Yes, that really resonated with me. It applies to performance improvement too. Performance improvement is finding ways for your future approach to compete with your current one.

If you look for ways to improve 10x you might find a way that should improve performance 3x, which when tested by brutal reality improves performance 2x. Still a good result.

But if you aim for a 2x performance improvement you’ll find a solution that should deliver 1.5x. And by the time you discover its practical weaknesses you’re back where you started.