Well-written and well-considered post, Paul. I think that your approach can achieve the goal, but not without a loss of velocity, which is a huge pressure at any startup, especially seed-funded companies with a relatively short cash runway. I'm a big proponent of pair programming, and think that when good developers pair, there is little to know loss of velocity. But pairing when serious mentoring is big goal is a lot slower in the short-term than forgoing the hire altogether. Start-ups who know they have a long runway can afford to go with your strategy. I personally wouldn't recommend it to most startups.