Thread by @mcclure111

Not good: This is a cryptocurrency, but they don't immediately disclose that

Good: It is apparently not a proof of work cryptocurrency, so it doesn't destroy the earth

Not good: Uses Intel SGX, which is a DRM tech and should therefore IMO be illegal

https://twitter.com/signalapp/status/1379472676905963523

If I'm reading this cryptocoin's documentation right, part of how their algorithm works is they literally gave Intel a hardware monopoly on servers operating the coin's network, via their use of SGX on the servers

I'm struggling to understand what the blockchain resolution method is. It appears to be byzantine agreement and they don't say they're using DBA, which to me means it's centralized (IE, it would mean every server has to be whitelisted by the mobilecoin company?)

Endlessly frustrating how even those very rare "blockchain" (*spits*) connected companies who appear to have a partially real product and not something 100% fake are incredibly obfuscatory about what their product is and does and how it works

With PayPal, if I send someone one dollar, I understand how that happened. My credit card company has a contractual relationship with PayPal who also has a contractual relationship with the recipient's bank.

With PayPal, I understand how these entities talk to each other and who they are allowed to share information with, and when a government search warrant might lead to disclosures to a government. I might not like any of this, or any of these companies, but I can comprehend it

On the other hand if I try to understand how "mobilecoin" used in Signal works, the answer is: a bunch of math, but not real math, just vague terms they define that interact with each other in other ways they define. They snow you with buzzwords.

I'm not sure I could understand the underlying mobilecoin math, but they don't explain the math, they just make a bunch of rapid-fire assurances about anonymity. The user-facing explanation linked by Signal makes constant references to "crates", a Rust specialist term

Also endlessly frustrating that groups appearing to create good-faith open-source public-benefit software do weird pivots like this where suddenly the app you've made yourself dependent on does something you didn't previously intend it to.

https://twitter.com/0x76_/status/1379505242342035456

There's a very good open source password manager called "Bitwarden". Last month, out of the blue, they added a secure messenger feature. That's incredibly alarming to me. I don't want my password manager transmitting information to anyone else. That is the opposite of its purpose

The literal point of a password manager is to make sure that the information in the password manager program *doesn't* get transmitted to anyone except me

(Getting lots of responses from Bitwarden users who didn't notice the messenger feature. It took me a bit to notice it also; it's an unlabeled button. This raises a question. *What good is a messenger feature your users don't know is there??*)

https://twitter.com/excogs/status/1379508523520905224

Anyway from an academic standpoint I'm still curious how mobilecoin reconciles these two sentences in their README. If you don't have a central authority then what set of entities are you reaching byzantine consensus between ?!? github.com/mobilecoinfoun…

(Overall their architecture seems similar in a lot of ways to the one used by Libra/Diem, Facebook's cryptocurrency, but Libra's protocol is a 30-page PDF linking two other 40-ish page PDFs. There are elements left unspecified, but there was at least *some* detail there.)

(By contrast, mobilecoin is several vague markdown files spread across 4-5 github repos. If there is a straightforward document simply explaining their algorithm with citations for background material I'm not finding it. How many words of this screenshot do you understand)

Basically this. If you look at the responses to that Signal tweet up top, it's just nothing but a flood of responses from crypto fans angry that Signal dared to base their payment system on something other than global warming dollars

https://twitter.com/olmaverick/status/1379512107150888963

[Thread descends into self-referential cryptocoin nonsense from here]
So this answers my question from above— they're not using "BFT" (conventional byzantine agreement), they're using "DBA" (distributed byzantine agreement, aka Stellar if I'm not confused)

https://twitter.com/tilde_cs/status/1379511798488051712

Why am I focusing on these weird acronyms? Well, because I actually believe DBA— the Stellar thingy— is the way a cryptocurrency could be "good", to the limited extent it is possible for money to be good. DBA, if it works, is the way to defeat the "rich get richer" problem.

What I mean by that: Every cryptocurrency in Bitcoin's intellectual lineage gas a situation where the people who buy in to a given coin early will endlessly gain benefits from further adoption of Bitcoin. If you invest in Bitcoin miners, you get the mined coins, you get the tips.

