Looking around a bit more, it seems that people have had good results using Wolftooth Roadlinks with MTB derailleurs in 1x8 and 1x9 drivetrains to drop the top jockey wheel down enough for there to be no rub with the 40 tooth cog. That would vary with the different derailleur hanger lengths of different models of bike.
Going to a 1x9 drivetrain remains only a possibility, but it is interesting to see how contentious the whole idea is. I only really came around to the idea myself after playing around with the HTML gear calculator and seeing just how much redundancy there is in a triple drivetrain. It was also quite unexpected as to how few gears you potentially lose off the top with various configurations. I guess I have accepted that the idea has merit, but I do understand why some don't.