first I'd like to indicate what you need to successfully accomplish the PoS attack you pointed to.
a) disable checkpoints and have users download your release.
Checkpoints completely centralise control of the chain, though.
They don't if they are hashes of hard fork blocks, coded into the software thousands blocks after the HF had happened by consensus.b) procure many PoS accounts keys with very large stakes.
Historical keys have no value, they should be easy to purchase.
There is no proof to this statement.c) have powerful / quantum hardware to build a myriad of chains to get one with the cumulative difficulty higher than the valid chain to be accepted by other nodes.
Since block production is so cheap in PoS, I don't need any such powerful hardware to do this.
Yes, you do. A well designed PoS has protections against this computing attack to notably slow you down.Now for the theoretical attack vectors on PoW, these are well known but I can't refuse since you asked.
1) voluntary collusion of miners.
Yes, 51% attack is well known.
2) involuntary - high vulnerability of centralized mining physical infrastructure and known locations to state level attacks.
The chain is highly resilient to such takedowns - anyone with a CPU can mine. In PoS, if you confiscate the stake from the stakers, the chain is totally unrecoverable.
a) anyone with a CPU can mine after how many years of difficulty adjustment and dysfunctional network? b) the protection of Bitcoin is ASIC network. If it's degraded to anyone with a CPU can mine, many state level actors can attack the network that had protection by ASICs removed.3) dependence on ever increasing hash rate and difficulty, security is weakened if price goes down, dependence on energy infrastructure.
Security is weakened if
hash rate goes down, not price.
Hash rate follows price. PoW depends on the price going up. If the price goes down, the hash rate goes down, PoW is at risk.4) quantum computing.
Does not exist.
You can't be sure. PoW and PoS have their advantages and flaws. Dismissing one or the other is ill-advised. Hedging your risks using both is wise.