WBD094 Audio Transcription 

Nic Carter & Pierre Rochard On The Economics Of The Lightning Network

Interview date: Tuesday 9th April 2019

Note: the following is a transcription of my interview with Nic Carter & Pierre Rochard. I have reviewed the transcription but if you find any mistakes, please feel free to email me. You can listen to the original recording here.

In this interview I talk about the economics of the network with Nic Carter and Pierre Rochard, We discuss the fee market, the future of the block subsidy and whether base chain fees will lead to a two-tier Bitcoin.

“To me, it is pretty staggering actually that Bitcoin has never had a real confirmed double spend.”

— Nic Carter

Interview Transcription

Peter McCormack: Hi Nic. Hi Pierre. How are you both doing?

Nic Carter:  Hi Peter. Doing great man. Thanks for having me on.

Pierre Rochard: Hi Peter. I'm doing well. How are you?

Peter McCormack: I'm doing very well thank you. Thank you both for coming on. Obviously a second time for you both. As you are aware, I'm doing a Lightning month, learning everything I can about Lightning and one of the things I did was I tried to keep away from Lightning quite a bit before I knew I was going to do this. Because I wanted to approach everything fresh. Now I've been looking at it. I've got so many different questions, so many things I want to explore. I don't know if my questions are dumb or fair, but I think we'll find a way through it through this session.

So we'll just roll with my questions and see where we come out. A good place to start with is, one of the things I'm trying to understand is the difference between the Bitcoin base chain and Lightning and whether we're going to get to a stage where we have a two tier Bitcoin where the base chain is only for those who can afford it and Lightning is for everyone else. Any thoughts on that Nic?

Nic Carter: Yeah, that is very good question. I don't view them as distinct. I see Lightning as essentially just encumbered Bitcoin with certain conditions required for its release. So conceptually to me, Lightning is really just a name for an elaborate system, smart contracts, which mediate what you can do with your Bitcoins. It's an arrangement that people enter into consensually, so that they can unlock some additional usage out of Bitcoin.

I'm not an expert on this, but the direction that things seem to be going is to further blur the lines between Bitcoin held in Lightning and Bitcoin itself. This notion of bifurcation seems strange to me, because in my opinion, for Lightning to work, you need to have periodic closings and settlements to the main chain. So if the main chain for some reason gets stupendously expensive to use, my guess would be that that cost is inherited on top. I don't think that's likely.

I mean, no one has ever really worked out what the sort of required level of fees is, but it doesn't look likely to me at all, that we're going to be in a really high fee regime anytime soon. So it's not necessarily something I'm too worried about it

Pierre Rochard: Long term, I think that every single Bitcoin will be on "Lightning". I don't think that anyone will be using the base chain except for closing a channel and simultaneously opening another one. I don't think the whales will be sending on chain transactions instead of sending off chain because I think that, if we think about Bitcoin versus Lightning or on chain versus off chain, we're thinking about a set of trade offs.

The most important trade off is that you have to be online for Lightning. Whereas with on chain Bitcoin, you could be offline forever and not have it be a problem. But I think that sadly, perhaps our society frowns upon people being offline for any reason, not just Lightning! So sadly, I don't think any of us are going on long Appalachian hikes for months at a time and thus being offline for long enough where this would be a problem.

So I think that everyone's online all the time and so that trade off is perfectly acceptable. Then the other trade off is the hot wallet situation. I think that the technology around this is going to continue to improve over the coming years where essentially we'll no longer worry about cold storage. I think that people will be completely okay with hot wallets and they'll be secure and the security problems won't be around the software or the hardware getting hacked.

It will be about people getting socially engineered into sending their Bitcoin. So I think that eventually all Bitcoin will be on Lightning and no one will go out of their way to use the base chain, because it's just a worse system than off chain.

Nic Carter: I might push back on that actually a little bit. I think I broadly agree with the thrust of what you're saying, but base layer Bitcoin still has pretty attractive properties and let's say you're like a political dissident and you're in jail for five years. You have a Brave wallet and you posted the QR code before you went to jail. I think you should still be able to collect donations in that time, even though you're totally offline. That's just one simple case where I think base layer Bitcoin is still preferable.

Peter McCormack: I think obviously though we're a long way from being at a stage where all transactions are done through Lightning and the base chain is just for opening and closing channels. We're obviously a long way away from that. One of the things I'm trying to get my head around with this is that I'm trying to understand the different trade offs because there's a few things that have come together for me at the same time I'm trying to figure out.

