Key Takeaways
- Tech circles hype a future where software engineers, designers, and PMs merge into a single "builder" role, but Josh Woodward sees the opposite at Google Labs.
- Erasing functional disciplines risks destroying the craft expertise and mutual respect that technical specialties build over years.
- Treat cross-functional AI skills like a college minor: team members should vibe code and prototype, but only to sharpen their core major.
- Falling execution costs shrink the ideal project team from 5-7 people down to agile pods of 2-3 specialists (at most 4).
- The best small teams behave like studio musicians: distinct virtuosos who team up to jam on specific tracks rather than generalists playing every instrument poorly.
The Myth of the Flat Builder
For the past year, startup founders have repeated the same prediction: AI coding tools will dissolve job descriptions. Engineers will write copy, designers will build backends, and product managers will ship production code alone. Everyone becomes a builder.
Woodward runs Google Labs, the Gemini app, and AI Studio. He watches this dynamic unfold across hundreds of prototype cycles, and he warns against the trend. “I think right now people are maybe overhyping like everybody's a builder and it's just this new job function and everything's kind of the same,” Woodward explains. “I actually feel like well at least what we're seeing is that people still have their specialty.”
When a company tells its employees that everyone does the same generic job, craft suffers. Great interface design requires taste, ergonomics, and visual restraint that an engineer prompt-engineering CSS rarely develops. Hard backend architecture demands deep knowledge of distributed systems and data contracts. “I feel like there's a risk if everybody becomes a builder that you lose a lot of respect and appreciation and obviously the expertise that comes from those specialties,” Woodward notes.
Keep the Major, Add the Minor
Woodward does not tell people to avoid AI tools. He encourages them, but with a strict mental model based on college coursework.
Your functional discipline is your major. A product manager must excel at prioritization, customer discovery, and strategy. A software engineer must excel at architecture, systems design, and code quality. AI tools allow you to pick up a minor in rapid prototyping or vibe coding.
Woodward gives this exact counsel to Associate Product Managers (APMs) entering Google Labs: “It's great you're vibe coding and you're making all these prototypes, but do that in service of becoming an extraordinary PM and don't lose that major.”
When a PM vibe codes an interactive mock on a weekend, that artifact should clarify user requirements for the team on Monday. It should not replace the staff engineer who knows how the database handles concurrency under load. The minor supports the major; it does not erase it.
The Two-to-Three Person Jam Session
If roles are not blending together, what actually changes when AI lowers execution costs? Team sizes shrink.
“We used to say like it's about five to seven people,” Woodward says. “Now it's about two to three people. Maybe four.”
Instead of large scrums with two frontend engineers, two backend engineers, a designer, a QA tester, and a PM, Woodward builds tight pods. These pods operate like experienced studio musicians: “You can almost imagine like super talented musicians who team up with different people and then they play different songs together.”
A bass player does not need to tune the drums, and the drummer does not need to sing lead vocals. Because both have elite timing and quick feedback loops, two or three musicians can record an entire record in a few afternoons.
What to Do With This
Look at your current sprint board and team rosters. Identify any project pod staffed with five or more people, and split it into pods of two or three specialists with clear primary majors. Tell your PMs and designers to prototype with AI, but evaluate them on their core functional craft rather than lines of prompt-generated code.