In Bitcoin, as long as you pay the fee, some miner will probably mine your transaction, based on purely economic incentive. Miners that refuse to serve certain customers based on "tainted" coins or illegal activity simply forego that revenue to miners who don't.
I'm going to go out on a limb here and conjecture that Facebook at the 100 entrenched industry players who are part of this new system will probably not consider censorship-resistance as a high-priority goal for the system. They will probably have a law-enforcement login portal that calls up Facebook's data profile of participants in a suspect transaction. There will probably be blacklists and fund freezes. There will probably be State Department liaisons to help comply with sanctions.
Farther afield you have stuff like Chia's proof of empty disk space or of course PoS.
Monero is currently pushing back with randomx which performs well on cpu's and not much else: https://github.com/tevador/RandomX
I wish Libra was building on an Ethereum scaling technology, like Plasma, but most likely it's a new standalone proof-of-authority blockchain where the validators are the hundred companies investing $10M each.
In terms of the sliding scale between centralized and decentralized, Ethereum v1 and Bitcoin are considered to be "pretty much impossible to change history"; Ethereum v2 is expected to be "extremely impossible to change history"; but Libra would require only that the hundred companies agree to a process by which history can be changed: Libra will be very centralized in practice and while it might and I hope help take crypto mainstream, Libra is not a substitute for Ethereum's secure base layer.