segwit makes individual transactions bigger.
It does not. (unless you want to count a byte or two from flagging in particular serialization which could simply be compressed out, if anyone cares).You should read the FAQ: https://bitcoincore.org/en/2016/01/26/segwit-benefits/
Key advantages include solving malleability, facilitating safe script upgrades, lowering resource usage for lite clients, allowing additional capacity without significant additional UTXO bloat risk (in fact, it largely corrects the issue that creating additional UTXO is significantly cheaper than cleaning them up). In doing so it significantly reduces some of the risk points in increased capacity.
Quote
saying you don't need the segwit data is the equivalent of saying SPV mining is OK
The network consists of more than just miners, segwit doesn't change the data miners need. Allowing a continuum of options in using the network is important for avoiding some users being 'priced out' by the operating cost of it.Quote
As for malleability, so anyone wanna say why that can't be done properly on it's own without segwit?
Segwit is the only proper solution for malleability. The only other alternative is long lists of canonical encoding rules-- a mess of bandaids which, at best, only solve it for a subset of transaction types-- and even then maybe not (it's very hard to be sure that a set of canonical encoding rules is actually sufficient to guarantee non-malleability). That kind of patchy approach is a huge pile of technical debt that would make it much harder to implement and maintain the consensus code in Bitcoin implementations.
