There are good points here, but some only in proper context. Where John's product is node-on-phone, it suffers from many of the same (or worse) problems by nature.

Unless you're using a custom ROM your phone, it's just a Google or Apple thin-client. Putting a node on a phone at all is a larp if you also claim to be concerned with censorship, security, control, social attacks, and performance.

Such concerns can only be mitigated on mobile if the mobile has limited scope access to the node. A thin-client made for a thin-client is the only sane architectural approach in my view.

Cost is another yet factor no one wants to mention with these unsustainable mobile nodes. As users look for alternatives to WoS in a high-fee environment, they are only now beginning to understand this.

It's ultimately the failure of node-on-phone architecture has pushed people to WoS, whether its delivered via PWA or Native is irrelevant.

Why are so many wallet teams doing this then? Frankly, it's a corner-cutting way for companies to sell LSP services and skirt perception of being a custodian.

In most cases these mobile nodes are still distributed in a completely trusted way, and can be rugged by the developer just as easily and with an even better shield of deniability.

On the other hand, bootstrapping your own WoS for yourself, apps, friends, family, etc, with architectures like we're modeling with ShockWallet, is the future. A holistic approach to mitigating as the above factors and unlocking usecases like cross-device wallets and so on is how we break out of this stagnation in wallet development.