So the security of the Bitcoin network is reliant upon miners and often the discussion comes up about what happens as the miner rewards drop. We're going to require a fee network. So if we require higher fees, therefore the base chain is going to have limitations to who actually uses it.

I'm getting lost in this circle now where people are saying, "we don't need to talk about the fee network because this is years away, for the mining rewards to drop so far." Yet at the same time there will be implications later on. Can you see where I'm getting lost in this circle?

Nic Carter: Yeah. It's tricky because the problem with very openly discussing potential shortcomings that you're worried about, is that they can then get magnified by people who might be operating in less than good faith, who are looking for the most compelling attack factors against Bitcoin to promote their own scheme, for instance.

So I would say that is part of the reason why discussion of potential issues is not as open as it might be otherwise. That said there is definitely a school of thought within Bitcoin land that actually this isn't a problem and that even as the subsidy goes away, maybe the absolute properties will change a little bit, but it's not as much of an existential threat. So there's still a debate on whether it's actually a serious long term threat or not. I'm sure Pierre has thoughts on that too?

Pierre Rochard: Yeah, I do think that it is a long term threat. I think the way that we think about it today is just, it's like going back in the 1920s and looking through a book about the future, "oh, there's going to be flying cars, like the Jetsons" and we try to project out, what it's going to be like in 30 years. But then in 30 years there's going to be some similarities, we're going to make some calls. But for the most part, our expectations about the future are completely unrealistic compared to what actually happened.

We just have to focus on what's within our control today, rather than trying to figure out a solution to something in 30 years. The way I see it is that basically we have to maximise the productivity of block space and that's what's going to make it so that people are going to pay high enough fees where we can avoid having to have tail emission or long term inflation or whatever people want to call it.

But the other thing too is that people when they're discussing this topic, they'll use the word "purity", which implies that if there's not enough security, then the system failed on a global basis. Whereas really what we're talking about is probabilistic transaction finality, where if a given person has an insufficient amount of finality, that actually doesn't cause the system globally to fail. It might mean that they get double spent on locally.

So what we've seen is that a local failure like that, and it happened with Ethereum Classic recently, but I mean with these things people say, "oh, well that's not the same" or whatever. But I really think the economics and the logic of it is pretty clear, which is that it's only person who got double spent on, that "loses faith" in the system. But really they're just going to wait for more confirmations.

To me that's the other solution to the blocks subsidy going away, is just "all right, well instead of waiting for that rule of thumb with confirmations, now we wait for 60 or 600 or 6,000". Even that's not a huge problem for people who aren't ordering Boeing 747s from China!

Nic Carter: Yeah. Just to add to that, I would say it would be extremely fruitful for some of the more mathematical, theory minded people in the industry to revisit that 6 confirmation rule of thumb and determine or devise a dynamic model for confirmations. Or just maybe a more justifiable... Because the six confirmation rule of thumb comes from almost a throwaway comment Satoshi made at one point.

I think there's other variables that come into play there, including the size of the transaction you're accepting, what has happened recently with hash rate and whether you have a non-repurposable hardware on your chain, whether there's a big secondary market for that hardware. In my opinion, there are many more factors that are highly salient there

Pierre Rochard: And the factors you just listed are just the endogenous factors to the chain. If you look at the exogenous factors, like who is it sending the transaction, do you have legal recourse if they do defraud you by doing a double spend. In which case that obviates a lot of induction of concern.

Nic Carter: That's true. But I don't know if we want to make Bitcoin's security model dependent on current institutions.

Pierre Rochard: I do think that... We're talking about local transaction finality, not Bitcoins global trustlessness properties. So in that regard it's always going to be the case that there's other concerns than just the on chain situation.

Nic Carter: Yeah, I would say that it's contextual and if you have other information, why not use that to build a pattern of evidence, which then plays into the amount of confirmations we're willing to accept. The other thing I'll say on this point broadly is that it's really weird to me, but people when they model out Bitcoin's longterm security etc or security budget.

Then they use that to guess at what fees are going to be, they just take today's security expenditure and model all their assumptions on that. What would it take to extend today's security budget into the indefinite future? To me that's totally crazy because the security budget today is just a function of the issuance and the price. The price is completely contingent. Nobody has any control over that. So it could be that we're vastly overpaying for security. We just don't know or underpaying. That seems unlikely to me.

Pierre Rochard: Yeah. I would argue that we're overpaying considering how we don't have any double spend. So to me underpaying would be, double spends are just so pervasive that it's making the chain unusable or getting close to that is unacceptable. But then having zero double spend, to me that's an indication of overpaying and that it wouldn't be the end of the world if we had a few a year.

