Segwit does not mess around with either in a dangerous sense which is what you're referring to.
I believe Bitcoin has a fundamental flaw, and that flaw is never addressed, so all these implementations are just a workaround that flaw.
The flaw is that the nodes are not incentivized, and all money goes to a centralized miner cartel.
Now because of this the number of nodes are shrinking, and you need to artificially keep the computational resources low (including blocksize) in order to satisfy current nodes. It's just horrible.
And because the block capacity is small, new users will hardly join bitcoin, and slowly bitcoin will fade away, due to more competitive altcoins.
This is a very accurate point, which can be fixed , if personal greed does not prevent it.
If a % of each BTC reward or a % of the BTC transactions were shared with anyone operating a Full Node, that would be incentive enough to keep active full nodes.
Something similar to Dash MasterNode's system would work.
But Core has proven , they are looking out for their personal interests and not the general interests of the BTC community.
FYI:
Since Proof of Stake was mention trying to insult me, here is how using Theymos idea of adding proof of stake could help the node problem.
If PoS was added to BTC, any full node could generate a PoS block and Receive the Transaction Fees for including them in the block.
At the current environment , that would be ~$1000 of fees for every block staked.
Now that is Incentive to run a Full Node!