I actually disagree with the perspective. Why would you bother running your own full node when you’re not participating in the consensus and don’t even hold a lot of capital on-chain? I think the fact that anyone with an adequate amount of capital can permissionlessly verify the chain data and even participate in the consensus itself is enough to prevent industrial node operators (wallets, block explorers, node indexers, centralized exchanges, etc) from malacting.
But let’s agree to disagree on this matter.
Then, what if we simply double the blobspace size? If the networking overhead becomes an issue, we may add an artificial cap to the block size (which won’t affect the L1 throughput in most cases since most of the blocks aren’t even using 0.3MB)