To answer your points:
Point 1: I think you’ve misread the context of that part. They have 33% safety in terms of being able to reach consensus, yes — but if you want to subvert them and force your own decisions to be validated, you need 51%. The 33% case is partially relevant, so I might include it, but there’s no incentive to gain 34% mining power just to stop the system working when that doesn’t benefit the 34% miner.
Point 2: Strictly speaking, you are correct. I explained it this way for simplicity, as most blockchains use the same simple consensus algorithm, and it just added an extra layer of complexity to explain that PoW, PoS etc are ways of determining how many votes each person has in a given blockchain’s consensus system. Still, perhaps it can be written differently to remain simple while explaining that; I’ll give that some thought, thanks for the input.
Point 3: Define “accountable” — it’s not clear in this context what you mean by this.
Point 4: No, it’s not. Markets make it very easy to do this; it’s just not something you do quickly in one big movement. If you want an example, look at rich vs poor distribution, company market share over time, etc. You slowly accumulate more and more of the share until you have a controlling stake without anyone noticing (as it won’t all be on one account). In the real world it’s a slow process, but for young POS currencies it wouldn’t take that long (and you could easily influence gullible crypto users to vote the way you want, or install your own special software to vote a particular way).
Point 5: It’s entirely plausible that they can scale it, but not without sacrificing security. Execution without data availability implies vulnerability by definition, for example.
Point 6: It doesn’t matter. The only way that matters is if renewable energy is unlimited (which it isn’t) and if miners are actually using it, both of which are not currently true. Until that’s true, no amount of eco-friendly algorithms will change anything.
Point 7: This relies on the false assumption that you can determine who the stupid actor is. I added a section recently about this, which you can see in the “Recovering from blockchain takeovers” part.
As for MEV, I have not looked into that; quite honestly there are so many weaknesses in blockchain systems that it barely seems worth mentioning, but I’ll research it anyway for my own curiosity.
(I would note that Avalanche, which you seem to be interested in, is even less secure than any of the above with the upside of being less consistent to break. It’s harder to subvert in a consistent manner, but it can be subverted by luck without penalty.)
