The lack of successful 67% attacks on major PoS chains is indeed a positive sign. However, this could also reflect social deterrence, early-stage economics, or insufficient adversarial incentives — not necessarily intrinsic robustness.

• Liveliness attacks (e.g., preventing finality without conflicting it) can happen with 1/3 of the stake. This happened on Ethereum in 2023.
• Censorship and delay attacks are feasible with much less than 67% if coordination and MEV strategies are used.

Anyway my initial argument was that Bitcoin have no interest to move away from its PoW and especially not to transition to PoS. PoW isn't perfect and has its drawbacks but PoS is not a better system, it's an alternative with other drawbacks that doesn't meet every chains needs.