Thank you @franky1, I'll read it. I don't consider LN an ideal solution for Bitcoin to scale, too. I think it's over-complicated, but for micropayments it may work (I consider it appropiate for payments where you, using fiat, would use a prepaid/gift card).
Franky is talking utter drivel. Whether LN is going to "work" is debatable, but the usual talking points against it (centralisation?! fees?!) are laughably far from reality.
What I was interested in, however, was an explanation in layman's terms if the Segwit concept itself (not LN and also not the "soft fork" itself as you point out in your other posting) had some negative impact on security or usability of Bitcoin. I know the Bitcoin Core page about
Costs and Risks, but it doesn't seem like this list is really a list of "disadvantages" as most of it could be applied to all larger "jumps" in software development. If you or another participant of this discussion have a link for me, I'd be grateful if you share it.
The segwit *concept* itself, as you put it, has no drawbacks because it is a fix of a fairly serious design flaw in Bitcoin. Any particular implementation will always have risks and might have drawbacks; the soft-fork implementation chosen has tremendous benefits and some fairly minor risks, heavily ameliorated by a year of testing. There's no reason to assume it must have a reasonable share of benefits and drawbacks; you'd expect people to choose an implementation with a lot more of the former than the latter!
Everyone on here saying "well segwit is a bit crap but i guess i'll vote for it as the best for now" is for sure talking from a perspective of ignorance, and having listened to equally ill informed people. It's actually a brilliant piece of work and a huge step forward, and immediately gives us ~2MB blocks (whether that's a good idea is up for debate, but if that's what you want, you get it).