Nic Carter: But yeah, totally agree. To me it's pretty, well staggering actually, that Bitcoin has never had a real confirmed double spend as far as I can tell, you know a malicious exploit. Even though there have been issues with the chain in the past. So it would be indicative of the fact that we might actually be overpaying.

Peter McCormack: So one of the things I try and do is rather than predict what will be with the on chain fees, is trying to imagine what the implication of higher fees are. So I've seen recently that the number of on chain transactions is going up and up and up. We have the potential that through another bull market, the growth of the technology, the use cases and of course the Lightning network, that that will continue to increase and perhaps on chain fees may be $10, $20, whatever they are.

A couple of notes that came back to me was one person said that, "$60 fee is a good price to pay for a transaction on a secure immutable ledger," and then there was another comment from someone else who said, "yes, valuable things are scarce". So there's almost a defense therefore of high on chain fees. But what I'm trying to understand is what is the implication for people who may be want to get involved in Bitcoin or they aren't Bitcoin rich? Does that mean they are therefore excluded from the base chain and the majority of their experience is on Lightning and Lightning only?

Pierre Rochard: Yeah, I mean just mathematically. If someone wants to buy $20 worth of Bitcoin and the on chain transaction fee to send it to their wallet, is $10. That's just an unrealistic situation where they would just go use Lightning instead. So it's on us to make sure that Lightning is suitable for that.

Nic Carter: Yeah, I think something people miss is the fact that you can onboard directly into Lightning. In the near future you will be able to with many exchanges and Bitcoin banks so to speak. So I happen to know that there are many active projects working on that, so you won't have to onboard into base layer Bitcoin and then do a cumbersome transition to Lightning. It'll just be much more seamless.

So of course transacting on Lightning is not free because you do depend on the ability to exit to the base layer potentially. But the other thing I'll say is that, in my opinion, fees are never low or high. They just are a certain amount. The system is this emergent dynamic system that outputs a certain level of fees in Satoshi terms. The reason people really think fees were super high in late 2017, was because the unit price was high. So it's high in dollar terms.

But if you look at the spikes and like sats per byte in terms of the system that doesn't know what the prices in dollars, three of those spikes... We had a spike in June 2017 that was about as high as the spike in November 2017. So the reason people really focus on the $50, $60 fees is because that was when the unit price was $19,000. So yeah, I would say that's kind of a distraction. People index into that, which doesn't make sense.

Everyone's been complaining about the high fees recently. If you actually look at the chart of fees, they're still minuscule and it's just ridiculous that a scarce spot on the world's most secure ledger can be yours for about a dollar right now or less, much less in practice. I don't think fees are ever low or high. It's just a function of your expectations and if your expectations are miscalibrated, that's almost your fault.

Pierre Rochard: Yeah. A bigger problem really in my mind is the volatility of the fees. So trying to have a good fee estimation so that you're not overpaying or underpaying and the experience around that has been a bigger issue than the absolute level of fees.

Nic Carter: Yeah, totally true. It's because fee estimation is a three body problem. It kind of reminds me of trading. You're necessarily trading against some pure piece of information. You're trading against the expectations of all the other fellow traders. So when you're estimating fees, you can look at historically what typical fee to enter the next block should be at this level of usage and you can build those kinds of models.

Everyone else is also looking at the same fee estimator. So you almost want to outbid what someone else looking at the same estimator would be bidding. So you have these weird situations with feedback loops in the fee estimators. I think poorly calibrated fee estimators were a cause of the crisis. You can always look back at historical Sats per byte in previous blocks and see that people were way overspending for fees.

Pierre Rochard: Yeah. What I'd really like to see develop, is a futures market for on chain transaction fees. So that if you open a channel today and you say, "okay, I want to lock in at certain Sats per byte, in six months when I'm going to close this channel", then I can enter into a futures contract to fix that, so that now the volatility doesn't affect me.

Peter McCormack: Do you not think charging fees on Sats per byte is a bit too confusing for most people?

Nic Carter: Yeah, for normal users. It should be up to the wallet providers to be sensible about it and to kind of abstract that away a little bit.

Pierre Rochard: We've seen wallets create UIs where basically they say, "low, medium, high priority". So there's ways of communicating what the underlying system is doing to the end user. But I also think that there's room for user education as well. So if we're able to learn this stuff and I think a lot of other people will as well, for mainstream consumers, the wallet really should be taking care of that.

