I highly doubt that Musk would have spent half a billion dollars developing the Falcon Heavy if he thought he was going to make it obsolete a few years after its maiden flight.
that wasn't the original plan
FH was supposed to be quicker to develop and the original ITS (or whatever it was called) was going to be further out
instead FH development kept dragging out longer and longer and then they 'rightsized' ITS to bring BFR in sooner (among other reasons)
there was probably a bunch of sunk-cost consideration going into it, they had already invested so much and it only needed 'a little bit' more to finish, and then 'a little bit' more and finally 'a little bit' more
Musk said that he almost cancelled it three different times, so I think it's clear if they knew then what they know now, they never would have gone the FH route.
what kept him from cancelling it?
- sunk cost: or, if you prefer to look at the other way, the (comparatively) small cost to bring it to completion
- morale: they had invested tons of work into making this work, to just cancel it would have devastating impacts. Nothing's worse than devoting years of your life to something and then just having it thrown away
- deadlines: he needed something to compete with delta iv heavy and he needed it yesterday so he doesn't just hand ULA a ton of lucrative national security contracts
- backup: BFR is a big project with much more complex engines. If it goes south and turns out to be more than they handle, at least now they can always return to the FH and develop it further
while those reasons were enough (barely) for him to keep it alive, they are nowhere near enough to justify the project as a whole. If he could go back in time, I suspect all the FH dev work would have been gone directly to the BFR
He considers it a test platform for technologies that the BFR will also need. Mainly dealing with double-digit numbers of engines, I think.
the main challenge of FH was how to tie the three boosters together and ensure their vibration doesn't interfere with each other and how to make sure they cleanly separate. Challenges that are largely irrelevant to BFR.
He likened the problem of FH to getting 3 missiles to fly in close formation to one another, because the struts can't actually do that much to hold them together
he also commented on how flexible the boosters are
so trying to get these big jiggly rockets to stay close, but not too close, is a nightmare