Just listened to another brilliant episode of LTB with Andreas and Adam. It went down into the weeds from a technical perspective but the explanations were thorough and clear.
Not wanting to rehash (couldn't help the pun) anything but my take-always were the following (and I'm sure there are plenty of places that will explain it better than I however this is just my simple plain English attempt at one listen):
1. Bundle up work that makes sense.
Segwit isn't just about scaling nor is it just about transaction malleability. It's a bundle of work that addresses issues with the architecture of the transaction itself (ie not the wrapper/header/block). It just wouldn't make sense to only do 1 task/item.
An analogy may be installing appropriate cabling throughout your house during a refurb rather than waiting till the new paint job and finish is complete and then doing it.
2. Soft Fork now seems to be less disruptive compared to a hardfork
Fixing the transaction architecture via a soft fork in the short term seems to give the network time to adjust and allow for segwit solutions to be implemented over time. This won't eliminate the need for a hard fork but might mean a movement towards greater acceptance/support of a hard fork in the future.
3. Market Forces beats Centralised Authorities for Decision making any day
Loved this point and the way it was conveyed through the example where one persons "junk data is another's viable application". Essentially there were concerns posted that alluded to tipping the scales in favour of diminishing costs associated with "complicated" transactions (ie P2SH? - more than just 1 input/two output). This was said to potentially result in high levels of spam hitting the network.
The counterpoint was clear - economic incentives delivered through network value by market forces. A user either decides their transaction is worth the cost of transmission or not - and in the process market forces will set the fees and value to all players.
The alternative would be the creation of some authority who'd have the power to make such decisions on behalf of the market - which is a feature killer of decentralised trust less systems.
4. Permissionless Innovation is Amazing
Bitcoin allows anyone with any idea for an application to - if they have a user base of at least 1 - realise it. No permission, no tendors, no request. Can't quite remember but I believe one features of segwit will be to include a version numbering for scripts which is supposed to help reduce cases for hard forks as well as facilitate other areas of efficiencies/innovation (eg side chains, smart contracts).
5. Andreas is a champ
Had to call this out and I really valued his epilogue. Andreas presents his opinions in a rational and methodical manner. And truely personal opinions they are with no monetary "cash for comments" by any entity on each side of the debate. (Frankly I think it's absurd how anyone could suggest otherwise).
Anyway, that's all off the top of my head for the LTB#337 review - I hope it helps in undertanding at least part of the story and debate that's going on, or at least can point people to other information sources.
Cheers,
Haso.