Thread by @el33th4xor

People are finally realizing what we have been saying for a long time: the capacity of the emerging LN network is likely going to be quite limited. In real life, most consumer flows are unidirectional, and channels will just get depleted.

https://twitter.com/Bitcoin/status/1099355610515210242

One thing that would help LN is if people started using custodial wallets, i.e. like bank accounts. That would help capacity, it'd kill privacy and financial independence.

The other thing that people are slow to catch on is the fundamental privacy trade-off involving LN routing.

To run the network efficiently, the protocol has to inform users of channel capacities. The more frequently it does this, the easier it is to monitor users.

The more frequently I can monitor your channel capacities, the better I can watch in real time as you receive payments. This is the antithesis of privacy.

And of course, there's the reliability and user experience issue that no amount of UPS's, extra disks, etc cannot fix. When we proposed Teechan two+ years ago, Core sicked its trolls on us, complaining about our prototype implementation back then.

We didn't develop Teechan into a commercial product because we don't believe there's a strong enough market case for L2, and also because we are working on more exciting projects than L2. But Teechan then was miles ahead of LN now, even the quick prototype was crash tolerant.

A few channel failures, and having to wait for channel timeouts, will sour the user experience to the point of wanting to open channels only with entities whose identities are well known and are technically competent. That's a network of big companies.

So the LN game is one of stringing the muggles along with empty promises while sweeping the big issues under the rug. I guess if your consensus layer is inherently limited, you don't know how to fix it, and you see fees as a positive thing, this is the best hand one can play.