Code has never been easier to write. With AI copilots and agentic coding tools, spinning up features feels almost effortless. But production systems don’t run on vibes, they run on reliability.

In this episode of Humans of Reliability, Swizec Teller, author of the bestselling Scaling Fast, argues that writing code was never the hard part. The hard part is making it run. We talk about the “Feynman computer disease” of AI coding addiction, why babysitting agents may be the new productivity trap, and what SRE instinct really looks like when you’re debugging a distributed system at 2AM.

Swizec explains why reliability is ultimately about ownership and accountability (“sign it with your phone number”) and why SLAs, system thinking, and production experience matter more in an AI-driven world. We also explore whether LLMs will automate SRE work, what current benchmarks actually show, and why the future may require more engineers, not fewer.

If code is becoming probabilistic and cheap, reliability might be the last real differentiator.