When historians look back they might well conclude that the 1MB limit resolution was Bitcoin's "betamax moment", the most important decision point deciding its ultimate fate.
coinmarketcap.com is interesting because it gives a market assessment of Bitcoin's future prospects.
Right now the total cryptocurrency market cap is $8,197,074,397 and Bitcoin's is $7,678,842,953 or 93.68%
(higher if MSC and XCP and other BTC-based alts are included).
So, the cryptocurrency market is effecting pricing a >93% probability of Bitcoin dominating this new asset class - long term. An asset class of such importance that many crypto investors consider could make all fiat obsolete and perhaps even demonetize precious metals as well.
If the 1MB limit is not handled properly, in good time, and is allowed to seriously disrupt organic growth, then that percentage will fall significantly. Perhaps even causing an irrecoverable erosion in the first-mover status, opening the window for alts like NXT and XMR.
This is the stakes in play.
Bitcoin can not disrupt fiat currencies unless it's a network that everyone on the planet is actually capable of transacting on.
Pulling out all the FUD/propaganda stops to keep the transaction throughput rate capped is the best possible strategy the defenders of he legacy system could employ to save themselves.
If they can force most transactions off chain, into the waiting arms of Coinbase and Circle, then most of the benefits of Bitcoin will remain inaccessible to most of the world's population.
Given the stakes involved, I don't trust Gavin to be capable of making good decisions even if he wanted to. A single point of control is a single target for blackmail and extortion.
they are very aware of issues like block size limits and bandwidth limitations.
What you're talking about are symptoms of a sub-optimal network layer.
The fact that the P2P network suffers from a tragedy of the commons effect because there is no price discovery for bandwidth or storage is nothing inherent to Bitcoin - it's inherent to the current design of the reference client.
If people who don't want Bitcoin to succeed can stop that problem from ever being fixed in the reference client, they can point to the tragedy of the commons problem as an example of why never to raise the transaction rate, forever.
Of course, that strategy only works when a single codebase has a monolopy on the implementation of the protocol.