He did say "In principle". The key questions are: - Is it fit to fly as is? - How long does inspection take? - How long do repairs take? - How much could be automated? (If re-usability lowers the cost, there will be more flights, if there are more flights, per flight cost of automation is lowered) Stages are subject to up extreme and down extreme. Dragon's been going up and down for a while so I hope they're learning on how better to protect it. Getting the first stage recovered consistently would be huge for the learning process. Shame about the sea this time.
Elon hasn't exactly said this, but here's how it's probably going to go: 1) Lots of testing with initial prototype - Done 2) Soft water landing attempt without landing legs - Done (failure) 3) Soft water landing attempt with landing legs - Done (success) 4) Soft water landing attempt with landing legs close to shore 5) Soft landing on land - analyze and determine changes that need to be made to make it fully and rapidly re-useable 6) Tweak some parts and test said tweaked parts on the prototype until it works (repeat steps 5 and 6 for a few flights) 7) Refly a used stage (without a customer) (possibly repeat steps 5, 6 and 7 based on outcome) 8) Begin pre-flown rocket flights with customers 9) Continue to tweak as time goes on to reduce price Now start all over for the second stage.
Maybe same day turnaround for cargo and slower turnaround with more thorough inspection for manned flights? Though I image if you went back in time to the dawn of commercial jet powered aviation and suggested that some day we'd see same day turnaound for long haul flights you'd be looked at like you were mad.