There are implications. For instance if foundry shit down unexpectedly the hash rate would drop significantly. It wouldn’t kill Bitcoin obviously, but it would be something people fud about I’m sure there are more threatening considerations too

It is a concern because it allows that one pool to alter transactions in the blockchain, while nodes create the rules avoiding a 51% attacks also helps keeps bitcoin decentralized.

why nodes are not truly in charge of the network your node is economically insignificant and your hashrate is a rounding error. it doesn't do anything. the toy SHA256 ASIC you purchased is a scam. if nodes like yours actually had any influence over what the rules are, bitcoin would be vulnerable to sybil attacks, and it isn't. "the block wars" refers to UASF, which was a stupid idea sold to the public as a way for users to have more control over the network. the narrative was backpeadled to "economically significant nodes" i.e. nodes operated by financial institutions. exchanges were supposed to reject blocks which didn't follow their desired set of consensus rules. if UASF really works at all, it enables exchanges to dictate the consensus simply because they handle most of the coins. fortunately UASF doesn't work. it was an empty threat and it literally never happened. segwit was activated the normal way after reaching 95% miner signaling. no soft forks have ever been activated by the UASF mechanism. UASF was pretend proof of stake larping as a pretend sybil attack. if you really like a network where miners aren't in control, I have some alts to sell you. View quoted note →

You're right that pool centralization doesn't make the deciding difference when it comes to defending against changes to the consensus rules via hard forks. Nodes are the deciding factor there, and specifically nodes that are actually being used by those who run them, not ones that are just sitting on a shelf gathering dust while the owner broadcasts transactions through whatever the default node is in Sparrow wallet, and checks to see where the transaction is in the mempools using mempool.space. However, pool centralization is a MASSIVE deal when it comes to what transactions actually make it into blocks. Foundry currently decides what transactions make it into 41% of blocks. Antpool and its slave-pools decide what make it into another 50% or so. That means effectively two entities control what transactions are included in roughly 91% of blocks. That's problematic, to say the least. Miners should move to pools that allow them to create their own block templates, instead of relying on the pool to dictate what is included in the block template. Currently, the only pool that allows this is @OCEAN that I am aware of, but more pools should adopt DATUM and do the same. It would not be NEARLY such a big deal that Foundry has 41% of the hash rate, if the individual miners on that pool could create their own block templates based on the transactions in their node's mempool.