Rethinking the R&D Budget
AI should change how we allocate the R&D budget, including the profile of the people we choose to hire.
I would increasingly allocate that budget toward engineers who combine strong technical judgment with product thinking. People who can understand a customer problem, make architectural decisions and carry a feature through to something a customer can actually use.
These people have always been valuable. As implementation becomes faster, I believe their relative importance grows. A team can produce much more code while still spending most of its time waiting for someone to define the problem, resolve a dependency or decide whether the result is good enough.
Adding more people to that process can also add more coordination.
A few implications for how I would organize an early-stage R&D team:
Give engineers broader ownership. An engineer should understand why a feature exists and take responsibility for its implementation, testing and readiness for review. The work should arrive at a reviewer with the important edge cases already considered.
This still requires independent review and specialist input where appropriate. But there should be a clear owner who carries the feature across these stages, instead of a series of handoffs where each person is responsible for only one small part.
Reconsider the default sprint length. For many well-scoped features, I would aim for a two-day execution cycle. The team should be able to build, review and learn without waiting for the end of a two- or three-week container.
Architecture, migrations and work with substantial dependencies will often require longer cycles. The point is to choose the cycle based on the work. A familiar planning ritual should not become the minimum amount of time it takes to deliver something useful.
Give product a clear monthly coordination role. Shorter cycles create their own problem: a team of capable people can move very quickly in slightly different directions.
Product leaders need to own meaningful monthly milestones and make sure these individual efforts fit together. Which customer outcome are we advancing? What must be ready together? Where does one team’s work depend on another’s? What should we stop building?
That shared direction becomes more important as engineers gain more autonomy.
The budget implication is straightforward. I would be willing to pay more for this combination of product judgment, technical depth and ownership, even if that means hiring fewer people.
The question I would ask before adding another R&D position is how much more independent, customer-relevant progress that hire will make possible.
_edited.png)