I agree with the general sentiment that this article doesn't make its case well - but there certainly is a valid case about the risk of any single instance getting too big.
I think that as long as Mastodon GmbH are still the good guys, a more effective strategy than trying to get individual users to change their behaviour, might be to get Mastodon GmbH themselves to see and agree that mastodon.social being this big in not in the interest of the overall long-term health of the network - and to build in automatic size limiting, where any instance with more than x% of users will close signups until they drop well below that threshold.