Tear down the walls!
We started our computing journey in the '80s and earlier with a very limited amount of CPU and Memory. We built abstractions to deal with these limitations. Such as having specific file types that can only be opened by specific applications. Nowadays compute and memory seem endless. Some of these limitations have been lifted. Some we are still stuck with. Some we have made worse.
The issue that got a lot worse, and one that has been bothering me a lot, is the silos we created between all of our information. It stated with the word document referencing the Q4 numbers to be seen in excel. An abstraction needed due to the fact that an 7MHz processor1 just could not handle a 16k row spreadsheet and perfect rendering of a 200 page report at the same time. But at least they were visually sitting next to each other in a folder. Through multiple steps of evolution this mutated to two links on a notion page, no one knew existed. Strictly speaking these two resources are still found next to each other, but physically and visually we couldn't have separated them more. By making them only accessible, not only within a certain set of applications, as we used to, but by locking them up in an external SaaS providers data-center. They now live in different tabs, on different websites, from different providers. Each provider with their own information navigation rules, access permission settings, and lack of opportunity to properly cross reference/embed infromation. With nothing left on your system, not even "spotlight" can help you find all your files anymore. You might forget that you own that information, and into which system you have entered it, but with a little bit of luck you will be reminded of it's pure existance by an invoice you received at months end in your inbox. To help us not forget where we put our information, or simply because we are currently working on it at the moment, we don't close those browser tabs. Leaving us with a sea of tabs.
More by accident a solution to reunite all your isolated information is emerging. To make LLMs really useful, they need to know what you know. So all the SaaS companies are now opening up and offering your data back to you in form of a MCP connection – all your data is finally connected by an uniform interface. But this is not for your consumption. But for the consumption of LLMs. Yes you can ask a chatbot a question and it will happily answer you. But this won't show you all your information belonging to the same project in one place. Instead you will tirelessly need to prompt for it. Every time you need it. And do not forget to ask it correctly otherwise critical information might be missing. As you can see the problem is not solved with this. It still is not easy to oversee, reorganise, take a stroll through it, get inspired or spark new ideas. Web apps, news pages, bookmarked blogs, are still alien amoung the rest of your data. No natural way to reference them, or store them along side your other resources. Visually and physically still separate from everything else you have.
Walled off.
Because no-one else has succeeded in tiering down these walls and create a unified interface for all your information, a friend and I have spent a year building towards this. Read about our experience here (article to come soon).
Footnotes
-
The first excel ran on a 1985 Macintosh, which was running an 7Mhz CPU at the time. ↩︎