No Memory, No Harness: Why the Database Is the Last Line of Defense — Kay Malcolm, Oracle
Source
youtube.com
Author
AI Engineer
Date
Context
Faster individual coding can leave a distributed team struggling with the same handoff problem: colleagues receive changes without the reasoning behind them. A conference account of a transatlantic engineering team uses that problem to argue for shared AI agentAn AI system that doesn't just answer once but works toward a goal in a loop — taking actions, reading the results, and deciding what to do next.Full definition → memory that preserves decisions across sessions and people.
The proposed benefit is continuity. Recording what happened, which procedure was followed, and what a decision meant gives the next person or agent context windowThe maximum amount of text a model can consider at once — its working memory for the current conversation or task.Full definition → to resume the work. The speaker also argues for consolidating storage to reduce reconciliation between separate systems. These are design arguments for preserving team knowledge, rather than measured productivity results.
Terms in this piece · Glossary
AI agent — An AI system that doesn't just answer once but works toward a goal in a loop — taking actions, reading the results, and deciding what to do next.
context window — The maximum amount of text a model can consider at once — its working memory for the current conversation or task.
videoWho Is Making This Request? Agent Auth 101 - Kamil Potrec, Keycard
videoStop Guessing Your vLLM Config: Benchmarking LLMs on Runpod - Chris Vela, Runpod
videoStop Giving Your Agent Tools. Give It a Sandbox - Mikyo King, Arize AI
videoWe Kept Rebuilding Our AI Agent. Here's What Survived - Fuad Ali, Arize AI