Peter McCormack: Say I have $100 of Bitcoin and to me that's a lot of money where I am. I've got it on the base chain, but over time the fees are going up on the base chain. So will I potentially have to make a consideration of moving that all onto Lightning so that I don't end up losing most of it in on chain fees?

Pierre Rochard: Yeah. To me it depends on to what end you're holding those coins. So I think that Bitcoin's on chain transaction fees are going to continue to be volatile and evolve. It's not clear to me that it's going to be a steady linear trend up or an exponential trend up or like sideways. I have no way of knowing where it's going to go. So if someone says, "well, I'm going to be holding these Bitcoin for several years", then they can wait for an opportunity to move those coins at very low transaction fees.

But if they're going to be sending it within the next few days and on chain fees are going up, then yeah, there's a sense of urgency to actually getting those coins onto Lightning. But the other part of the equation too is that we don't know what future protocol changes will allow from the perspective of... Essentially like what you're describing is someone being concerned about their on chain funds becoming blow the dust limit, sort of like where it costs more to move them, than the value of...

Peter McCormack: Or even just moving part of it, will chew up such a high percentage of their Bitcoin, because of the fees.

Pierre Rochard: So there may be protocol changes in the future where the cost of moving those Bitcoin goes down a lot. So we have Schnorr signature aggregation on the horizon. Then there's everything that we don't know that we don't know yet. That might mean that it's not a problem. But generally speaking, yeah I think that people should be thinking about opportunities to get onto Lightning and to consolidate their outputs on chain.

Nic Carter: Yeah. I would echo what Pierre said. The development trend in Bitcoin seems to be towards more parsimonious usages of block space, with Schnorr being a great example. Just the mere adoption of SegWit really moves needle a lot. So SegWit gave us a 2.53x increase, if people actually use SegWit a lot in block space. So a really big efficiency increase. I mean it was an effective block size increase.

The effect of that is very stark. Right now, we're at a period of very high utilisation of block space, almost the highest it's ever been in fact. If you look back to the last time block space had this much data through-putting through the Bitcoin system, it was between about December 2017 to February 2018. Fees were never below $5 back then.

Now fees on a weekly basis are about a dollar. So we've already seen a increase in capacity and the next set of innovations they're going to be an increase in density. So there is considerably more slack built into the system now. It can tolerate a lot more usage and there's more education to back then.

A lot of the big power users and Bitcoin exchanges are not matching, they weren't using SegWit, they're being extremely wasteful with block space. So things are really different now and everyone is kind of shell-shocked by what happened in late 2017, but conditions are really different now. So I really don't expect fees to climb anytime soon.

Peter McCormack: So really it's just a case of putting the trust in the future technologies, optimizing the use of block space and at the same time in parallel, Lightning improving and having a better experience for people just to live entirely on Lightning?

Nic Carter: That and just to finish up, I would say calibrating expectations. If you want a spot on basically on base layer Bitcoin, it's probably going to be a couple of bucks. That would be my guess. If fees are basically zero, then you do run into those instability problems. So I think if you want to have your cake and eat it too, you want zero fees, but you want longterm security and you want scarcity, you're probably not going to get that.

Pierre Rochard: Conceptually, the way to think about on chain versus off chain, is that on chain you're paying by the amount of data that you're consuming. So there you can move hundreds of millions of dollars worth of Bitcoin with $2. Whereas off chain you're paying by the amount of value that you're moving because you're exhausting channel capacity in one direction or another.

So there, you can move $20 worth of Bitcoin for a fraction of a penny, but if you're trying to move $600 worth of Bitcoin, then it can become very expensive. Where to the point where it's less expensive to move it on chain instead. So there's kind of a trade off.

Nic Carter: Just to butt in there, I just listened to your podcast with Christian Decker and this is why it really resonated with me. He calls Lightning complimentary to Bitcoin. They are priced in different ways. Lightning is great for small to medium sized payments. It's priced, basically on a bandwidth basis and Bitcoin is great for really colossal transactions which maybe you can wait a little bit longer to settle. So they work extremely well together.

Peter McCormack: See one of the things that Christian talked about, which I'm just slightly trying to get my head around, is splicing, where he talks about the wallet, when you choose to make a payment, it will figure out whether it will be best to be on chain or off chain. How much do you understand about splicing Pierre and how much can you add to this?

Pierre Rochard: I definitely don't understand anymore... I understand a lot less than Christian Decker understands! But conceptually it makes sense to me that you could add capacity to a channel with just one on chain transaction, instead of having to close the channel and then open a new, bigger one.

