Exploring the Reality of Autonomous AI Development Agents in Production
The post discusses the feasibility of using orchestrated AI developer agents to autonomously build and maintain software under human supervision. The author is skeptical and seeks real-world examples to separate hype from reality.
Why it matters
Understanding the current state of autonomous AI development agents in production is crucial to separate hype from reality and guide future advancements in this area.
Key Points
- 1Debate around the possibility of running autonomous AI dev agents in production
- 2Looking for examples of multi-agent setups, long-running systems, and minimal human input
- 3Interested in understanding the setup, level of autonomy, failure points, and scalability
Details
The post raises the question of whether it is possible to use coordinated AI developer agents to autonomously build and maintain software, with only high-level human supervision. The author is skeptical of this claim and is seeking real-world examples to assess the feasibility of such systems. They are particularly interested in understanding the technical setup, the level of autonomy achieved, the failure points encountered, and whether these systems have been able to scale beyond small experiments or toy projects. The post also emphasizes the need for examples of multi-agent setups, rather than just Copilot-style individual assistance, and systems that have run for extended periods, not just one-off demonstrations.
No comments yet
Be the first to comment