inamerrata

At the start of the year, I wrote out some thoughts about Bitcoin priorities, probably most simply summed up as:

it’s probably more important to work on things that reinforce [Bitcoin’s] existing foundations, than neat new ideas to change them

In that post, I also wrote:

I’m particularly optimistic about an as yet unannounced approach that DCI has been exploring, which (if I’ve understood correctly) aims to provide long term funding for a moderate sized team of senior devs and researchers to focus on keeping Bitcoin stable and secure […]

It wasn’t something I’d even considered as a possibility at the time, but the world works in mysterious ways, and as it turns out, I’m now joining the Digital Currency Initiative to work on making that approach live up to its promise.

There are, I think, two ways to make systemic improvements in security. One is in code and tooling improvements — reworking the code directly to remove bugs and make it more robust, and building tools like linters, continuous integration and fuzz testers, that will then automatically eliminate potential bugs before they’re written or merged. I expect that will be where we’ll devote most of the effort.

But I think an equally important part of doing security well is having it be an integral part of development, not an add-on — while certainly some people will have more expertise than others, you want everyone thinking about security; in a similar way to wanting everyone to be thinking about performance if you want your system to work efficiently, or wanting everyone to be thinking about user experience if you want a smooth and consistent experience. That is, the other part of making systemic improvements in security is maintaining a culture that deems security a critical priority, and worth thinking about at all levels.

That may mean that I want to walk back my earlier conclusion that “neat new ideas [that] change [Bitcoin’s existing foundations]” are something to deprioritise. Because it certainly seems like people do want exciting new features, and given that, it quickly becomes super important that the people working on those features aren’t a separate group from the people who are deeply security-conscious, if we want to ensure those new features don’t end up compromising Bitcoin’s foundations. The alternative is to continually fight a rearguard action to either debug or prevent adoption of each neat new idea that hasn’t been developed with an appropriately adversarial mindset.

In particular, that may mean that working on things like ANYPREVOUT and TAPLEAFUPDATEVERIFY might have two ways of fitting into the “improve Bitcoin’s security” framework: it makes it easier to use bitcoin securely (ANYPREVOUT hopefully makes lightning safer by enabling eltoo and thus reduces the risks of toxic state; TAPLEAFUPDATEVERIFY may make improvements in cold storage possible, making on-chain funds safer from theft), but developing them in a way that puts security as a core goal (as compared to other priorities, eg “time to market”) might help establish traditions that improve security more broadly too.

(And I don’t mean to criticise the way things are going in Bitcoin core so far — it’s a great project where security does take a front row seat pretty much all the time. The question I’m thinking about is how to make sure things stay that way as we scale up)

Also, just to get it on the record: “security” means, in some sense, “the system works the way it’s intended to”, at least in regard to who can access/control what; but “who is intended to have what level of access/control” is a question you need to answer first. For me, Bitcoin’s fundamentals are that it’s decentralised, and that it’s a store of value that you, personally, can keep secure and choose to transfer if and when you please — which is really just another way of saying that it’s “peer-to-peer electronic cash”.

I don’t think Bitcoin gets anywhere by compromising on decentralisation: better to leave that to competing moneys whether that be Central Bank issued or altcoin tokens on the one hand, and higher layers that build on Bitcoin, like Liquid or exchanges, on the other. If those things succeed, that’s great — but having a money that’s an even playing field for everyone, powerful or not, is a fundamentally different thing that’s worth trying to make work.

There are plenty of details that go into that, and plenty of other things that are also important (for instance, I think you could also argue that many of Bitcoin’s other priorities, such as the fixed supply, or privacy or censorship resistance can only be obtained by having a decentralised system); but I think it’s worth trying to pick the principles you’re going to stand for early, and for Bitcoin, I think the best place to start is decentralisation.