Peter McCormack: So let's talk a little bit about Lightning as well. When I interviewed Jack Mallers, he sees a scenario where there's going to be plenty of people who will use Lightning and Lightning only and never touch the base chain. They will have a Lightning wallet, they will buy some Satoshis and use them in any way they want.

But they will stay entirely on Lightning and never come off that. One of the things I didn't understand about that, does that mean there will be channels that will never close? Then secondly, as I've moved Bitcoin onto Lightning myself, my Satoshis in my Lightning wallet are represented by my Bitcoin, which are held on the base chain, which opened up with the multisig. Who owns those Bitcoin for the person who bought the Satoshis? Do you understand the question?

Pierre Rochard: Yeah. You still own the Bitcoins. The metaphor I use is, when we're using the legal system and we have contracts, we don't go to court every time we rely on a contract, we only go to court when we have a dispute on the contract. It's still the case that we have a contract, even if we don't go to court and legally prove it.

That's kind of the way I think about off chain, is that we don't go back on chain unless there's a dispute or in the case of Lightning, you need to close the channel because you need to open up liquidity in a different direction. But the idea being that you still have a perfectly valid Bitcoin transaction that you're sitting on, so you still own all those Bitcoin, it just hasn't been broadcast to the network.

Peter McCormack: But am I essentially buying that Bitcoin of an exchange then? But I only have it represented as the Satoshis or is there some part of a multisig contract that represents that part of that Bitcoin I own?

Pierre Rochard: Well both. So let's say you were to go use an exchange and you're buying on Lightning, you're buying Bitcoins on Lightning and receiving them with the Lightning Wallet. So they would open a channel with your Lightning Wallet, where they have outbound capacity and you just have inbound capacity. Then at that point they would send over Satoshis to you.

Now that transaction or that kind of flow of payments is the same in any kind of situation, if you just leave out the exchange. So it is the case that you are updating valid Bitcoin transactions every time that you update a channel. A channel is balancing. So you definitely still have an on chain transaction that you could broadcast to the network if you wanted to.

Nic Carter: Another way of putting it would be that if you buy Lightning from an exchange, you buy the option to transact base layer Bitcoin.

Peter McCormack: Okay, so am I at any risk that that channel gets closed by somebody and therefore my Satoshis on Lightning therefore become a Bitcoin base chain transaction?

Pierre Rochard: Yes, absolutely. There we run into some of the issues with what's being developed, which is called Autopilot, which allows for automatic management of channel. So a channel could hypothetically be open forever. There's nothing about it that makes it expire after a certain amount of time. So ideally all channels would just stay open forever. But we don't live in an ideal world and the people who are on the other side of the channel at any point can close the channel on you.

That's where the software in the future, will automatically figure out, "okay, this person has this actuarial expectation of closing a channel. So we're not going to open a channel with them, because they're unreliable. We're going to find someone who almost never closes channels on others". So that might be the way that the software optimises to avoid that situation. But I just want to reiterate that this will all be automated so that the end user does not have to worry about it themself.

Nic Carter: But if you are dealing with a malicious counter party to take recourse, you might have to do an on chain transaction.

Pierre Rochard: Yeah. So if the person in the channel decides to breach, so that is they broadcast an old channel state or they're just offline and thus tying up your liquidity in an inactive channel. You are kind of forced into a situation where... Well with the breach transaction, they've already broadcasted an on chain transaction. So you have to go and punish them.

Then you get all the funds from the channel. So in that situation of punishing a malicious person, you actually profit off of that, assuming you're online. If you're offline, then you lose the money. But in the case of someone who is just unresponsive and is a useless peer on the network, then you do have to close channels. What I would be worried about isn't so much an individual having this problem.

If we have a network wide issue where, for example, right now there's a player called LNbig and they've been opening up hundreds of channels with different peers on the network. So if they were to go offline or if they were to suddenly decide they're going to force close all the channels on everyone, then that would be disruptive to the network. I think that network wide issues like that are a bigger concern in my mind then a local problem because someone generally should have like eight channels open with different peers, so that if one goes down, then you're still resilient for sending payments on Lightning.

Nic Carter: Not to second guess Jack Mallers, who knows a lot more about Lightning than me. But this is why I'm a little hesitant to say that users will always transact on Lightning. These situations where you need on chain recourse are likely vanishingly rare, because the attackers can basically be punished. But you still, to protect your interests, you still need to be able to transact on chain under certain circumstances.

