This is FUD.
The counterparty to miners are the nodes, and they are the ones in control of the network. There are about 47000 nodes and anyone can run one (this is why the block size is small) on like $200 worth of hardware.
An entity controlling the majority of mining power can produce the longest chain. But nodes respect the longest valid chain (by their definition of validity), so they can reject a change supported by miners, by considering that chain invalid. That supports your idea that the 47,000 nodes, which anybody can run, control the network.
But if there are competing definitions of validity, what really matters to me is the version of "validity" accepted by the counterparty I wish to send coins to. In today's world that is pretty much exchanges, since not very many businesses accept Bitcoin directly.
If there is a fork of Bitcoin where miners are in contention with node operators, the winner is the side of the fork that Binance+Coinbase+FTX+Kraken decide to treat as "BTC" and the loser is the side they treat as "Bitcoin Kinda-Sorta".
I wish he'd spent some time discussing the fact that bitcoin's 'value' is proportional to the electricity needed for hashing, and that property will favor aggregating control to a few actors at whatever geographic local has the cheapest electricity. The political and environmental shenanigans exploited by these actors is a deep dive, too, but out of scope for his thesis.
[1] https://www.tomshardware.com/news/study-finds-bitcoin-more-c...