No doubt the fees and privacy is a downside.
I don't necessarily have an issue with the fees. They are high by bitcoin standards, but low when compared to using a credit card. My concern is more that the wallet could charge fees higher than what is advertised via intermediary LN nodes.
But I guess these are the sacrifices for having an ultra mega nooby friendly service, as so the user doesn't need to care about nodes and servers and such, just easily make LN transactions almost like doing on-chain transactions. What do you think? Are there possible ways of handling things without sacrificing ease-of-use?
I do see LN wallets in the future that are user friendly. I would predict we will see user-friendly desktop LN wallets before mobile wallets because coding mobile software is somewhat more complex than desktop software, and mobile OS will restrict what apps can do more than what macOS and Windows.
Over the long term, I would predict there are a handful of fairly large LN node operators that most end users will connect to, and will compete with, and connect to each other. The topology would somewhat resemble that of airport routes, but with many more routes.
With the above being said, I would envision future wallets LN wallets to connect to nodes with some kind of standardized URI that is unique to LN. I think most users will need to connect to one of the major hubs, and those that connect to multiple hubs will force each hub to compete on price. Something similar to what Phoenix does, except the user can choose which nodes to connect to.
I think LN wallets will use seeds, similar to what Phoenix uses, but will also automatically backup channel states to cloud storage during and after each transaction. This is important because without this, if you lose access to your wallet file, you will have no way to protect yourself against your counterparty from closing a channel using an old closing transaction.
Unfortunately, none of the above involves the creator of the software earning any kind of income.