Pierre Rochard: Right, and so the channels maintain like a reserve balance for making an on chain transaction and the peers are constantly renegotiating that reserve balance. What I noticed with the latest fees spike was that if someone had a channel open that was for 50,000 Satoshis, the reserve kept growing until it took up the whole channel. So we have to watch out for that as well. We have to open channels that are large enough to justify having a larger reserve that is for the fallback mechanism of falling back on chain.

Peter McCormack: So one of the things that stands out to me, in my world of looking at everything as dumb as I can, is that in the simple world of Bitcoin, I have a wallet. If I have to make a payment to somebody, I put their address in, set the amount and put low, medium, high with my price and it sends. At some point in the next hour or two it confirms. It's very simple and it's very easy to understand.

In the world of Lightning now with my limited exposure, there's so many other things I have to think about, that I don't have to think about with Bitcoin. I have to think about channels and channel capacity and what happens if it gets closed? It feels very, very complex. Does that concern you at all? Because for me, I think it's almost, it's putting me off using Lightning.

Nic Carter: In my view, the complexity is partly due to the newness. It's hard and unfamiliar to us to think about payments this way, as essentially bandwidth that can be consumed. I think partly or epistemically very comfortable with Bitcoin itself because you understand it very well. So you don't really think about sha256 when you're transacting. You've fully grokked the system and you don't have to worry about the minutia.

Lightning certainly is more complex. I think about Lightning in terms of a different settlement system, which is one of the basic primitives of finance is just achieving efficiency through different settlements with counterparties that you might trust a little bit more. I don't really worry about all the complexity. I would say that's the challenge to the developers, is to abstract it away a little bit. But certainly it's new to most of us and hence maybe more challenging.

Peter McCormack: So these complexities are really just building blocks to a future final easier solution as probably Bitcoin was probably very complex in the first few years and has got itself to a point where it's really easy to understand. Is that what you are getting at?

Nic Carter: Of course it was. Comparing the difference between using Squares, Cash app to get Bitcoin, to the way you had to get in the early days. You had to run your own node etc. I mean I still recommend running a node, but acquiring and using Bitcoin was way more complex and required much more sophistication back in 2010.

Pierre Rochard: I still hear people complain about addresses being like random numbers and letters instead of having an email address or something like that. For people there's still concerns about Bitcoin's base layer UX. Then on the way that payments are done on Bitcoin, that's not great either from the merchant's perspective.

When I worked at BitPay, overpayments, underpayments, payments that were late essentially and people were trying to game the exchange rate and all of that. Those were problems and those problems are largely solved with Lightning's invoicing system. So I think that there's UX gains with Lightning as well, aside from just the cool experience of seeing a payment instantly get confirmed. Sorry, not get confirmed, but you know what I mean.

Nic Carter: The other thing is that we can afford to indulge in some more complexity because we are progressing in this layered way. So I think it's important that the base layer system is fairly simple. But we can experiment and try really complex features on Lightning because the costs of failure are less. If Lightning fails, Bitcoin itself is totally intact.

Peter McCormack: Pierre, how much can you tell me about how the fees in Lightning work and there's been a lot of talk about the need perhaps for mega hubs. Can you talk to me about that and explain how that works and what the risks are and what the fears are?

Pierre Rochard: Yeah, sure. So if you have a channel open directly with the person that you're sending money to, then you don't pay any Lightning routing fees. But you do pay an on chain transaction fee to open that channel and then eventually to close it. The problem with doing is that now you have to open a channel with every single person that you're sending money to. That's not very useful.

So Lightning allows you to route payments where you can be connected to a node and that node is connected to the person who's receiving the money. You're able to use them to route your payment and they charge you a routing fee for that service. They actually can set that fee themselves and it's also publicly advertised. So if you open up eight different channels, that means that you have eight potentially different routes to get to someone else, which means that you can pick whoever is advertising the lowest fee and be able to send the payment that way. So that means that this should be a fairly competitive market.

Now the advantage of having centralised routing nodes is that then you're minimising the number of hops. So if everyone is connected to one routing node, then everything is going to only be one hop and there's only going to be one routing node charging fee. So that would be very efficient. The problem with that is that it's not very resilient. If that one routing node goes down, then the whole network essentially falls apart because no one can route anything to anyone else.

So what we're seeing right now with the current evolution of the Lightning Network topology is that we're in between those two extremes and so there are big routing nodes that are emerging, but at the same time, most people are connected to many different nodes. So it's not like those big routing nodes have some kind of large size, economic influence or are a big risk to the network if they were to go down. Now what I think we'll see with the routing fee, is that right now it's a very inefficient market.

