Charity Majors, the founder and CTO of observability platform Honeycomb, recently joined the Scaling Dev Tools podcast for a wide-ranging conversation about where software development is headed in an AI-saturated world. The episode digs into one of the most pressing questions facing engineers today: what happens when the cost of generating code approaches zero?

When Code Becomes Cheap, What Matters?

Majors argues that as AI tools make it trivially easy to produce vast quantities of code, observability transforms from a nice-to-have into the definitive source of truth for your systems. "If you can't observe it, you don't own it," she notesβ€”a provocative stance that flips traditional thinking about what deserves engineering investment. The conversation explores how teams caught up in AI-assisted development often ship faster but lose visibility into what's actually running in production, creating a dangerous blind spot.

Taste Still Beats Automation

Perhaps the most contrarian take in the episode: great products still depend fundamentally on human taste. Majors pushes back hard against the narrative that AI will soon replace product thinking and design sensibility. "Code generation is commoditizing," she explains, "but knowing what to build and whyβ€”that requires judgment machines simply don't have yet." This echoes a growing sentiment in engineering circles that the bottleneck has shifted from implementation to decision-making.

Fast Feedback Loops as Competitive Advantage

The episode also covers how high-performing teams use fast feedback loops to ship frequently without accumulating technical debt or breaking production. Majors breaks down specific practices around testing, deployment, and monitoring that separate teams shipping confidently from those perpetually afraid of their own code. Engineering leaders looking to build credibility with business stakeholders get tactical advice on translating technical metrics into business value language.

Key Takeaways

  • Observability becomes your source of truth when code generation is cheapβ€”invest accordingly
  • Human taste and product judgment remain irreplaceable despite AI advances
  • Fast feedback loops are the differentiator between teams that ship fast and teams that ship recklessly
  • Technical founders must learn to speak in business terms to build organizational credibility

The Bottom Line

Majors isn't buying the hype that AI makes engineering effortlessβ€”she's focused on what actually gets harder when you have infinite code generation capacity. That's a perspective worth paying attention to, coming from someone who built her career making complex distributed systems debuggable.