On the scalability issues of onboarding millions of LN mobile clients

The discussion centers around the potential for consensus capture in Bitcoin, not just by miners, but by any subset of users whose interests diverge from the overall consensus. The concern is that the costs of security are being externalized from light clients onto full nodes, making it harder to run full nodes and potentially pricing out some node operators. As node counts drop, the set of node operators will increasingly represent economic actors with extreme weight, and their interests may diverge from the population at large, leading to a natural instability in the system. This consensus capture could be carried out by exchanges or HNWI's as easily as miners, so the problem is not limited to a single group. While there are potential solutions such as halting the wallet, fallback connection to a trusted server, or invalidity proofs, it is clear that fostering node adoption is necessary to maintain a wide-enough, sane backbone network to build on top of. Despite the potential benefits of fast, affordable, confidential, censorship-resistant payment services like LN, the security model of Bitcoin still relies first and foremost on running a full node. BIP 157 improves privacy of full-node-less usage, but provides no real benefits to full node users compared to more efficient protocols like Stratum/Electrum. Therefore, all efforts to improve the "full node-less" experience are harmful, and should be actively avoided.