If you haven't seen this before it's worth a watch.
Andreas explains why Big Blocks are not going to work. (watch time: 5-6mins, Petabyte Blocks & Streaming Money)
https://youtu.be/AecPrwqjbGw?t=11m39sSegwit in its current form can't handle hundreds of millions of users either. That was never what it was about. We needed a step up, a year earlier than we got it, and we had a few options to do so. For the current leg, and I repeat myself yet again, why was segwit better than simply doubling the blocksize? Nobody seems willing to explain that bit, for whatever reason.
Because Segwit is an upgrade that enables many new technolgies, L2 & L3. Technologies that will enable millions of TX's per second.
We don't need an incremental step in the wrong direction, as perfectly illustrated by Andreas, big blocks can never fullfill the Bitcoin dream, because PetaByte blocks are effectively impossible.
I guess if you think doing the wrong thing as a temporary fix, and incurring the associated Technical debt is OK. There is little I can do to defend segwit, but in my mind and the minds of many other developers, it's not acceptable, so Big Blocks are not an option at all.
If bigger blocks are a bad idea then why are you not advocating for a blocksize decrease? Why do we have Precisely the Correct Blocksize as things now stand? Also I'll need to see some convincing numbers for the millions claim.
Most people who see LN and other Layer 2+ solutions as being the way forward for scaling also realize bigger blocks are an eventual necessity. But I would want to be conservative. Since shit expands to fill the space available we should see how we can do with efficiency BEFORE we add resources.
People keep saying that and I call bullshit on it. Why would people start making more transactions just because the network has more capacity? It makes no sense at all.
Hash Time Locked Contract (HTLC) payment networks including Lightning Networks and the Layer 2 and 3 technologies have three major problems:
1. They force centralization onto hubs. Those who think this is not true, I can defeat you in a debate even though I will not expend my time on that debate now. Because of this centralization, they destroy the security of both the off-chain transactions, and they weaken the security of the on-chain Nash equilibrium because they create large hubs that hold economic control of the network. Remember as the block reward decreases, Nakamoto proof-of-work looses incentives compatibility to converge on a single longest chain and will fork off into innumerable forks unless theres a stable oligarchy cooperating. See the Decentralization section of
my blog post. This off-chain crap is the antithesis of the utmost priority on the security of the game theory of the chain which is what gives Bitcoin its status as the reserve currency of crypto (and eventually of the world). So there is no way that shit is going to be allowed on Bitcoin. That stuff will end up pushes off on a Core fork or other fork off from Satoshis protocol where the $billionaire economic majority will remain steadfast.
2. None of that shit scales because it drives up the need for larger blocks. Yet larger blocks are not a scaling solution (not the in correct mathematical) definition of scalability.
3. The transaction fees will become too high for users to open a payment on channel on-chain (as they will be competing against the hubs who dominate the available MBs per block), thus LN is going to be a fractional reserve system where users open accounts at hubs as they do with exchanges today.
I explained some of this is more detail at the following linked post (including my critique of layer 2 proposals):
https://bitcointalk.org/index.php?topic=4435201.0#msg39666724I figured out a long time ago that on-chain transaction volume scalability is going to come from an altcoin. And that is what I have been working on. If I successfully mitigate the nothing-at-stake, censorship, liveness, and economic capture vulnerabilities of proof-of-stake, then what I will have is far superior to that off-chain crap. Bitcoin will remain the reserve currency. So I hope you understand I need to STFU and work.
I dont have enough time to get involved in another long argument. I
might not reply even if someone thinks Im incorrect and even if I possess a correct rebuttal. So take this FWIW to you. Im dont claim to be omniscient, yet Im reasonably well studied/researched on technological, game theory, and political-economics issues of consensus ledgers. I roughly guess that I have more specific knowledge on this than Andreas, but I havent spoken with him nor watched all of his videos.
Well, I will try my luck to see a reply.
About your points:
1- Yes, some L2, like LN might be more or less centralised. Let's assume a hub and spoke topology for LN.... What's the problem? It's just one of the second layers of BTC. That doesn't make Bitcoin more or less centralised. The alternative, in case of mass adoption, would be corporate (coinbase, etc) online user wallets with no blockchain impact. I do prefer LN which, at least, settle channels on Blockchain and reduce the risk of "evil" hubs incurring in fractional reserve banking practices. In that, LN becomes a more DECENTRALISED alternative to online wallets. It does not substitute blockchain tx's but online wallets.
If you want to (really) scale, you either choose between L2 settled to blockchain or "L2" online wallets with no settling at all. My choice is clear in this case.
2- Segwit+L2 is a scaling solution as it makes it possible to handle several times more tx's for the same blocksize. Segwit does so by moving signatures to a secondary place and L2 by offloading main chain. And both do work as capacity multipliers for each linear blocksize increase. That's the very meaning of scalability!
Of course one of the capacity factors is the underlaying block size, which will need to be increased as needed if adoption growths in the future.
Multipliers/scaling factors (L2, Segwit, etc) are needed, but so is blocksize increase if we want to reach capacities several orders of magnitude higher than currently.
I envision a future in which 80% or more of tx's will be on L2 (small stuff mainly... cofee anyone?) and THAT will leave more room for other (bigger) tx's to be directly on blockchain. That is scaling.... and I don't care at all if my small change stuff is more or less decentralised. Not at all as soon as my main reserves are safely stored ON-CHAIN.
3- Again, somewhat correct, BUT you seem to forget that as more tx's get routed through other L2 channels/sidechains there is more capacity available on main chain. I don't see how you see LN to be or allow a fractional reserve system... All LN channels need to be PRE-FUNDED by real BTC on the main chain. I really don't follow your thinking here. The alternative (online wallets on main providers) IS a fractional reserve risk and that is what we want to avoid with L2 solutions like LN.