Several issues.

The claim that PoW is "inefficient" is economically naive, at least in the form presented here.  The exact cost per transaction is difficult to determine at the moment because of price fluctuations, but ultimately the efficiency of the network will settle to a value determined by the total value transferred over total cost of running the network. The cheeky claim that it "costs hundreds of time more than a credit card transaction" can hardly be taken serious if it is presented without any kind of calculation, as was done here.

Re: network security. PoW at the scale of the current Bitcoin network is de facto secure (no attack ever succeeded, no realistic attack vector that does not involve massive clandestine hardware production is known), and it is getting more secure each day as additional hashpower comes online. "Increased" security through PoS is an advantage that holds little weight then.

On the other hand, there are numerous disadvantages of PoS. The most significant ones I can think of:

Coin distribution by stake, instead of coin distribution by work. There is something intuitively unappealing about handing the highest rewards to the already largest holders of an asset, and there is something intuitively satisfying about rewarding those who "work the hardest". Don't take my word for it, ask around. This is probably the most basal reason why a majority of the Bitcoin community dislikes PoS.

Consensus forming in the case of competing chains. From a technical perspective, probably the strongest argument. Summarized in the well know form as "The trouble with Proof-of-stake is that there is nothing at stake" if there is a need to form a consensus which of two or more competing chains is the canonical one. I'm not informed enough on the current state of the discussione to know if there have been, generally accepted, solutions to the problem, but the entire problem is alien to PoW because hashpower/energy is a limited resource.