Out-of-band transactions are probably inevitable

The latest internal controversy with Bitcoin surrounds the upcoming release of Bitcoin Core, v30, which will by default, relay transactions containing large amounts of data as OP_RETURN.

For the non-experts: There's an option to put small amounts of data inside a Bitcoin transaction, using something called OP_RETURN. For example, see this transaction here, in which somebody defied a presidential executive order. There's no protocol limit, but transactions are relayed by a network of nodes. The software on these nodes comes with a default policy that won't relay the transaction if you try to include over 80 bytes of data. The reason for this is that people don't want stupid crap on the blockchain. But if you really wanted to submit a transaction with more than 80 bytes, you had to route your transaction around the network, for example, you could send it directly to the miner who could then mine it.

The stated reason for the change is, in a nutshell, "People will do this anyways, and when they do this outside the gossip network, it encourages people to establish patterns of ways to route transactions around the network, which eventually leads to centralization."

I'm not sure I agree with this: People can do this, but it's a different thing if it's super easy (and cheap.) For example, I recently decided I want to write some silly stuff on the blockchain. I run a node, and it wasn't that hard to figure out how to construct a transaction with OP_RETURN data using a few commands and then submit that to the network. I had sent myself $1.04 and ended up creating 4 transactions with this. Very cheap and easy. I wanted to mess around with a larger payload of data. My node wouldn't let me relay that. There are other places to submit such a transaction, but the ones I found were asking a minimum 4 sat/vbyte which would have gone over my $1.04 budget, so I stopped and got back to my day job.

What I think is going on now is that there is a slowly growing panic around the complete absence of fee-paying transactions. The new rules may invite a rush of transaction spam that might temporarily raise the fee rates to make the fee market seem "robust."

I do think that when v30 gains some adoption there will be lots of folks who are willing to spend a couple minutes and $10 (I might go for it myself) to upload much longer pieces of data to the blockchain. It's fun. It feels cool. This is going to lead to full blocks and higher transaction fees, and Bitcoin can ignore the security funding crisis for a little while longer.

So I disagree that "people will do this anyways" based on the observation that nobody is doing it right now.

The problem with the argument is that it concedes that there is very little hope to organically encourage people to use the network - the goal is just to hope the waves of spam never stop rolling in. Implicit is the understanding that large data payload transaction will come to play an important part in Bitcoin's security model.

The fear seems to be that someone with a real data-on-the-blockchain business model which requires higher payloads might come along and realize that they can't just hook up their node, instead they have to partner with a miner. This means miners will make this part of their service, giving them an advantage, and this leads to centralization. I'm not sure I agree with this either: Almost all mining is done by a handful of miners and pools. (It's silly to solo-mine, these days it is a noteworthy event in Bitcoin if a solo miner gets a block. ) Most of these pools have some sort of public face. It should be relatively trivial to offer a portal for submitting these transactions. If you're requiring a minimum fee, that's where the economic practice of undercutting comes in. Quickly we get to a place where all of the major pools are offering a place to submit these transactions.

However, there are actually very good reasons to worry about out-of-band transactions.

To be clear, there's two types of out-of-band events.

1. Out-of-band submission. This refers to a transaction being submitted outside the usual node-to-node gossip, outside the usual idea of mempool. If submitted only to certain miners, this gives the miner a strictly larger set of transactions to construct a block with, giving them an advantage over other miners. If other miners hear about the transaction, they can include it.
2. Out-of-band fees. This is more interesting: I can submit a transaction via any method, via the mempool, or directly to miners, however, I arrange to pay miner directly without using the built-in fee mechanism. For example, I could send a very little transaction with tiny fee. Most miners will not touch this, but I can pay a specific miner via credit card or lightning or something else to include the transaction.

Obviously neither of the above are illegal or against protocol rules, although they do seem somewhat against the spirit of Bitcoin. The security model has assumed that miners are always in competition to mine a full set of pending transactions.

You don't have to be incredibly creative to think of reasons why it might be preferable to make out-of-band payments to miners. The most obvious is that it gives both miners and transactors a way to hedge against hashprice volatility.

For example, suppose that I decide to start a business offering Bluesky users the opportunity to write the text of any skeet into the Bitcoin blockchain, using the OP_RETURN. Clearly the best way to market this is for a flat easy fee, say $15.99. Click the skeet, submit the credit card payment, and go tell your comrades that you just etched a Mark Hamill banger into the Bitcoin blockchain FOREVER.

It's easy the imagine the arrangement I would seek out: I would want to work out to pay them, say, $5000 today for the right to 1000 of these types of transactions. We may have done the math to get that today it only costs me $3.50 per transaction, but we both want to lock in a fee of $5. If the miner has more than a few percent of the total hashrate, I can be almost certain they will find a block every day, and anything I give them will be in it. Now that my fees are locked in I can focus on marketing.

If Bitcoin is to develop a robust fee market, necessarily there will repeat customers. Repeat costumers are going to want to lock in costs. Miners are going to want to lock in revenue. Yes, there will be derivatives markets on electricity, fee rates, hashrate, Bitcoin price, and all these things, but why not cut out the derivatives brokers and go straight to the miner?

In a couple halvings when transaction fees make up a larger portion of the revenue, things could get more interesting. By negotiating fees directly, larger miners can exercise much more control over the market, undercutting and forcing others out. In this case miners will have an incentive to aggressively market themselves to customers, and the advantage of centralization will become massive. Branding will be everything. (At risk of stating the obvious; Rob Gronkowski will absolutely be involved in the branding.)

Publicly traded mining companies, though partnerships with AI//HPC and involvement in energy markets, are beginning to have their fingers in more pies (or are becoming the pies getting fingered). If a partner has the ability to not submit a transaction to the network but have it mined directly, they will opt for that. (It's still unclear to me longterm what the time concerns on a transaction will be, if we go with Saylor's notion that they're like real estate payments, waiting days to "settle" won't be a dealbreaker. It's not like major corporations will be trying to double-spend on each other.) This can allow miners who may be struggling to mine profitably block to block to mine intermittently, only when they have enough transactions built up to "profitable."

Long term, it Bitcoin is to be successful, folks very broadly need to organically use L1 to transact, and then as fees start to grow, organically move to L2s and this has to happen on a large scale. Hoping for a short term dataspam bailout may make it appear possible to kick the can down the road a while, but this is not a longterm solution.