Maybe it would be helpful to consider what it means to have a crypto wallet (because that's the likely endpoint here, otherwise why would they have to partner with Stellar?) on production servers. Consider explaining to a client that, no, it's totally fine, it doesn't do anything. I promise. Doable, I'll grant you. Consider explaining that to an auditor. No, nobody can ever compromise this, nobody could use this for shenanigans, it just doesn't do anything!

An auditor who did not bug his or her eyes out at that and ask some very hard questions would be doing a bad job as an auditor.

It is harsh--but, one, I thought they were legitimately better than this, and two, it's a problem. Keybase wants to be infrastructure, and people have different needs for infrastructure. That's an argument for different tools, not a multi-tool that makes its use in some situations problematic.

> Maybe it would be helpful to consider what it means to have a crypto wallet ... on production servers.

Why would anyone install the Keybase app on production servers?

Given that the protocol is open and the app code is available, I imagine there will be plenty of cli clients that just do not implement the currency stuff. Or at least I hope so, given that I share the above concern.

If this is really a huge deal for you, it sounds like you're just not the target audience. I'm a keybase user and I think this is really cool.

Yeah, main thing is that part can be easily taken out for those who do not want it. Should not be an issue at all imho.

If it were aboveboard, it would be opt-in, not opt-out, yeah?

Keybase started out with an identity system because it was the key foundational module to create their entire e2e encrypted software system.

After the user system, they created a network filesystem, which their git system is based upon, and I'm guessing it's also something their team chat system uses too. The entire groups system is also based on their user system.

It looked like they were building out a dependency tree bit by bit.