---
title: "Comment by Kartik N V J K on \"AI Agents Are Distributed Systems. We Just Gave the Problem a New Name.\" | Hashnode"
description: "The mapping holds, and the model-is-10%, harness-is-90% number is the part teams underweight when they pick a framework. The one that bites earliest for me is i"
url: "https://hashnode.com/posts/ai-agents-are-distributed-systems/6a53c8fb489799a821294947/comment/6a55336ef90af637f0354d4e"
---

[Discussion](https://hashnode.com/posts/ai-agents-are-distributed-systems/6a53c8fb489799a821294947)

Comment [AI Agents Are Distributed Systems. We Just Gave the Problem a New Name.](https://hashnode.com/posts/ai-agents-are-distributed-systems/6a53c8fb489799a821294947)

[Kartik N V J K](https://hashnode.com/@kartik-nvj)

[Jul 13](https://hashnode.com/posts/ai-agents-are-distributed-systems/6a53c8fb489799a821294947/comment/6a55336ef90af637f0354d4e)

The mapping holds, and the model-is-10%, harness-is-90% number is the part teams underweight when they pick a framework. The one that bites earliest for me is idempotency, people who'd never ship a service without idempotency keys will happily let an agent retry a non-idempotent tool call, then debug the double-executed action. If agent memory really is an event log, do you key tool-call results by an idempotency token so a retry replays instead of re-executes?
