Thanks. This I'm somewhat familiar with, it's a theoretical fundamental issue in any decentralized and distributed routing network so applies for any interesting L2 solution.

In practice this needn't be an issue:

* Most nodes will rebalance, open and closed channels based on neighbour uptime and reliability

* In practice the LN network is expected to have a little but of a hub-and-spokes topology. So in practice you will almost always have a couple of well-known highly connected peers along your route

So it's a fundamental issue that will become a reality in worst-case scenarios but something that can be mitigated.

As one of the comments noted, we're somehow able to do this for TCP and TOR.

You linking to the /btc subreddit makes me think nothing anyone says or any developments will ever convince you (correct me if I'm wrong). In that case, I think your efforts are better spent in focusing on bringing value to the efforts you do believe in rather than telling others they're bound to fail.