To the user, eventual UX is similar. Direct CEX on/off ramps to rollup, trading, DeFi, NFTs, smart contracts, etc. It is all "on-chain" by virtue of it being cryptographically verified on mainnet Ethereum (and, likely in less time than the finality of a single BTC transfer).
Payment and State Channels are the only true way to scale. Off-chain txs. Lightning network works quite well today for Bitcoin and I look forward to seeing entirely new paradigms evolve in time for Eth.
The Eth core devs and users are more interested in rollups due to them being general-purpose[2], i.e. suitable for applications and smart contracts to build on top of (for e.g. Uniswap on a rollup via zkEVM[3]).
[1] https://ethereum.org/en/developers/docs/scaling/
Not everything needs to be recorded for all eternity.
I do not anticipate a resolution to the scaling problem any time soon [2]. But maybe they, and me, are not seeing the flaw in their argument. Should the scaling problem remain intractable, I think that the existing major blockchains have peaked. Both Ethereum [3] and Bitcoin [4] have been pinned near their maximum transactions per second for years now. Further adoption can only come by displacing some other transaction on-chain.
(I'm operating on the assumption that off-chain non-blockchain handling of blockchain data is not actually blockchain for this comment. No keys, no ownership as the BTC maximalists like to say.)
[1] https://en.wikipedia.org/wiki/Bitcoin_scalability_problem
[2] https://datafinnovation.medium.com/sharding-is-also-np-compl...
[3] https://etherscan.io/chart/tx
[4] https://www.blockchain.com/charts/n-transactions-per-block