Gather.dev

Engineering leadership for the agentic age

What we're learning about the agentic transition.

Every month, hundreds of senior engineering leaders convene at Gather events to figure out what's actually working for AI adoption in engineering. The conversation has moved past copilots. We're all thinking about how to go from the modest gains of augmented engineering to the multiples that come from deploying a reliable, trustworthy agentic SDLC.

This is where I share what I'm learning.

Three conversations, one picture.

I'm interviewing three groups of experts to build the most complete view of what's happening:

Agentic pioneers

Building autonomous software factories. What are they building, what's working, and where are the limits?

Engineering leaders

Managing agentic adoption in organizations large and small. How are they scaling pilots, supporting their teams, and measuring impact?

Dev tool founders

Solving specific problems that impact the adoption of an agentic SDLC. What are they seeing from inside the toolchain?

The patterns I find across all three groups become articles, book chapters, and community discussion topics.

The O'Reilly book.

I'm writing Scaling AI Adoption in Engineering for O'Reilly. It's built on real case studies from real companies. The Substack is where the ideas get tested before they become chapters.

Building what I write about.

I'm building my own multi-agent orchestrator with durable execution, configurable pipelines, and a compounding memory system. I'm moving my role from instigator to advisor as the agents take on more of the business. I write about what I learn along the way: what works, what breaks, and what surprised me.

What to expect.

  • Agentic engineering is the future of software development, and it will reshape organizations, not just tooling
  • This transition requires hands-on leadership from CTOs. You cannot delegate it to your head of platform
  • If you are not getting great results from AI, it is you, not the models. The investment is in learning to specify, verify, and continually iterate
  • Less is more. I will only post when there is something to say, and I will maximize information density so you do not need your AI to summarize it for you

Everything the community learns gets synthesized and shared here. The knowledge flows out. The rooms stay intimate.

Want to be in the room? Learn about the community.