Why vague specs punish agents

If you are a manager, a lead engineer, or anyone growing into more responsibility, this throwback episode is built for you. We keep hearing the same question, now louder than ever: "Why can't this go faster?" AI and agentic coding have made the literal coding step dramatically cheaper, so product leaders reasonably expect the whole pipeline to speed up. But it hasn't—and in today's short, focused episode I explore why. The answer isn't new at all. It's the theory of constraints, and it has everything to do with friction you may not be looking at.

• Speed Isn't the Story—Friction Is: When a fast component gets introduced into the pipeline, the instinct is to celebrate the velocity. But pay attention to what comes after. The real question is what keeps work from naturally flowing faster, and that lives in the friction, not the energy you're pouring in upfront.
• The Universal Bottleneck: I rarely claim universal truths on this show, but here's one: anything that looks like a pipeline will have a bottleneck. If you're not paying attention to it, it doesn't matter how fast every other step gets. Faster coding just exposes where the constraint really sits.
• The Two Places Friction Shows Up: For teams fully adopting agentic coding, the bottlenecks cluster in two spots—requirements gathering at the front, and verification, validation, and testing at the back. Rushed requirements upstream create even more painful rework downstream.
• Why Agents Punish Vague Specs: Human engineers fill in gaps by being close to the work. Agents fill in gaps too, but sometimes incorrectly. If your requirements aren't detailed, the agent guesses, and you pay for it in review. Spend more time in the planning phase, not less.
• The Foundation You Build On: Agents glob extra code onto a weak structure—unnecessary models, redundant endpoints, patterns that don't fit. A code base organized with clear conventions, good documentation in your CLAUDE.md or AGENTS.md, and dependable patterns lets the agent discover and extend rather than guess and hope.
• Specification and Validation Are Bookends: Good requirements translate directly into good tests. Acceptance criteria on one end, changes in the middle, validation on the other end—directly connected. Poor specification sitting on a poor structure guarantees poor execution and poor validation.
• Reframing Your Objections: Think scope creep is the problem? That's a requirements issue. Think you lack the talent? That's a foundation issue—because the engineer's job now is to cultivate the foundation so generated code enriches it instead of toppling it.

This is not a new problem. We asked it of the internet, of web frameworks, of CSS. Now it's time to apply the same principles to agentic velocity: look at your requirements, your foundation, and your validation. Somewhere in those three is your bottleneck. I guarantee it.

📮 Ask a Question

If you enjoyed this episode and would like me to discuss a question that you have on the show, drop it over at: developertea.com.

📮 Join the Discord

If you want to be a part of a supportive community of engineers (non-engineers welcome!) working to improve their lives and careers, join us on the Developer Tea Discord community today!

🗞️ Subscribe to The Tea Break

We are developing a brand new newsletter called The Tea Break! You can be the first in line to receive it by entering your email directly over at developertea.com.

🧡 Leave a Review

If you're enjoying the show and want to support the content head over to iTunes and leave a review!