Members-Only
Recent Talks & Demos are for members only
You must be an AI Tinkerers active member to view these talks and demos.
Weave: turning doomscroll into a 2nd brain
Explore Weave, a spatial canvas transforming saved content into a second brain. Discover how conversational AI surfaces surprising connections across diverse sources, offering substantive insights over speed.
Weave is a spatial canvas for the content you’re already saving: tweets, YouTube transcripts, notes, articles. Ingested sources get embedded (Gemini Embedding 2), stored in pgvector on Supabase, and rendered as a spatial graph where cosine similarity surfaces connections across formats and authors. Ingestion pipelines use Microlink plus oEmbed for tweets and tweet images, and Supadata for YouTube transcripts and Twitter video. A voice layer lets you query the graph conversationally. The LLM reasons over retrieved neighbors and pulls connections you wouldn’t have spotted yourself. Started as a project at the AI Tinkerers UI hackathon two months ago.