Our files are in the cloud now. If the browser is the OS, we need a new file system.
A place to save, share, and be up to date. Users should be able to quickly understand, with pertinent context, what's going on with their favorite services, websites, apps, and creators. Users should be able to have a high level of visibility on what other users, whether it's their friends, people within pertinent organizations, or relevant people they're interested in. Users should be able to take a website or app, and store its context to be able to "listen" to the updates made inside of those systems.
If our system is built correctly, it will the quickest and most powerful way for individuals, groups, and organizations to control, retrieve, and index their usage and context on the web. By doing so, we'll build the world's connected file system.
Most operating systems file systems have typically mimicked the rows and rows of file cabinets, stacks of folders, and piles of paper from insurance companies and government buildings. These objects and information were mostly static, a file was written and stored not to be updated, but as record keeping. As computers moved from mainframes to personal machines in our homes, that paradigm mostly remained as applications and how users interacted with and created on computers were mostly limited. How we use computers and what we do on the web has reframed what we do on our computing machines, and the mechanisms for storage and retrieval haven't kept up with that shift.
All of what we do digitally is moving to the web, from work (word → docs), where we talk (AIM → Zoom), where we play (pc → stadia), where we read (books → blogs), to where we create (adobe → figma). This means that the previous mechanism that exist for us to control and see how data is stored and retrieved in our digital systems, mainly the desktop's operation system, fails.
Turns out, the paradigm of a file doesn't stretch well on the modern web.
A modern file system acknowledges the web has grown to a short tail of apps (Notion, Figma, Docs, LinkedIn, Facebook, Linear, etc) and a long tail of text or multimedia heavy content (blogs, newspapers, tumblr, videos, pdf, etc). The former mostly contain their own form of context, and inherently have their own embedded "files".
The problem is that most use cases require context from multiple apps and content (links to flights, links to airbnbs, links to blog with travel suggestions, etc). The app files are often "live", where the context inside of each "file" updates because of user interaction or because the app system itself is updates. It's extremely expensive for a user to understand at any given time what's going on inside of these files quickly without opening the files. Lucky for us, static and active files use the http linking system, allowing us to extend general bookmarks.