Replying to Avatar Lyn Alden

“We should change Bitcoin now in a contentious way to fix the security budget” is basically the same tinkering mentality that central bankers have.

It begins with an overconfident assumption that they know fees won’t be sufficient in the future and that a certain “fix” is going to generate more fees. But some “fixes” could even backfire and create less fees, or introduce bugs, or damage the incentive structure.

The Bitcoin fee market a couple decades out will primarily be a function of adoption or lack thereof. In a world of eight billion people, only a couple hundred million can do an on chain transaction per year, or a bit more with maximal batching. The number of people who could do a monthly transaction is 1/12th of that number. In order to be concerned that bitcoin fees will be too low to prevent censorship in the future, we have to start with the assumption that not many people use bitcoin decades out.

Fedwire has about 100x the gross volume that Bitcoin currently does, with a similar number of transactions. What will Bitcoin’s fee market be if volumes go up 5x or 10x, let alone 50x or 100x? Who wants to raise their hand with a confident model of what bitcoin volumes will be in 2040?

What will someone pay to send a ten million dollar equivalent on chain settlement internationally? $100 in fees per million dollar settlement transaction would be .01%. $300 to get it in a quicker block would be 0.03%. That type of environment can generate tens of billions of dollars of fees annually. The fees that people pay to ship millions of dollars of gold long distances, or to perform a real estate transaction worth millions of dollars, are extremely high. Even if bitcoin is a fraction of that, it would be high by today’s standards. And in a world of billions of people, if nobody wants to pay $100 to send a million dollar settlement bearer asset transaction, then that’s a world where not many people use bitcoin period.

In some months the “security budget” concern trends. In other months, the “fees will be so high that only rich people can transact on chain” concern trends. These are so wildly contradictory and the fact that both are common concerns shows how little we know about the long term future.

I don’t think the fee market can be fixed by gimmicks. Either the network is desirable to use in a couple decades or it’s not. If 3 or 4 decades into bitcoin’s life it can’t generate significant settlement volumes, and gets easily censored due to low fees, then it’s just not a very desirable network at that point for one reason or another.

Some soft forks like covenants can be thoughtfully considered for scaling and fee density, and it’s good for smart developers to always be thinking about low risk improvements to the network that the node network and miners might have a high consensus positive view toward over time. But trying to rush VC-backed softforks, and using security budget FUD to push them, is pretty disingenuous imo.

Anyway, good morning.

To be fair, we do absolutely need to do fee smoothing to reduce spikes.

Reply to this note

Please Login to reply.

Discussion

Do you mean some kind of futures market for fees?

Why and how?🤔

Maybe spikes are useful information that helps regulate user behavior.

How spikey or smooth will transaction fees be in 15 years, and what should we do about it today with any reasonable amount of conviction?

Yea I mean I don’t think they’ll be *that* spikey, but having the ability to eat a single huge-fee transaction without miners trying to reorg each other would be nice. I think there’s some designs with a “pot” of fee money that miners can contribute to or take from in exchange for bigger/smaller/easier/harder blocks that can eat such a spike and smooth it out a bit. Would also reduce weekly cycle impact on miner revenue.

Notably, such a design would still allow market forces to increase block rate/size during fee/block space demand increases, but even without miners turning off, which is cool.

I read that study a while back that discussed how incentives could become unstable in a fee-only scenario, with extra block re-orgs. There have probably been more studies since then. My analysis and research at the time pointed to that there seem to be various fee-smoothing mechanisms that are somewhat straightforward and could be implemented, but it also seems like not an existential risk and something that could be implemented if it starts being a problem.

One of my opinions of the paper I read was that they didn’t sufficiently take into account economic incentives. Miners put serious capex into ASICS with long lifespans, and so even if block-by-block incentives can incentivize more re-orgs, the long term revenue comes from the reliability of the network. So it’s unclear that miners would cause enough block re-orgs to make the network stability shitty. But if that’s how it works out, there are fixes as you pointed out. Seems like something that can be done if it actually materializes as a problem one day, rather than something that needs to be done years or decades in advance of there *maybe* being a problem.

To be fair, we do absolutely need to improve on-chain privacy.

You mean Lightning?

Adopt a low time preference. The end.