Memorable found a way to optimize memory with embeddings instead of more tokens which is a powerful new way to do memory

1// Each successful run is distilled into procedure nodes: the conditions under which a workflow applies, the actions the agent took, and the outcomes it produced. 2// When a task arrives, Memorable retrieves relevant nodes and constructs an ordered path through them. Because procedures can be composed, the agent can execute workflows that no previous agent has run end to end. This essentially turns agent reasoning into a graph search problem. 3// The Memorable CLI is live across Claude Code, Codex, Cursor, and Claude Cowork. Memorable is also deployed in @garrytan's GBrain, GStack, and YC's QM Harness. 4// The long-term goal: systems agents whose operational competence compounds across sessions. Models will change. Context windows will reset. The operational experience your agents earn should persist and remain yours. Check us out at: memorable.sh Demo: youtu.be/3m9I1zHSMPk

I'm running my company with it

hey! I'm presenting on how founders can win with Grok Bot tmr at 4PM PST. you should probably come watch (even if you're not a founder). we're ~completely full on in-person capacity in SF, but live streaming it with @bot. signup link in comments. LFG

We’re inviting one Grok Bot user to a Starship launch. You could win the invite by sharing the ways you've integrated @Bot into your work. The invite includes a plus one. How to submit and some examples we love: