1. Purists believed the blockchain was good enough for all transactions and the proverbial "cup of coffee" purchases could be done on-chain simply through micro bitcoin transactions. It's clear now that if we had all of VISA's transactions worldwide, in its current form the blockchain would need to have 2GB sized blocks to sustain that volume of transactions. Current technology and even anything we can perceive in the immediate future makes that impossible with transmission of new blocks generated, propagated and stored in a distributed manner. The counterargument is we'll only slowly get up to that size; it won't happen overnight and we'll find a way between now and then to do it.
2. Transactions being done on LN will remove fees from the blockchain by doing all transactions off-chain. LN transactions are actually limited to micro bitcoin values only and the fees all go to the network based on transaction value rather than size (unlike the blockchain) though opening and closing a LN channel requires an actual blockchain transaction and suitable fee. This one particularly had the miners riled up though segwit2x means they no longer believe in it.
3. LN is centralised instead of distributed and blockstream will will run it. Lightning nodes can be run by "someone" but that doesn't mean it will all be held by one entity. It is possible to run a node of your own though it is highly likely that several competing services will come out using the same protocol and people will use that in preference to running a complete full blockchain node and lightning node on top.
Where segwit comes in and where the objection to segwit comes from is that segwit makes LN easier and better featured so most LN development has been done with segwit existing in mind. The people who object to the above were against segwit based on the perception that segwit makes LN possible.
I expect the large exchanges will probably end up providing LN nodes of their own and compete on fees.
Thank you for that useful info. This clears up a few things!
In light of that I still think something like it - if not LN itself - is an inevitable requirement if we ever want BTC to compete with the FIAT payment systems. If the BTC community wants for BTC to remain just a sort of digital gold store of value and large value transfer (once in a while) then doing everything on the blockchain will still work.
I fully agree with you. I too expect this tech to be used primarily by exchanges, large traders and perhaps those btc gambling sites and whatnot.