First of all, it is true that the price of Ethereum seems to have roughly followed the price of Bitcoin in the past. However, (as User Stompix pointed out in the mentioned AltcoinsTalks thread) it is not at all impossible that Ethereum will be able to spin it as battle between 'PoS vs. PoW' (and 'Ethereum vs Bitcoin'), sort of like a 'Nokia vs Samsung' or 'Pepsi vs Coca-Cola' situation. Once the public hears that there is a contest between the two technologies, the price of the two might no longer remain correlated.
I think at this point anyone that is interested enough in crypto to be invested (or to consider investing) is already well aware of the PoW vs PoS situation.
Regardless of that it seems impossible to guess how the market would react in practice, but given that history suggests both that (1) Bitcoin crashing takes the rest of the market with it and that (2) markets react surprisingly little to 51% attacks (though those were admittedly rather minor alts) it seems like the odds are not especially in an attacker's favor -- at least in the scenario of Ethereum trying to take Bitcoin's throne by means of a 51% attack.
Your second point is also very valid. But first of all, if some of the Ethereum stakeholders start funding an attack, all they and other stakeholders would have to do is to in order to mitigate a reverse attack is to collectively stake a larger amount of the total Ether. And what's more, if one reads the Ethereum docs, see in particular
https://ethereum.org/en/developers/docs/consensus-mechanisms/pos/#pos-and-security, it appears that the Ethereum stakeholders believe (and maybe rightfully so) that they will easily be able to revert a 51% attack, even if one happens.
Please note that I'm merely comparing the two cost estimates given by the paper, assuming that both cost estimates are equally reasonable and that both attack scenarios are of similar impact. Discussing the validity of these conclusions are a different matter, but would require taking a closer look at both estimates.
6-20M to reverse how many blocks exactly? The way PoW deterrence works is that the cost is disproportionately higher than the result of such attack. Spending such high percentage of total market cap would be viewed as a loss for Ethereum, because it means that after a couple dozens of such attacks Ethereum's value will collapse, will Bitcoin will just suffer from a temporary set back. And the practical damage of 51% attack can be mitigated by requiring more confirmations.
6–20 billion, so hundreds of blocks, really.
According to the paper the range of 6-20 billion would apply to an attack that lasts for one hour. So that's about 6 blocks, not hundreds.