I think, his only mistake was underestimating how fast we would reach the limit, he expected improvement in internet speeds to be faster than adoption.
If Nash says “ sound money cannot be engineered” i take that as a point for removing the limit and letting the market find its own equilibrium. if a miner tried to mine a block 1GB his likelihood of being ophened due to slow propagation rate would be 100% because it takes to long to move that much data across the network. we do not need to engineer a limit, a “free market limit” will emerge because the of technological constraint.
Satoshi expected that technological constraints would never be a problem becuase improves would out pace adoption. in my view he has not been proven wrong, 1MB is not indicative of today's technological constraints, the most conservative research has shows that 4MB would be starting to hit such a limit. by the time the network grew 4X technology will have improved.