Welcome to the Seed Team lobby.

This document is intentionally light on content. We use it as a shared place for open-ended team discussions: early ideas, quick questions, product thoughts, design notes, technical explorations, and conversations that do not yet need their own dedicated document.

We compare four configurations to assess whether the proposed framework maintains competitive verification performance while enabling richer explanations: (i) a simple prompt-based baseline equipped with web search; (ii) LoCal [

16], a multi-agent system based on logical-causal analysis, adapted to our tri-class setting; (iii) our proposed solution; and (iv) a variant of our approach, focused exclusively on linguistic agents (i.e., excluding web-search capabilities). To ensure fair comparison, all methods use the same Tavily retrieval mechanism and are evaluated under both knowledge-complete (with-KC) and knowledge-constrained (no-KC) regimes.

The actual conversations happen in the comments panel.

To follow or join the discussion, open the comments for this document. Each comment thread is a separate conversation. Some threads may be casual or exploratory, while others may later turn into issues, project notes, design documents, or more structured proposals.

Think of this page as the hallway: the document gives people a place to gather, and the comments are where the conversations happen.

Seth Godin on Marketing

Subdocument testing

Document Collection

Are we ok if, for document collections we don’t render the editor but render a special document collection UI that will edit the content of the document?

Critical decision about document Collections

we just released 2026.8.9, and this introduced the new document collections. In order for that to work, we are saving a new attribute called type with the value Collection. We need to decide this now since is in production and we don't want wrong data everywhere. Conclusion