window-based graphical user interface,' which conceptualized windows as memory pages and task switching as knowledge faulting. More recently, the WindowScape paper proposed a task-oriented window manager that abandoned explicit grouping in favor of a photograph metaphor—allowing windows to appear in multiple contexts without requiring upfront placement decisions. Various implemented solutions have fallen short. Tiling window managers maximize screen real estate through predefined layouts but remain 'way too hard to use. Like an order of magnitude too hard for normal people to use,' often requiring users to edit configuration files before achieving basic functionality. The author specifically cites Niri (https://github.com/niri-wm/niri) as an existing scrollable tiling window manager that addresses core needs but requires improved accessibility. Attempts to treat browsers as operating systems have proven misguided. Both Arc browser and Chrome OS attempted to make browsers more OS-like rather than elevating web content to first-class citizens within the host operating system. As the author contends, 'If a web application is operating as an application, it should be its own window. If it is complimentary to another application, it should be a window associated with another application and be allowed to contain many tabs.' Historical efforts by major platforms also failed to gain traction. Windows Timeline (2017–2019) tried to organize cross-app activity but required opt-in logging that felt invasive. Windows Sets (2018–19, canceled) grouped apps and web content into task-based tabbed sets but saw limited adoption. macOS Stage Manager (2022) emphasized rapid switching between single tasks—suited for iPad but inadequate for desktop workflows requiring simultaneous use of multiple applications. KDE Activities have offered task-scoped desktop state for over a decade but remain hyper-manual and complex to configure. Progressive Web App installs provide browser-chromeless windows for web applications but retain a clunky user experience. # The Path Forward The author's vision combines the strengths of existing approaches while addressing their weaknesses. A scrollable tiling window manager would allow users to configure distinct workspace arrangements (e.g., private work on the left, public work on the right) and navigate between them spatially rather than through complex modal interfaces. This solves both the screen real estate problem (infinite workspace without subdivision) and the public/private window dichotomy through simple scrolling. Crucially, this approach must prioritize accessibility for normal users, avoiding the configuration complexity that has doomed previous tiling window managers. The author admits lacking UI/UX design skills but frames the proposal as a thought exercise to inspire others: 'I'm making this mostly as a thought exercise and hopefully a prompt to get other people to think about this same problem.' While the technical foundations exist in projects like Niri, the primary barrier remains usability—transmitting the power of tmux-inspired window management to users who neither want nor should need to edit configuration files to achieve productive, personalized desktop experiences. Uncertainties remain regarding how such a system would handle legacy applications designed for traditional windowing models, whether the spatial navigation metaphor would scale to complex multi-project workflows, and how to balance discoverability with efficiency for novice users. The author does not provide specific implementation details or timelines, positioning the article as a catalyst for discussion rather than a technical specification.