Why raise blocksize if blocks aren't utilizing the recent blocksize increase of SegWit? Why raise blocksize if companies still don't batch their transactions? Why raise blocksize if a lot of software still uses horrible fee estimation algorithms? We have enough tools to keep fees reasonably low if only they were used more. But a lot of transaction traffic comes from centralized services like exchanges, and they are too incompetent or malicious to implement those technologies.
I know raising maximum block size/weight is risking decentralization, but i think it won't be hurt to increase it not too much as long as there's data/consensus which says most nodes still can run even after raising maximum block size/weight?
Some studies suggest that 8 MB would significantly decrease the network node count. We haven't even seen how SegWit's blocksize increase would affect it, so it would be very reckless to do any blocksize increases soon.
Sure, it's bad that fees are high, but high fees are not going to kill Bitcoin, because they are driven by open market, they are not installed by one side only like it happens in some centralized systems.
But centralization can kill Bitcoin, you can't undo the damage of big blocks if the blockchain became bloated past certain point.