There's very rudimentary tools for managing, "what should my routing fees on channels be" and then also, "how do I rebound things so that I continue to route efficiently as capacity gets depleted in certain directions". So all of these things that really need to be automated, there's a massive amount of work to be done there. So to me Lightning is still very much in the Stone Age, but we have a very... I can see the horizon and I can see what we can do imminently with the existing technology.

There's just a lot of infrastructure that needs to be built up around that. Now, I was just talking about the routing fees, but I think that actually a bigger source of fee revenue or key costs depending on which end you're on, so far has been about getting inbound capacity. So ideally you would just get inbound capacity for free, because other nodes are randomly connecting to you. But that actually takes quite a long time because the current iteration of Autopilot, has this model called "Preferential Attachment" where it'll preferentially attached to larger nodes and then also attach to the smaller nodes.

So that means that the larger nodes are accruing more and more channels and it can take a little bit of time if you're a smaller node to just get inbound capacity for free, randomly from other peers. So services have emerged where if you need to receive payments or you want to be routing payments and you need an inbound capacity, you can go and pay someone to open a channel with you and provide that inbound capacity. So there's Bitrefill for products that does that, Lightning Loop does it in a trustless atomic manner with one on chain transaction and we can get into how that their product works called Loop.

Then I have lightningpowerusers.com of course, which is my website where you can buy that in that back capacity as well. So I think that we'll continue to see new pain points where there's an opportunity for some people to make money and it's going to cost other people money. In this case it costs merchant's money because merchants have to go pay someone to provide inbound capacity to their node so that they can be shipping orders and receiving payments.

Eventually what I'd like to see is for there to be a protocol level marketplace essentially for capacity, where we can have a decentralised exchange for inbound capacity. But we'll see how realistic that is.

Peter McCormack: Another thing Pierre that I wanted to ask you about is obviously there's an on chain fee when you open a channel. I guess for you to benefit from the low cost of using the Lightning network, you've got to hit a certain number of transactions for that to become economically worthwhile? Because if you're opening the channel just to send a single transaction, unless you need a fast transaction, there's no real benefit.

Pierre Rochard: Yeah, that's correct. So it really has to be the case that you plan on reusing this repeatedly. Otherwise it's like getting a rechargeable battery and throwing it away after the first use.

Peter McCormack: Are the fees going to be high enough within the network to make it economical for these mega hubs to be out there providing capacity?

Pierre Rochard: We have no idea of knowing really. I think that there's different theories about whether it's going to be hyper competitive, where the revenue is going to be very small or it's going to be like a loss leader where people are running a routing node but they're providing other services that are ancillary to that. The benefit from having access to a routing node and that's where they make money.

Or is it that routing is actually going to be profitable specifically for people who specialize in it and actually have skills about essentially being a trader? Where they are trying to figure out, "all right, where is there a high demand for capacity on the network? Let me connect to a node in that area of the network. Then I know that they'll connect to a node in this opposite side of the network and then I'll be able to route lots of payments."

There's different views on how profitable that's going to be. Then on a macro level, we were talking about the trade off between on chain and off chain. So the point at which it makes more sense to send a payment off chain versus sending an on chain transaction, that equilibrium is going to constantly be changing.

So we'll see as on chain fees go up, will that cause routing fees to go up, where essentially now a routing node is saying, "well, the cost of me opening and closing channels went up, so now I have to pass that cost on to people who are using my routing node and so I'm going to charge higher routing fees." So I think that's definitely a plausible scenario. We just have to see it in practice and run the experiments.

Peter McCormack: So it feels like everything is just very exploratory. We're still very early. We actually need to see how people are using Lightning and seeing how if handles the pressure and the stress of lots of people on it. So that either, those who are running these mega nodes can see what the economics are for themselves, but also for developers to react and prioritize developments that improve the experience.

Pierre Rochard: Yeah, that's right. I think that if you talk with the developers, it's interesting as they see their job is making it so that it's not profitable to run a big routing node. Because that introduces this issue of centralisation, now that this network is not as decentralised as we wanted it to be and it kind of plays into the hands of the critics of Lightning.

Peter McCormack: Okay. All right. Well listen, this has been great! It's been useful because I'm doing so many interviews this month, I'm trying to not leave them too long. Is there anything either of you want to add? Also just as a final question, is there any parts of the Lightning network that you have larger concerns about?

