"Claude Code displayed strong preferences for particular developer tools when given open-ended implementation tasks, introducing another distribution channel relevant to the DevRel discussion from part two."
Ong and Vikati studied 2,430 Claude Code responses across three models, four project types, and twenty tool categories, obtaining identifiable primary tool choices from 2,073 responses. Their results provide a measurable connection to the earlier discussion of DevRel and product visibility, showing how an AI coding assistant can establish preferred tooling defaults when users delegate implementation choices.
View Source