In contrast, mining can drastically improve the return on capital by operating at larger and larger scales so as to get better and better deals on equipment and electricity. That kind scaling doesn't happen with proof of stake.
Yes and No. We saw what happened to very large mining farms that got wiped out in a matter of weeks as mining was cracked down on in china. Large ETH stakers wont be forced to liquidate their stake in the event they get kicked out of their current jurisdiction. In short, uncertainty about the future of crypto regulation makes it easier to be a staker than it does a miner.
Yes, this ensures a limited supply with PoW, right?
> If you try to remove real-world costs from that equation, you lose the counterbalancing effect.
I assume you can't remove real-world costs in the case of PoS, but yes I can see how the real-world costs will be much, much lower.
Are you saying that there is a good reason to keep those real-world costs higher? (pardon my ignorance, I haven't looked into the principle you describe here before).
There's an argument to be made that there is a centralizing force in the role of the consensus/security layer for the chain, because the asset being earned (ETH) can be staked and earn further returns - However, this role and phenomenon is mirrored in the PoW world. The difference is that the centralization happens one step removed from the on-chain asset - PoW miners consolidate profit into further mining investments, such that an increasing amount of the hash rate is owned by the largest miners, who can acquire improved access to electricity and/or equipment relative to smaller miners. One could argue that a PoS system actually has less room for exploitation, however, since you can't restrict a solo-stakers access to ETH, while you could restrict (or provide severe barriers to entry) on the competition of electricity/equipment.
The concern of the "top staking provider eventually holding the vast majority of Ether" is highly unlikely. Given that the asset will soon have ETH issuance cut by 90+% as part of the merge, and the fact that there is no mechanism by which top staking providers are incentivized more than the small solo-stakers, this would soon enter the realms of the purely theoretical, and seemingly take lifetimes to happen if one could even envision it happening at all without a reallocation into other investments.
MEV is being democratized as well (whether a good thing or not, will let you be the judge), with even solo-stakers being able to use an MEV client alongside validation clients in order to benefit from additional income on block proposals (see Flashbots mev-boost client, releasing alongside the merge).
Would love to better understand whether that better informs your perspectives on the subject, or what other concerns you still have.