At best a cabal of miners could censor you by refusing to include your transactions into blocks, and refusing to mine on top of any block that includes your transaction. But only while they maintain a majority of hashing power.
But that's exactly what we were talking about: hard forks that change how the protocol works. The parent comment that you responded to is correct in that theoretically a majority of miners could fork Ethereum to steal $100M from a single person.
In the event of a contested hard fork both chains would continue to live, because some miners would mine chain A, and some miners would mine chain B. Yes, miners who mine the old chain would not accept blocks from miners who mine the new chain, you are correct about that, but the same statement holds true in reverse: miners who run the new chain would not accept blocks from miners who mine the old chain.
If a majority of miners decided to steal $100M by forking the chain, they absolutely could do that, even if the majority of users would support the old chain. In that event the tokens (which used to be worth $100M) would now exist on both chains, and it would be up to the market to decide how much the tokens are worth on each chain.
Nope, I was very clear that in this context there is more than one network, so expressions like your "by definition the network has rejected invalid blocks" doesn't make sense. Network A rejects blocks by network B, and network B rejects blocks by network A. There is more than one network. You don't need to look further than Ethereum Classic to see that this can happen in practice. The "minority chain" doesn't just disappear magically.
The original implication was you don't want to piss miners off otherwise they can unilaterally decide to take your money away. That is false.
I think it's not impossible to imagine miners and nodes voting to hard-fork a coin because some genocidal villain, Mr LiterallyHitler, had a large amount of wealth wrapped up in it.
If you are retreating your argument to that goal post, then we are in agreement: pissing off miners will not result in miners stealing your money. Your original argument was stronger, however. You were asserting that a minority-supported fork would not be able to steal money, which is not true.
It is a common misconception that miners, because they are the ones who produce new blocks, have power over the network and set the rules. That is not the case. Miners determine the order and inclusion of __valid__ transactions. Deleting your money would require a fork. Even if a majority of miners mined their own fork but didn't get users/nodes on board they would be wasting their time on what is essentially an altcoin that they alone care about. The network would ignore them and your $100M has not been deleted because it still exists on the chain everyone actually cares about.
There are only so many different ways I can state the above idea (going-on 4 times in this thread alone). If you still disagree, then I'm sorry I wasn't able to explain it any better. I'm going to politely decline from responding any further to this thread because I think we are just wasting each others' time.
[1]
"I guess you'd better hope you're never on the wrong side, politically, from a majority of ETH miners."
https://news.ycombinator.com/item?id=28678451
[2]