It's a good thread. Nostr has a _long_ way to go before it's safe to use as an identity. Or even, arguably, safe to use at all.

Vaxxed by a company the CDC didn't approve of, with an approved vaccine, or a COVID vaccine they didn't approve of?

Sunburns inside your nostrils are the worst.

Signet is testnet without work. All work and no play makes signet a dull boy. Signet has no work at all. Signet is all play. A play at a trusted centralize system. A system with no work. No work at all. Just sigs. A network of sigs.

Then testnet will become worthless. Because testnet is only valuable if it's worthless. Making it valuable makes it worthless. Testnet has to be worthless to be valuable. If you make it valuable it becomes worthless. That's why you must not trade testnet for value or the value of testnet will be worthless. Testnet is worthless if it becomes valuable. Value is anthetical to the value of testnet. Testnet is…

Full-RBFFFFFFFFFFFF

Who cares about bullish? It's easy to pay with cash and makes a difference. And other people see you do it too, including total normies who aren't going to be reached with nostr. Pay with cash and stop making excuses.

You should always pay with cash. It's the #1 thing we can do to stop CBDCs.

Worst sunburn I've ever had I got while skiing.

Demurrage however has a reasonable chance of happening.

Tipped you 22 sats, because tail emission. 😁

In Bitcoin, yes. 😂

I'm one of those node operators admittedly. 😂

It'd be good for LN payment software to have the option of not always using the most efficient routes too.

2x is likely to be enough for anonymity in a lot of situations.

Do you know if 2x had a technical reason, or was a safety limit? If I'm not mistaken any limit would unnecessarily cause payment failures in cases where fees had been lowered; the sender thought the higher fee was acceptable so there's no reason to reject it.

Question: does the Lightning protocol allow you to overpay fees? I'd assume so: checking that the fee is identical to the required fee isn't necessary.

Why am I asking? Because the combination of round payment amounts and public routing can be used to deanonymize payments. Eg if Alice sees a payment of 1005 sats, and Alice knows that the only route that would require a fee of 5 sats on 1000 is to Charlie via Bob, Alice can make a pretty good guess that the payment is intended for Charlie.

Since LN fees are so cheap, it'd be perfectly reasonable for wallets to intentionally overpay fees a bit to increase the anonymity set of where payments might be going.

Related: I recently changed my default tipping settings to be not-quite-round numbers.

It's not so simple to just say "LM infra runs on clearnet"

Lightning nodes don't depend on DNS or certificate authorities, and Lightning routing doesn't depend on uncensored end-to-end connectivity. Eg if your ISP censors IP connections to Charlie, you but you have a route to Bob and Bob has a route to Charlie, you can still pay Charlie with Bolt12 payments because on LN there exists a route via Bob.

Historically whole swaths of onions have been effectively blocked due to Tor being DoS attacked. While ~zero LN related stuff has DNS blocked (or blocked by certificate authorities. So Tor loses on that count.

Overall, DNS itself is more decentralized than Tor in terms of number of points of failure (there's a _lot_ more root servers than Tor directory authorities, and also a lot more domain registrars to choose from). OTOH, Tor tends towards "all-or-nothing" failure modes, as it's much harder for the people running Tor to block a specific service.

In practice of course the internet itself is arguably a centralized system due to how IP addresses and routing work. But no-one has really tested that politically so it's quite possible that attempts to censor via that level fail and the internet fragments. We really don't know.

Bolt12 is clearly superior here, when implemented, as it re-uses Lightning infrastructure that you have no choice but to depend on anyway. A lot less moving parts too.

The claim was that LNURL over Tor had the property of not needing "centralized DNS":

#[0]

Replacing centralized DNS with centralized Tor isn't a change in centralization level. That's why presenting it as such is misleading.

Nope. Onion addresses are useless without the Tor consensus, which is centralized.

If the Tor consensus signers go down, Tor itself goes down.