3. It forces high fees all the time, because even if blocks will be full, then your self-transfer will always fill the remaining space, and make some fee market.
Doesn't that mean that the attacker has to frequently pay high fees to sustain this attack? Moreover, why does "wasting time on dummy transactions" centralizes the network, and why do we call it "wasted time" since they are regular transactions, just like any other honest nodes would either way verify.
You can avoid paying high fees if you never broadcast those transaction. Although with website such as
https://mempool.space/ it's trivial to detect that.
There should have been a 4-year upgrade in terms of block size incrimination in a similar fashion to how network difficulty rises with time.
Difficulty doesn't always rise, see
https://blockchair.com/bitcoin/charts/difficulty.
What I mean is, that Satoshi was able to identify the supply and demand issue earlier through its halving event. However, they did not consider the scaling of block sizes or maybe they never thought that Bitcoin would go to such a high level anytime soon.
Yeah, Satoshi isn't perfect human being. Satoshi also seems fail to predict about specialization of mining hardware and pools.