A valid contribution, but one that disrupts the project’s pace, isn’t seen as an improvement; it’s seen as a hindrance to the team.
In the long run, whoever slows down the group’s pace is no longer invited to participate in decision-making, and the expert’s input always arrives after the improvement is already implemented or nearly finished. Arriving late and out of sync turns a valuable contribution into a problem.
Hilary Atkisson Normanha, writing about how localization work is adapting to AI, precisely defines the trap: “We need four months to do a proper linguistic evaluation,” and the troubling part is that four months can be exactly how long a proper evaluation takes. The answer isn’t wrong, but it stalls the entire process.
What works best is a piecemeal approach: identifying where the result is most likely to fail, where it’s likely to be good, and what issues will truly be problematic if left unresolved.
It’s the same order of operations we advocate for in every project: we ask what problem is being solved before proposing what to build, and the expert we want to hire must adapt to the project’s pace to be effective.
Sources
Hilary Atkisson Normanha — Perfectionism Is a Localization Problem: Your Best Skill Has a Blind Spot
