At Apple, Matuschak was much better at engineering work than he is now at research work, because engineering drew on deeply familiar knowledge in a well-understood domain requiring flow and execution, while research requires constantly answering 'what should be done' questions that demand synthesizing unfamiliar literature.