Nic Carter: No, the main thing I'm worried about is actually just expectations management. There is always a curious divergence between the very cautionary things that the actual developers working on Lightning will say, about its ability to mainstream itself, to absorb a huge amount of capacity and work well for everyone, even specifically on new users.

There's a contrast there between what they say and then between what just "boosters" are essentially saying. There is a really significant risk I would say, which is the risk of Bitcoiners or outsiders becoming disappointed by the pace of progress there. I think people like Christian Decker do a really good service where they bridge that gap and they'll qualify the pace of development and make it clear what the constraints are, in that Lightning is maybe not a silver bullet to solve all Bitcoin's problems.

That's fine because that was never the expectation. But if that is the expectation, then it becomes kind of a crisis for Bitcoin. In part the block size wars were due to super inflated expectations about what base layer Bitcoin could do. Then the delta between those expectations and reality. It's not that the reality was wrong, its that the expectations are too high. I've seen a little bit of this on Lightning, it being presented as this final solution to all of Bitcoin's issues.

I think it's a really vital piece of the puzzle and a really sophisticated network, which is way beyond its years already in terms of the amount of adoption it's getting. But I think it's also worth just recalling that Bitcoin itself works fine. If Lightning doesn't grow to the desired size, that's not fatal to Bitcoin. We're not necessarily indexing our hopes for Bitcoin on Lightning, although I think Lightning is the most promising development in Bitcoin right now. So that's my note of caution.

Pierre Rochard: Yeah, I completely agree. I found the best way to manage people's expectations, is just have them use it. So that's why I created the Node Launcher. So if you want to try out Lightning and let reality set your expectations instead of Twitter. Then that way you can see what the promise is, which is people who have been able to send payments are raving about it and then what some of the pitfalls are. I do get lots of support requests from people who are having technical difficulties with using the Lightning network.

So to me, there's no better way of adjusting people's expectations than sticking their nose in it. So I'd highly encourage people who are curious about it, to not spend too much time reading about it and actually go use it and then read about it! That way you can have a taste of where it currently stands. So I've written a Medium post, which actually I just edited. I deleted two thirds of it because it has gotten so much simpler with new versions of the Node Launcher. So definitely go check that out.

Then I guess my closing thought would just be that, I do think that having a system, whether it's Lightning or something like it, where you create a lot of demand for on chain transactions and any elastic demand, where essentially there's no substitution effect if on chain transaction fees go up, where people are going to start using Litecoin instead.

So having any elastic demand for block space I think is going to be key to driving up transaction fees to a level where they are providing suitable transaction finality and we don't end up in double spending hell, when the block subsidy has an increasing amount.

Peter McCormack: My summary is that it's just super early. If anything, it's still a bit of a toy, it's a bit of a play thing. I think you have to have a certain amount of knowledge and you have to have a certain amount of experience. I don't think it's a commercially ready technology for the masses yet, because there's too many complexities around using it.

That's my experience so far. But at the same time, the handful of times I have used it, it's been mainly a pleasure and I haven't really had any problems. I don't think it's a smooth journey from the Bitcoin base chain to using Lightning just yet.

Nic Carter: No and many earth shaking technologies start off looking like toys. The reason I think people are so excited about Lightning in spite of some of the UX difficulties of using it, is because it speaks to the layered scaling approach on Bitcoin. So far it's proven itself to work, which you know for the longest time, it was just a far off, deeply held belief and hope, but it wasn't a reality.

So the fact that in these last few years we've come from discussions on mailing lists, to paper and then a specification and now many competing viable implementations. That is pretty astonishing! It reflects the maturity of thought within Bitcoinland, that we have conceded that we can't just open that block space throttle indefinitely. We're now in the process of building out this overlay network, which makes Bitcoin true to maybe some of the early visions for what it would be. That is extremely exciting and important!

Pierre Rochard: I think that Bitcoin as an investment and as something that people buy to speculate on, is going to have much faster and has had, arguably mainstream adoption, like with Cash app, with Robinhood and I think that's going to continue. That's going to move much further ahead than Lightning will and I think that there's always going to be... Lightning's never going to be finished, so it's always going to be getting improved and it might be a few years before I would say Lightning is ready for mainstream, of someone just downloading an app on their phone and using it. But that's okay because Bitcoin itself is also not really mainstream, even though the news talks about it all the time, but it's not like all 300 million Americans are using Bitcoin.

Peter McCormack: Right. Well I appreciate both of your time as ever. It's very useful for me to put some of my dumb questions to you. So thank you both for coming on again.

Pierre Rochard: Thanks for having me on.

Nic Carter: Thanks so much, Peter.