A Conversation With Kun Chen @unsupervised-learning
A Conversation With Kun Chen  @unsupervised-learning
Uploaded January 2026 | Updated September 2026, 3 weeks ago
Check Out RovoDev here!
https://ul.live/atlassian

In this video we interview Atlassian's lead architect about RovoDev, an AI agent designed specifically for professional software teams that leverages organizational context to optimize the 84% of time developers spend on non-coding tasks.

What we talk about:
Building for Professional Teams: Why Atlassian designed RovoDev to solve the complex lifecycle problems of large professional teams—such as pipelines, testing, and security—rather than focusing on simple demos or hobbyist projects.

The 84% Problem: The finding that professional developers only spend 16% of their time coding, and how AI is being used to tackle the remaining 84% of time spent on non-coding friction and repetitive tasks.

Leveraging Organizational Context: How connecting the AI to internal data sources like Jira, Slack, and documentation allows it to retrieve the right context, enforce team standards, and prevent recurring incidents.

The Rise of Slower Agents: The trend where AI agents are actually becoming slower—not faster—because they are taking on more autonomous responsibility for running tests, self-correcting, and ensuring quality.

Future AI Capabilities: Upcoming concepts like "full context mode" and the idea of agents "sleeping" to reorganize their memory and extract insights from past sessions to improve future performance.

00:00 Introduction and Guest Background
02:25 Solving Problems for Professional Teams vs. Hobby Demos
04:04 The Architect's Role and Internal "Dogfooding"
05:52 Connecting Cloud Agents with the Command Line
06:41 Measuring Success: Reducing PR Cycle Time by 40%
08:19 Integrating with Backend Knowledge Bases and Guardrails
11:39 Data Sources: Capturing Context Beyond Public Models
14:05 "Paved Roads": Identifying and Enforcing Golden Patterns
22:02 Triggering Context to Prevent Browser Switching
23:59 The Evolution of "Full Context Mode" and Deep Research
25:17 The Future of Coding: Why Agents Are Getting Slower
30:04 Upcoming Features and Events

Subscribe to the newsletter at:
danielmiessler.com/subscribe

Join the UL community at:
danielmiessler.com/upgrade

Follow on X:
https://x.com/danielmiessler

Follow on LinkedIn:
linkedin.com/in/danielmiessler
A Conversation With Kun ChenThe Ideal Number of Corporate Employees is ZeroHow Much AI Do We Need?  (Answer: Weve Only Just Begun)A Conversation With Jeremy EplingTrumps New Robot Security DogsDebating AI with Marcus Hutchins: Part 1Unified Entity Context: The AI Stack Everyone is Building Without Realizing ItWall Street Journal on Small StartUpsWatch Out For This New Type Of ScamThe AI Bubble is Real, But Not The Way Most ThinkA Conversation With Brian VecciHow to give your Claude Code agents custom Eleven Labs voices
Unsupervised Learning |

A Conversation With Kun Chen

SHARE TO X SHARE TO REDDIT SHARE TO FACEBOOK WALLPAPER