That's "proof of work"— whoever buys the most video cards gets the inherent profit from the Bitcoin network. Proof of work also has this unfortunate side effect where it causes global warming and therefore destroys all life on earth, or at least, all human life. Suboptimal.

So cryptocoins like Ethereum are moving to something called "proof of stake" where control of the chain— and therefore the benefits, mined coins, "tips"/gas fees etc— are [details vary] based on owning Ethereum coins. This sidesteps the global warming problem.

*But...*

A weird ideological issue comes in at this point.
"Proof of Stake" coins, in some sense, gain a sort of interest over time. If you have a lot of POS coins, like ETH wants to be, you can sorta collect taxes from lesser users. (I am not fully certain ETH2 works precisely this way.)

So there's no "people will literally die" urgency to *stopping* people with big ETH investments from collecting rent (in the form of gas fees) from everyone else. But like. If we were designing a society or economic system from scratch. Maybe that's not how we'd prefer to do it.

Fortunately, there's another solution! There's the unfortunately named "Byzantine" approaches. In these, the network members essentially hold a vote on the next blockchain head. No weird games, no hashbreaking, no lottery. Straightforward. Uses well-established database tech.

But! The Byzantine approaches have a problem: How do you decide who gets a vote? In federated BFT, like Facebook/Libra/"Diem", there's just an approved list. Meaning it's not actually decentralized. And the whole point of this absurd "blockchain" mess was to decentralize, right?

(In fact, the mere centralization of the Facebook/Libra/"Diem" approach creates a rich gets richer problem: Libra doesn't have "mining", it mints coins by literally investing all money stored in the system in bonds. Good plan IMO. But guess who keeps the interest? Facebook.)

So then there's DBA/Stellar, which mobilecoin is using. It avoids centralization by using really complex graph theory.
New problem: This creates a risk the blockchain might "fork", split into two.
One solution to this is to re-centralize, and just… lie about it, potentially.

My friend named @keisisqrl asserts Stellar is using "trusted nodes"-- if the Stellar network ever broke in two, those trusted nodes would serve as a tiebreaker.
I'm not sure she's right. Stellar itself claims she's wrong.
But I can't *prove* she's wrong?

https://twitter.com/keisisqrl/status/1379514053383520257

I make a sort of mental argument to myself: There's *some* entry point to the Stellar network, somewhere. You download an app, or sign up on a wallet site. If the Stellar network ever split in two, whoever had the most popular app or wallet site, that quorum slice would "win".

...because when new people joined the network, they'd join that wallet/app's quorum slice. I guess.

So I try to hunt down what determines your entry point to the Stellar network. And I find…documents about "Stellar", "Horizon" and "Stellar Core", vaguely referencing each other.

It's hard to find out what's actually happening. I can "Understand" Paypal, or even Signal. But Understanding Stellar, or Mobilecoin— I'm basically internalizing someone's entire EC2 network architecture. I can't grasp it enough to disprove a claim like "there are trusted nodes".

I don't think a centralized system would be *bad*. I'd be happy just for a Paypal that isn't a transphobic monopoly. "Paypal, but it's cryptographically federated like domain name registration" sounds nice.

But if they *say* it's decentralized they should be telling the truth!

Anyway, I'm not sure anything I just said matters. I apologize for writing a thread about cryptocurrency.

Here's the TLDR:

1. The ideologically & ecologically best way to run a distributed-ledger payment processor is DBA/Stellar
2. That's what mobilecoin/the Signal trial program uses
3. But I'm not sure Stellar even works
4. And mobilecoin seems incapable of clearly explaining how their system works

Galaxy brain take: This would be a socially beneficial feature for a password manager app to offer, because it would incentivize people to stop using those ten passwords

https://twitter.com/seneschaaal/status/1379518415874912256

Okay so mobilecoin somehow found this thread and politely linked me the exact thing I was looking for, a technical PDF with background citations, explaining the entire system. And it's…140 pages. So I guess I'll give you my opinion on it in 3 weeks lolsob

https://twitter.com/mobilecoin/status/1379527021915594753