Author rightly points out that layer2 is constrained in part by layer1. We can't open enough lightning channels for every person on the planet to have one within a reasonable time, let alone 2 or more for censorship resistance, routing, etc.

I'm personally ok with trying to optimize base layer to support more tx, but it's still going to be fee market driven. I use layer1 and layer2 solutions. Since nostr, my use of layer2 with custodial wallets has eclipsed all my past transactions and yet, we are still testing the tech as trailblazers, and very early.

I really think that most people will be custodial, and they'll have the flexibility to move funds with greater ease than traditional banking as a tradeoff.

Yes. People will lose funds over time. I've lost access to sats to a custodian provider. But it's important not to lose an amount that would be life changing detrimental. That's where we need to build more protective solutions.

For layer1 though, there's little that can be done. New address types? Taproot and segwit are already quite compact. More frequent blocks? Complete non starter as it affects supply issuance. Larger blocks? Maybe in time but still too early to consider IMHO. Until muun swaps and inscriptions, blocks weren't regularly getting full and we have to consider network costs and ability to maintain decentralization. It's certainly easier to try things in a layer 2 way. Lightning, Liquid, Cashu, and Fedimints are all layer 2.