Also, i have never ever seen a scientific, measurable, technical, economical, psychological research/prognosis with hard data and testing about what a limit increase would ACTUALLY do to decentralization - i bet all the miner pools and their thousands of contributors would still run a node, the node hardware can not possibly be more expensive than 0,01% of the mining costs...
As a scientist i am basically agnostic in the matter, but i see no real life data showing doom is coming with greater (much greater than 2MB in the long run) blocks.
But, i see - and everyone with common sense - the math, that not increasing in several magnitudes in blocksize in the future is mathematically//physically hindering the growing of the network before we hit that sweetspot of "global network effect".