Why Byblos
Value Thesis
Problem Statement
The modern browser is phenomenally good at what it was designed to do. It loads webpages quickly, gets out of your way, and makes it easy to move from one destination to the next. If your goal is simply to find information, the browser works extremely well. The problem is that many important activities do not end when the page loads.
Buying a car, planning a vacation, choosing a school, evaluating software vendors, researching a medical condition, hiring an employee, investigating a technical problem, or furnishing a home all involve the same basic pattern. You gather information from many webpages, compare alternatives, discuss them with other people, and gradually arrive at a decision.
At that point, the challenge is no longer finding information. The challenge is managing everything you've learned. The browser's answer to complexity is almost always to open more tabs.
At first this works well enough. Then you have fifteen tabs. Then forty. Then eighty. You vaguely remember that one page had the best price, another had the feature you wanted, and a third looked promising but required more research. You know you discussed one option with your spouse and another with a colleague. You remember rejecting something important, but not why.
The webpages are still there. The context surrounding them is not.
Most people compensate by introducing additional tools. Notes move into one application. Discussions move into another. Reminders live somewhere else. Spreadsheets appear to track comparisons. Bookmarks preserve links for later. None of these tools are bad. They are just being co-opted to do something they were not designed to do natively: evaluate webpages.
Additional tools fragment the decision-making process. The pages live in one place, your notes live somewhere else, and the conversation that led to the final decision lives somewhere else again. Weeks later, reconstructing your reasoning can be harder than finding the information in the first place.
Consider a vacation search. You spend several evenings looking through Airbnb listings. Some are too expensive. Some are too far away. Some have incredible views. Some look perfect until you discover a noisy road nearby. A week later you return to continue the search. The listings are easy to find again. What is harder to recover is everything you learned about them. Which property was your spouse excited about? Which one had the hidden fees? Which one seemed expensive until you compared it to the alternatives? Which one did you ultimately reject, and why?
The same thing happens during research. Months later, you can usually find the article, paper, video, or source again. What is harder to recover is the thinking that surrounded it. Did you trust it? Did you disagree with it? Was it evidence for your conclusion or evidence against it? Did someone else on your team already review it? The page survived. The context often did not.
Solution: treat webpages as the primary object
Byblos starts with a different assumption: for many internet-native activities, the webpage is the thing people are actually working on. The notes, discussions, ratings, measurements, reminders, annotations, and decisions are not separate objects. They are context surrounding a webpage.
Instead of scattering that context across multiple systems, Byblos keeps it attached to the pages that generated it. The page that started a discussion remains connected to the discussion. The Airbnb listing that Alice loved retains the heart that she left on it. The restaurant that you rated 5 stars still has them months later.
Many of Byblos's features follow naturally from this idea. Pages can be organized into trees because pages often have relationships to one another. Annotations exist on pages because that is where observations about pages most naturally belong. Discussions exist on pages when the pages themselves are the topic. Sorting and filtering exist because large collections of pages eventually need structure.
Byblos is not a collection of independent features. It demonstrates the consequences of a simple idea: if webpages are the primary objects people work with on the internet, then the surrounding context should accumulate around them.