[xur17]

> I associate crypto with shonk, which puts keybase.io in the "this is a crypto thing? Hmmm.... shonky." category.

I'm sorry to hear that. There is a lot of scammy behavior in the space, but there is also a lot of interesting work going on. It just turns out that the shitty people speak the loudest, so they're the ones you hear about, rather than all of the people heads down working hard on interesting problems.

I don't think it is intellectually honest to dismiss the entire space as a scam due to the bad actors.

[api]

The defense is to point out some genuinely useful work going on in the space that doesn't revolve around speculation.

[xur17]

* Bitcoin's lightning network

* Stable coins (Dai, USDC, etc)

* What Keybase just built (much easier way to manage keys for wallets)

* Makerdao's decentralized loan system

The first 2 enable you to transfer value over the internet in a cheap and non-revokable way (internet cash). The third one simplifies wallet management for the first two. The last one opens up a bunch of new opportunities for borrowing money without having to trust your counterparty. All of this enables people to transfer value without any middlemen.

And separate from this, projects like IPFS / Filecoin (if they actually work) would let you store data without using a centralized third party.

[drawnwren]

I agree that crypto could be valuable.

But, (1) the lightning network has been repeatedly attacked as unworkable trash and (2) there needs to be a reason to make crypto transactions other than drugs or vaguely crypto-anarchist philosophy.

(b) Tether's recent scandal has made all stable coins fairly suspect.

(d) The makerdao system has some pretty serious problems. High liquidation rates and skyrocketing fees both come up.

If anything, crypto has taught us (through the plethora of 2018/19 crypto scams) that revokability is a feature of transactions that is _valuable_.

I'd say my biggest takeaways from crypto, currently, are that there is not such thing as 'without having to trust the other party' and that all of the middlemen in the current financial system do provide more value than just taking a cut (while that value is probably not proportional to the cut that they take).

All that said, I'm still optimistic on crypto. However, I think it's unrealistic to expect it to replace a significant part of the web in its current form.

[vermilingua]

As to point (2), I think this keybase/stellar integration is the solution to the problem that crypto has been looking for.

If I want to give a friend money, I can hand them cash, but I have to be in the same room as them. I can transfer that cash in other ways, but I need to trust the middlemen. I can also use a bank to send them money, but I have to know their details, which are sensitive information. I can use one of the plethora of money-transfer apps, but that involves putting my money under someone else’s direct control (someone who may go insolvent by the end of the year).

Or, if I have them on keybase (or any number of platforms, ideally) I can now just send it with a simple text message.

Additionally, what if I really appreciate something you said on HN, or twitter; or you made a really valuable contribution to one of my projects on GH? If you have your keybase proof on there, I can send you my appreciation in a matter of seconds. Imagine trying to do that using cash, PayPal, etc.

Despite the fact that until now, crypto has had far more overhead than cash and credit transactions, this and successor programs may have the potential to make transfers far more convenient.

[shakna]

> I can also use a bank to send them money, but I have to know their details, which are sensitive information.

PayID [0] was standardised in Australia a while back, standardising what the banks were already doing. At least here, knowing someone's details is something from the past.

A PayID can _only_ receive.

A PayID is a number - not necessarily a phone number you own, but many people opted for that route, but keep the same number even if their phone number does change.

A PayID does reveal the name of the account holder.

Transfer happens as part of Instant Payments, so at it's longest it'll take 60 seconds to arrive in an account, but usually within 2.

PayID is also integrated into Apply Pay and Google's offering.

... Is this still an unsolved thing overseas? I've had instant transfers to friends without knowing their banking details for nearly a decade.

[0] https://payid.com.au/

[vermilingua]

I know, and I use PayID frequently; but it has shortcomings, mostly that nobody seems to know about it. Additionally, PayID only works within Australia, and effectively only to people you know personally (mobile numbers are not typically public information).

I love PayID for what it is, but it could be well improved in my first example, and doesn’t work at all for the latter.

