AI-assisted programming has advanced beyond simple autocomplete functions. Today, large language models can edit entire codebases, coordinate prolonged tasks, and work collaboratively across multiple systems. As these capabilities grow, the central focus in software development is transitioning from code writing to orchestrating work, managing context, and ensuring a shared understanding among numerous agents.
Steve Yegge, a seasoned software engineer and writer, has influenced many developers with his writings. Over the past year, he has been delving into agentic software development, creating tools like Beads and Gas Town to test multi-agent coordination, shared memory, and AI-driven workflows.
In this discussion, Steve joins Kevin Ball to explore the progression of AI coding from chat-based assistance to advanced agent orchestration. They address the technical and cognitive challenges of managing fleets of agents, discuss how task graphs and Git-backed ledgers alter the work landscape, and consider what these changes mean for software teams, tools, and the industry’s future.
Kevin Ball, also known as KBall, is the vice president of engineering at Mento and an independent coach for engineers and engineering leaders. He has co-founded and acted as CTO for two companies, started the San Diego JavaScript meetup, and organizes the AI inaction discussion group through Latent Space.
Please click here to see the transcript of this episode.