Have you noticed that every conversation in tech leadership now circles back to AI, and that none of those conversations actually feel resolved? That's because AI adoption is not a tooling problem. It is an organizational redesign happening in real time, whether you lead it or not.

In this episode, Kevin gets off the fence about a topic he has been deliberately avoiding. He walks through how his thinking about AI in engineering organizations has evolved over the last two years, and introduces a four-layer model for understanding where AI adoption actually breaks down: tooling, process, structure, and judgment. Most companies are working at layer one and quietly ignoring the rest.

Kevin also breaks down the four types of people you will encounter on your team during this transition (eager adopters, skeptics, the quietly worried, and the early adopters who fell behind), the new and growing problem of CEOs handing engineering teams "finished" AI prototypes and expecting them to ship, and why the leadership skill we need most right now is judgment about output quality, which nobody is hiring for yet.

If you are an engineering leader trying to figure out where to put your energy, this episode will help you stop solving tooling problems with more tools and start identifying the layer where your real work actually lives.

• Listener Survey: https://forms.gle/JVeKHsHJKhEM3dvK6
• The book: It Depends: Writing on Technology 2012-2022 (https://itdependsbook.net)
• Your host: Kevin Goldsmith (https://kevingoldsmith.com)