[RustyRussell]

> (1) the lightning network has been repeatedly attacked as unworkable trash

[ Disclosure: I work on the lightning specification and one of the implementations ]

I'm sure it has been attacked (because "cryptos, bro"), I'm disappointed that it made you dismiss it.

FWIW, I switched from Linux Kernel development to developing Lightning after almost 20 years: I find it ambitious, high-potential and fascinating, as well as challenging.

[drawnwren]

I don't mean to attack you. It's just that there is a large gap between well-engineered and easy to use.

I am 100% certain you're a better developer than I, and that lightning is not a result of poor engineering. But if a layman can't pick it up and use it in under 5 minutes, it's unworkable (if you're targeting laymen).

Crypto is full of genius engineers, and product folks, but I think there is (and always has been) a usability crisis.

Thanks for your work on both Linux and Lightning. I look forward to the day you can point to this comment on HN as being ridiculously dated.

[arshbot]

> I am 100% certain you're a better developer than I, and that lightning is not a result of poor engineering. But if a layman can't pick it up and use it in under 5 minutes, it's unworkable (if you're targeting laymen).

In this scenario linux is unworkable since a cli is probably too much for a layman. But that's why people built gnome and other window managers on top of the "unworkable" but useful core.

All we have to do is wait for some people to build that nice pretty layer on top - but the core must be an efficient tool

[uncletammy]

Sorry but calling Bitcoin's Lightning network "internet cash" is incredibly misleading. Bitcoin is "internet cash". At least it used to be.

Bitcoin's version of the lightning network is a Rube Goldberg machine bolted onto a payment network and dressed up to look like a scaling solution.

Lightning network is a super cool technology but BTC's implementation of it suffers from crippling flaws that make it useless for Lightning's most promising use cases. Lightning network was designed to be a micropayment solution built on top of a low-fee base layer. It requires that low-fee base layer to deliver on it's promises. Since it's been built on a high-fee base layer (BTC), transaction routing is now a problem that can only be solved by sending your lightning payment through a large and highly connected lightning node and never closing your payment channels (because re-opening them costs too much). This effectively eliminates the p2p aspect that is required for "lightning bitcoin" to function as cash.

Instead, BTC lightning network has to function more like traditional banks where the node operator is the bank and the payment channel is the user's account. The transactions are like really fast SWIFT/ACH transfers except they only go through if the sender's account meets the necessary liquidity constraints. Those liquidity constraints actually make btc lightning transactions more like purchases made using a secured/prepaid credit card because you must have money tied up so that you can spend money. And like a bank, the lightning node gets to decide if they want to route your transaction to it's final destination.

It's also incredibly likely ( and arguably confirmed based on https://www.fincen.gov/sites/default/files/2019-05/FinCEN%20... ) that BTC lightning nodes will be subject to the same rules and regulations as a traditional bank. If so, it's yet another centralizing force on BTC's version of Lightning. Don't get me wrong, I'm not knocking banks here. Banks are useful and KYC/AML are often good things. I'm just saying, "lightning btc" couldn't be further from cash.

[ChainOfFools]

I would add to that some explanation as to why the proposed decentralization of custody, control and leverage inherent in any cryptocurrency proposition should be expected to remain stable indefinitely rather than collapse into a conventional, de facto centralized model under the inexorable pressure of competition to derive profit from such a system

[michaelsbradley]

How about pure development work to build tools that can be used to experiment with and build atop decentralized technologies such as IPFS and Ethereum blockchain, etc.?

https://embark.status.im/

https://github.com/embark-framework/embark

[k__]

IPFS seems to solve an actual problem without too much overhead tho...

[michaelsbradley]

See: https://github.com/MetaMask/metamask-extension/pull/4405.

[ericflo]

I have a bunch of data backed up on the Sia file storage marketplace. Works fine for that purpose for me today.

[fourstar]

You just described literally anything new.