Volume, Ambition, Clarity
There’s a ladder that every engineering team climbs with AI. The burnout posts, the productivity debates, the 10x discourse. That’s all the first rung.
I think of it as three stages: volume, ambition, and clarity. Where you are on that ladder changes what AI feels like and what you need from it.
Volume
This is where most teams are right now. You’ve learned the tools, found a rhythm, and you’re shipping two or three times what you used to. More PRs, more features, more tasks knocked out in a day.
This is where you burn off the backlog. You kick off 50 tasks you never had time to do, 1-shot them with a background agent, click around the branch, and ship it.
It’s also where the burnout lives. Steve Yegge called this the “AI Vampire” a few weeks ago. You’re doing more, you’re tired, and the value capture feels off. You’re putting in 1.5x the cognitive effort for 10x the output at the same salary. It feels weird.
But I don’t think it’s the whole story. Volume is the first stage, not the destination. The engineers I’m closest to have already moved past it.
Ambition
This is where it gets fun. You stop using AI to do more small things and start using it to do harder things. I talked with my old VP Eng counterpart yesterday, and he’s seeing a steady drumbeat of projects cutting their timelines by 50%. Those projects used to lose the prioritization game, because it’s better to do low-effort, medium-value stuff, or worse. Or, if it seemed important enough, they’d get started, stall because they were hard, and killed. High-effort, high-reward takes a certain kind of conviction to sustain.
But now? There’s time. You spend ten rounds in the planning ring with the AI. You explain everything you’ve picked up while you’ve been on this team, and it does it right. You fix things at the root instead of patching around them. Adoption is cheap! You send your team PRs and an explanation: we finally did the thing we’ve been talking about since the company was 50 people.
This is where AI starts to feel transformative, not just fast. The system starts to feel different, because it matches how they’ve always wanted it to look.
The engineers at this stage are the happiest people I talk to.
Clarity
Then you run out of ideas. You’ve burned through the wishlist.
And now you need to figure out what’s next.
This is a different problem than volume or ambition, because the obvious wins were obvious partly because everyone already agreed they mattered. Three years of shared context made prioritization easy. The next tier of improvements doesn’t come with that built-in consensus. Someone has to look at the codebase, which has changed enormously and fast, and figure out what’s worth building next.
The teams I’m watching reach this stage are the ones I’m most excited about. They’re not slowing down. They’re developing the ability to read their own systems clearly enough to keep getting bolder. What’s actually in the codebase now? What changed in the last three months, and what does that open up? Where are the new pressure points that didn’t exist before the ambition phase?
This is where AI-assisted development becomes a team capability instead of an individual one. A team that can see the system, find what matters, and move on it together at the speed of thought.
Most teams aren’t here yet. But the ones that get here are going to be extraordinary.
This is what we’re building at Tern. AI writes code. We help your team see what’s worth writing.