i have worked on both TLS and blockchain tech for years.
my point is that the purpose of the security architecture proposed in the article is to authenticate data from the chain. this is a protocol-level detail: like TLS secures transport between a client and a server, light clients authenticate data between a validator set and a user. whether a particular person or group decides to use such a protocol, or if they decide to censor data, is a separate concern.
It isn't trying to force all servers to return an identical response like Web3 is.