Activity + Discussion panel fixes documentToText Activity Filtering Feed sorting Video

Week: @2025-10-21: Hypertuesday. During the past week I've been doing work with the feed and trying to find the memory leak. SQLite Memory leak Feed Next Week

Week: . During the past week I've been switching between debugging random issues, and continuing the work on . SQLite Memory Leak Private Documents Next Week

Email Notifications System Documentation Current User Stories Implementation Receive Email Notifications please read this document with user stories for email notifications. Current Limitations Future Development I think we can bring back the accounts table with a flag to determine if the (accountId, email) entry is a subscription or a linked account to email. -- Add back the accounts table with email linking CREATE TABLE accounts ( accountId TEXT NOT NULL, email TEXT NOT NULL, isAccountLinked BOOLEAN NOT NULL DEFAULT FALSE, createdAt DATETIME NOT NULL DEFAULT CURRENT_TIMESTAMP, notifyMentions BOOLEAN NOT NULL DEFAULT FALSE, notifyReplies BOOLEAN NOT NULL DEFAULT FALSE, notifyDocChanges BOOLEAN NOT NULL DEFAULT FALSE, notifySiteDiscussions BOOLEAN NOT NULL DEFAULT FALSE, PRIMARY KEY (accountId, email) ); The only issue is that email preferences are different for own account and subscriptions. Because of that I would instead propose creating a separate table for linking an account to email. CREATE TABLE email_account_links ( email TEXT NOT NULL, accountId TEXT NOT NULL, createdAt DATETIME NOT NULL DEFAULT CURRENT_TIMESTAMP, notifyMentions BOOLEAN NOT NULL DEFAULT FALSE, notifyReplies BOOLEAN NOT NULL DEFAULT FALSE, notifyDocChanges BOOLEAN NOT NULL DEFAULT FALSE, notifySiteDiscussions BOOLEAN NOT NULL DEFAULT FALSE, PRIMARY KEY (email, accountId) ); And we could add creation of linking tokens to allow verifying the email address in the welcome email. CREATE TABLE email_account_links ( email TEXT NOT NULL, accountId TEXT NOT NULL, linkingToken TEXT UNIQUE NOT NULL, expiresAt DATETIME NOT NULL, isVerified BOOLEAN NOT NULL DEFAULT FALSE, verifiedAt DATETIME, -- NULL until verified notifyMentions BOOLEAN NOT NULL DEFAULT FALSE, notifyReplies BOOLEAN NOT NULL DEFAULT FALSE, notifyDocChanges BOOLEAN NOT NULL DEFAULT FALSE, notifySiteDiscussions BOOLEAN NOT NULL DEFAULT FALSE, PRIMARY KEY (email, accountId) );      

Some time ago we enabled the ability to create Web Crypto keys in the browser on all Seed sites, for users to be able to sign their comments right on the site they are visiting. Previously they were redirected to hyper.media because this was the only place where their identity key lived, as Web Crypto keys are tied to a single domain and can't be shared securely. Now, this creates fragmentation. Essentially you have a separate account on each web site you are commenting on. Our vision for solving has always been the idea of linking your keys. You account is essentially the set of all the keys you have on you different devices and web sites. So during the past few weeks I've been working on implementing key linking. We already used to have to link desktop keys to hyper.media, but it wasn't suitable as-is for multi-origin world we now live in. So, I'll explain some of the changes I did. It All Begins on the Web Linking Instructions Conclusions

Update: Next week I will continue working on notifications refactor and implementing new subscribe feature for readers without an account.

Problems Solution Rabbit Holes

Problem Solution Scope Rabbit Holes No Gos

Subdocuments are part of the content of its parent

Installation Deployment Script Architecture Operational Modes Configuration Management Dependencies Cron Configuration

Full page Search

Import content into seed ecosystem

I was watching this video from Alan Kay. I was going to add some of his ideas to the text, but I can't because the video stops playing when I edit the document. Is there a reason for this? How easy is it to play YouTube videos in Drafts?

Yes — it’s possible, and it’s a pretty natural extension if you’re willing to treat “documents” as a universal substrate and add a few missing pieces (identity, types, security, and interoperability). A document-CRDT gives you the hardest part for the web already: offline-first, concurrent edits, and deterministic convergence without central coordination. Here’s what it takes to turn a document CRDT system into a distributed object framework.

Following a conversation I had with  at lunch, about limiting interactions to a single action—Follow—I don’t believe this approach works. We have accounts and sites. Although they share a common ID, in practice, they represent different things. An account signs content. For example, my Gabo account may sign content across many different sites. When another user follows my account, they want to follow what my account is doing across the entire hyper.media network. A site, on the other hand, represents a shared space of activity. When I join a site, it’s because I’m interested in participating in the collective activity taking place there, involving many different accounts. If I subscribe to or join Ethosfera, I want to follow what the Ethosfera community is saying to itself as it pursues a shared goal. In this case, I’m subscribing to a knowledge repository and an ongoing conversation—not to a specific user. Because of this, a single “Follow” action is insufficient. Following an account and joining a site are fundamentally different intentions and should be treated as such.

What if we include how a document shows the Document Metadata Block and the content, comments, and collaborators configuration within the formatting and viewspecs settings logic?

https://x.com/emollick/status/2020303173362012667 Ethan Mollick @emollick I think agentic AI would work much better if people took lessons from organizational theory, which has actually spent a lot of time understanding how to deal with complex hierarchies, information limits, and spans of control. Right now most agentic AI systems seem to pretend that models have basically unlimited ability to manage subagents when that is clearly not true. We need measures of spans of control for AI. A human tops out at less than 10 direct reports. I am pretty sure that 100 subagents is too much for an orchestrator agent - suspect we need middle management agents (yes, I get it, insert middle management joke here). Similarly, we need more attention to boundary objects. These are what is handed between groups (marketing to IT to sales) in organizations to convey meaning as a project crosses group boundaries, like a prototype or a user story. Right now agents pass raw text & maybe code back and forth. Structured boundary objects that multiple agents of different ability levels can read and write to would solve a huge number of coordination failures & reduce token use. I also think about coupling, which is how tightly units inside organizations are bound. Most agentic systems are either too tightly coupled (every step needs approval) or too loose (Moltbook). This tradeoff is well-studied in organizations, I bet a lot would apply to agents. Other known issues like bounded rationality also apply, I suspect. Everyone is rushing towards the (terribly named) agent swarm, but the issue won’t just be how good the model is, it will be org design choices. I am not sure the labs see this, but we definitely need a lot more experiments with organizing agents done by people who understand real coordination issues.

@cesar shared the idea of how chrome works and I liked it!