---
title: "Agent Memory Is a Solved Problem. Agent Learning Is Not."
category: "talks"
date: "2026-07-01"
time: "3:20pm-3:40pm"
track: "Expo Stage 1 NE"
room: "Expo Stage 1 NE"
speakers: ["Karthik Ranganathan", "Heather Downing"]
sourceLabels: ["Official conference schedule", "Public YouTube metadata"]
scheduleTrack: ""
scheduleRoom: "Expo Stage 1 NE"
scheduleLabels: ["Expo Stage 1 NE", "session", "confirmed"]
---
# Agent Memory Is a Solved Problem. Agent Learning Is Not.

## Conference Context
- Date/time: 2026-07-01 · 3:20pm-3:40pm
- Track/room: track TBD · Expo Stage 1 NE
- Speaker(s): Karthik Ranganathan, Heather Downing
- Session type/status: session · confirmed

- Track: track TBD
- Room: Expo Stage 1 NE
- Session type: session
- Status: confirmed

## Session Description
The failures that break multi-agent systems are not reasoning failures, they are handoff failures. One agent works something out and the knowledge dies in its private context, because the only thing that crosses the boundary is output. Memory made each agent better in isolation and changed nothing about what the group knows. The missing primitive is supervised promotion: a deliberate decision about which private learning is worth sharing, moved into common knowledge with the reasoning attached, so trust survives the handoff. Today a human makes that call, and promoted knowledge resolves on read, in any tool, with no retrain or reindex. Those calls are also the training signal for what comes next: orchestrator agents, trained on what matters to the people they serve, that promote on their own. This talk covers how our collective knowledge grew as we approached memory promotion, including what the first build got wrong, and a live look at it working between humans and agents.

## Media Evidence
No related AI Engineer channel video found yet.

## Evidence Graph
This evidence graph is generated from currently linked source material: official schedule text, related video pages, cached transcripts, visible slide text, dense/reconstructed slide pages, and AI slide-classification audits.

### Media Signals
No linked video, transcript, or slide source has been attached yet.

### Agent Reading Notes
Use these signals to refine the synopsis, topic links, people/company context, and method notes. If a source is a related external video rather than an exact official recording, keep it framed as supporting evidence.

## Transcript Status
No official session recording transcript was found by exact title match on the AI Engineer YouTube channel during this run.

## People
- [[karthik-ranganathan]]
- [[heather-downing]]

## Notes
- Pending transcript synthesis when an official recording or confirmed matching video is available.

## Synthesis
### Synthesized Breakdown
# Agent Memory Is a Solved Problem. Agent Learning Is Not. ## Conference Context - Date/time: 2026-07-01 · 3:20pm-3:40pm - Track/room: track TBD · Expo Stage 1 NE - Speaker(s): Karthik Ranganathan, Heather Downing - Session type/status: session · confirmed - Track: track TBD - Room: Expo Stage 1 NE - Session type: session - Status: confirmed ## Session Description The failures that break multi-agent systems are not reasoning failures, they are handoff failures. One agent works something out and the knowledge dies in its private context, because the only thing that crosses the boundary is output.

### Speaker And Company Context
- [[karthik-ranganathan|Karthik Ranganathan]] — Co-founder and Co-CEO at [[yugabyte|Yugabyte]].
- [[heather-downing|Heather Downing]] — Developer Advocate at [[yugabyte|Yugabyte]].

### Topics Covered
- [[agent-security]]
- [[coding-agents]]

### Derived Links And Source Material

### Novel Concepts / Clever Methods
- No highlighted novel concept has been detected yet.

### Evidence Boundary
This synthesis is based on the official schedule and linked source pages. It should be revisited when exact session recordings or transcript-backed secondary sources are available.