I don't know if it can be fully solved. It can be mitigated to some extent, but I don't know that it's a fully solvable problem.
Monero, for example, uses proof of work for the security and dynamic block sizes for the scalability, but eventually that could do some small damage to decentralization, although hopefully it should not be much since storage is decently cheap
Scaling block size is not a long-term scalability solution. Bitcoin Cash forked from Bitcoin to double the block size. Then they doubled it again. They continued this and it's now 16x Bitcoin's block size and there are calls to double it again because "fees are too high". Increased block size=increased resources to run a node. It's why most of Ethereum's nodes are hosted in one of three corporate datacenters. Very dangerous for decentralization.
All of humanity's transactions shouldn't be stored on the ledger, permanently, forever. That is a waste of resources and totally madness. L2s are the solution, they use the main chain for security but store transaction data off-chain. They also inherently increase privacy. Bitcoin has lightning and (soon) Ark. Lightning is secure, mature technology and it works without sacrificing decentralization. In the last two months alone, it has been used by Nostr users (decentralized twitter clone similar to mastodon) to send over three million tips in the last two months. None of those tips were on the Bitcoin ledger because none of them need to be.
Lightning doesn't absolutely require centralization, but it does cause it. Who's lightning nodes have the most liquidity? That would be places like corporations, BlackRock, etc.
It doesn't. You can run a node on a raspberry pi. The only time having a large amount of liquidity matters is for sending large payments, but multi-channel payments are becoming a thing (break payments up until several smaller payments) so even that is not a problem long-term.
And what about the problem of force closing channels and causing the people to pay the fees on chain which can be quite expensive?
Edit: Also, there are no lightning wallets on fdroid that have been updated within the past year, and some of them specifically say that they have anti-features. If it's not open source, I want nothing to do with it.
And what about the problem of force closing channels and causing the people to pay the fees on chain which can be quite expensive?
On-chain fees are like $.50-$1 most of the time. This only matters if you operate a routing node. If you do, it's something you need to take into account and plan your channels wisely. There's no incentive to force close, non-forced-close is cheaper for everybody, so forced close only really happens if the other node just goes awol for whatever reason. People who are using "regular" lightning wallets never have to worry about this since a lightning service provider (LSP) handles everything. The number of LSPs continues to increase over time and wallets are talking about adding the ability to automatically select LSPs based on published pricing. Importantly, LSPs do not custody funds so there is no rug risk there.
The other situation that gets talked about in relation to channel closes is if a malicious party broadcasts an old channel state to chain (and you have to step in and say hey no actually this is the newest most correct channel state). This is an attack that exists in theory but in practice there is anti-incentive to do it. You lose funds trying it and most lightning wallets and all lightning service providers (LSPs) automatically monitor the chain for this so your chance of actually accomplishing this are basically zero. I have never seen this happen in the wild nor have I ever heard of it happening.
Zeus gets a new version like every month and it's open source. There's plenty of FOSS lightning wallets. Electrum is a good desktop one.
It sounds like you know a good deal about the tech and are interested in it. Encourage you to research more on lightning as well as Ark.
In the last two months alone, Nostr users (decentralized twitter clone like Mastodon) sent each other 3 million tips over Bitcoin lightning. It works, it scales, it's been out for 5+ years now and continues to improve. I can send money to anybody on planet earth in under a second for a penny in fees which i can't even do with my bank account. But it's a failure. Lol.
Just saying, it does have legitimate problems. I have, for example, interacted with the polygon chain and have never interacted with the ethereum chain at all personally.