Great study on what actually predicts vibe-coding ability.
100 tertiary-level students completed measures of computer-science achievement, domain-general cognitive skills and written-communication proficiency, then took a vibe-coding assessment.
The study was preregistered, which matters because the intuitive answer here, that writing ability carries everything, is what a post-hoc analysis would tend to find.
Tasks were curated through an eight-expert consensus process and run in a purpose-built vibe-coding environment that mirrors commercial tools while allowing controlled measurement.
They find that writing skill predicts performance. So does computer-science achievement, and it stays significant after controlling for domain-general cognitive skills, so it is measuring something specific to CS training rather than general ability.
The practical question the authors point at is a curriculum one. How much time to spend teaching people to write good prompts, against how much to spend on computer-science fundamentals, for the people who will build software this way.