I generally dislike stablecoins today, due to centralization and trust issues. See #945661

stablechannels don't have those problems, but you have the free options problem, and worse: people really want a stablechannel to avoid massive (80%) drawdowns, but that's exactly when you'd expect them to fail. They risk becoming a complete scam, where the stablechannel peer keeps the upside and walks away on the downside. And you'd never know as long as the price is rising...

I'm emacs, but I'm not religious about it; it's just that I've used it for long enough to adapt it (or me) to my habits. I use lots of shell, and I use "guilt" on top of git, and multiple helper shell scripts.

I'm eclectic: I pick up random artists from anytime in the last 50 years. I have pre-teen kids, so I get exposed to their music, and of course I'm old enough that the 80s was my own teen years, so I've a soft spot for bad music from that era!

But to give a somewhat factual answer, the most recent song I played in the car was Disturbed's version of Sound of Silence, if that helps!

I joined, um, mid-2023? But indeed, this process has brought me back!

Now I just need to figure out how to connect my CLN instance here: there's neither a commando nor a BOLT12 option :(

CTV is a short-cut, and I have trouble evaluating whether it's the right shortcut! Interestingly, CSFS is more straightforward: it's just one part of an existing opcode, after all.

ln symmetry would be a rework, but not an unimaginable one. For CLN, we have a specific daemon which handles channels in the "normal" case (channeld), and we'd probably just write a completely new one for LN symmetry.

There are two things I'm aware of: one is the idea of batching withdrawls to lightning channels which is made less interactive by a covenant (https://x.com/Polyd_/status/1714566295813841117). There was also an issue with requiring another round of interaction during HTLCs which Bastien Teinturier believed could be avoided, but I don't remember the exact details either, sorry!

I've never thought about it!

I think if people want it badly enough, they would figure out a way to bootstrap again. But if they wanted it so badly, why did it fail the first time?

You'd prefer an exception? What's the difference?

We deliberately crash on bugs. It's the safest thing to do. We isolate the different sections into completely separate binaries, any one of which can go down without breaking the rest. This is also why we use sqlite3, which means that we need to get to the end of a transaction commitment, otherwise it Didn't Happen.

My "vibe coding" is when I write a shell script with "cat >".

When I've tried to use AI assist it's been very mixed. There's a point where it goes from being a useful assistant to a massive PITA, and it takes care to avoid crossing over that, especially in areas where you know least, and thus are relying on it most!

I think it will be similar, but with more large players. There will be more people offering hosting and liquidity provision for businesses; I hope via a combination of liquidity ads and the LSP spec. There will be technical improvements, and I hope we're on gossip v2 and deprecated v1, for much efficiency win.

On the payment side, recurring BOLT12 should be standard by then, which will open a host of new uses and conveniences for regular payments. This might help crack some subscription-style business models which are hard with the current lightning network...

Not much! I'm usually happy if things are directionally correct, not that every single thing is going right.

Generally I'm happy to set about 10-15% of the network (though obviously it's hard to tell!). From a specification perspective, this is a healthy number, which ensures everyone stays friendly and interoperable, and makes it worth developing new features together with other implementations.

To some extent, more users, more problems!

I'll be back answering questions later! But I will get to them!

I grew up here, and my family are here. I have an entire non-Bitcoin life, in fact! Adelaide is the easy life: 20 minutes to everywhere, great weather and beaches, good wine country.

Australia has its problems, as do all places, but these are mine to help sort out.

Lightning was overhyped before it existed, then boring while people focused on Store of Value, and then back in vogue in some parts. I still think people with functional banking systems don't care about spending Bitcoin (and thus, Lightning).

Progress often works this way: my Linux experience was the same. In 1998/99 there was an unjustified hype, at the tail end of the dot com boom. Then it was tumbleweeds: all those "trial projects" vanished, and so did the headlines. But the developers barely slowed down, and in the 2008 crisis when businesses were actually trying to save money, Linux was ready for them.

Perhaps bitcoin payments today are for the weird: those with strange tastes, or unloved by the current systems. That's OK, we're here for you if (when?) that changes.

It would be Rust. The language is maturing fast, and I'd be prepared to place a bet on it. The performance is nice.

Re: memory ownership, we use tal which provides many handrails with memory management (I wrote it based on the ideas from Andrew Tridgell's talloc library). I would not write a significant C program without it.

Hmm, there are so many. Probably the sql plugin! It turns the list* commands into SQL tables, and you can write arbitrary SQL queries, and it's all done in a way which is safe and read-only. This is vastly underutilized, and there is also more optimization we can do inside the sql plugin to cache stuff and make it even faster...

Wildness gives a degree of strength and violence which is a definite advantage over me and my little keyboard hands. Have you ever tried to wrestle an unwilling housecat? I mean, you win, but you're not coming out of that unscathed!

And I'm in Australia, so even small things can kill you. I think I'm walking away from this question, and not making eye contact...

I wouldn't. I'd put them on spec work. And it would be Andrew Tridgell, who would surely go and write his own implementation in his spare time anyway :)

"Smart" isn't a single dimension, and at some level I lose the ability to evaluate: you might be a brilliant genius linguist, but how would I evaluate that?

In the realm of coding I would definitely put Andrew Tridgell (Samba, rsync etc) and Linus Torvalds (Linux, git, etc) up there as some of the smartest people I've worked with, and hands down Andrew Tridgell was the nicest to work with.

But I don't really think like that, I use a simple yardstick: are they smarter than me? If so, and they're nice, I want to work with them! In the lightning world, there's no doubt in my mind that Lisa Neigut, Bastien Teinturier and Olaoluwa Osuntokun are smarter than me, for example: my ego simply serves to make me strive to keep up!

Hmm. From my POV, working on big project where there were multiple eager employers who would pay you to work however hard you wanted on software they were giving away is such a dream come true that I find it hard to see a real problem here.

On the other hand, my life has been blessed. I never expected this to happen; I expected this FOSS stuff to be a hobby project at best!

I started hacking on g++, the GNU C++ compiler, because I was using SGI machines in my first job and the boss was too tight to pay for enough compiler licences!

It was the early days of the internet and I got exposed to some genuine next-level devs. In particular, I went to a conference and spend two days with Linux developers, and I knew that these were the people I wanted to work with. So I started hacking on a side-project I'd begun, which became ipchains, and next thing I know I'm being paid to write Free Software! Had no idea that would last for over 25 years...