Discuss Production Outage

Ran out of log space

Performance Improvements

Immediacy

First load performance

searxng

crawl4ai

Lots of questions

Agents creating tools

Agents calling tools

Agents exposing tools to other agents

Tools require input/output schemas

We will understand other systems, of course

Datomic Schemas + Roam Research + Datalog

Tana (similar to Notion DBs)

Codegen is required. We don't want to manually create JSON schemas

User defined schemas with GUI, how to connect with hardcoded schemas

Custom blobs and resources (resource has metadata+hierarchal CRDTs)

We want triplets that are part of the doc change DAG, OR external triplets

Probably want to preserve document/resource semantics in a more generic way

Difficult considerations between static/hardcoded features and dynamically loaded mechanisms

Our schema system should be forgiving of "invalid" data

Let's move on this! We have to be careful about breaking changes

Enhance queries against our current schema

Query based on arbitrary metadata!

Filtering+Sorting can happen:

SQLite layer

Daemon/Golang

TS API

Dynamic/Generic Indexer for arbitrary efficient queries on document and block attributes.

WASM/WASI/Extism ?

Related to the dynamic querying system

Related to agent tools