I've always found your posts insightful and I know that you are a very intelligent person; I'm interested in your rationale for keeping the blocksize static.  

The argument that "if the blocksize doesn't really matter, then why was a limit added?" seems valid to me.  A smaller blocksize also limits growth of the UTXO database, which is actually the more critical resource.  I'd like to see a healthy fee market develop along with a push towards off-chain transactions (perhaps using m-of-n Oracle sidechains).  A smaller blocksize would incentivize that.  That being said, Solex recently presented a helpful graphic that showed the increase in internet bandwidth since bitcoin's inception, indicating that 3MB blocks now would represent the same % of typically-available BW as 1 MB blocks did in 2009.  So the argument can be made that "we can safely increase the blocksize in parallel with internet speeds."  The Metcalfe value chart also shows a strong correlation between TXs per day and market cap, so an empirically-based argument can be made that limiting bitcoin's TX bandwidth would also limit its value.  

One point you made that I think is not valid is the notion that a 1 MB blocksize is a core part of the coin's definition (i.e., part of the "Satoshi Social Contract" in the same way the 21 M coin limit is).  There's plenty of evidence that indicates the 1 MB limit was never intended to be permanent.  Of course, this doesn't mean that it's necessarily a good idea to increase it without careful thought and debate either.