This story is part of BREAKER’s Social Good Week, a series looking at ways blockchain technology can engineer progress and help humanity.

Let’s say, hypothetically, that you had $1,000 dollars worth of bitcoin, and you wanted to give it away as charity. How do you do this? Where do you send the money?

This is trickier than it sounds. Your average homeless person is not walking around with a Ledger Nano wallet. Refugees are not hodlers. The needier the recipient, perhaps the slimmer the odds that they’re crypto-savvy.

For Joe Waltman, a San Francisco-based start-up entrepreneur, this question was not theoretical. He was given $1,000 in crypto, along with a challenge: Give this to people in need, and do it as efficiently as possible.

So Waltman conducted a few experiments. In one, he posted in a Venezuelan sub-reddit, and said he would give $10 in bitcoin to anyone who could prove that they’re Venezuelan, explain why they needed the money and what they would do with it (like buy food or clothes), and then take a photo of what they bought. This struck people as confusing and weird. “We shut that down pretty quickly,” he admits. Next idea: He tried to give it to Venezuelan restaurants who were already accepting crypto. “I basically told them, ‘I’m some dumb American. I want to send you $100 bucks, but I want you to turn it into as much food as you can.’” This time he had better luck. Waltman cold-called one restaurant on a Thursday, zapped them crypto on a Friday, and by Saturday, 300 people got a meal.

Related: When Hyperinflation Means You Can’t Buy Bread or Tylenol, Crypto Beckons

Then there’s the Venezuelan brewery. It turns out that in addition to beer, a brewery creates byproducts like hops, grains, and barley. Much of this stuff is nutritious, and each day it’s tossed in the garbage. The brewery told Waltman that if he gave them $100 in crypto, they could transform these leftovers into 10,000 snack bars—10,000!—and give them to hungry kids at a local school. “This was a no-brainer,” says Waltman. “The bang for the buck is incredible. These kids are probably going to bed hungry. For $100, you can feed around 600 kids for a month. I told them that we’ll give you $100 for the rest of our existence.”

The $1,000-crypto experiment was something of a proof of concept for GiveCrypto, the charity launched by Brian Armstrong, CEO of Coinbase. It was also, in a sense, a job interview. Armstrong had raised $4 million in donations, and he wanted to treat the organization like a start-up. So he hired a tech guy, Waltman, to figure out how to spend it.

Waltman was new to the space. “I didn’t have a lot of experience dealing with people in terrible situations,” he says. “I’m still learning how to maintain a safe emotional distance.” At the end of the $1,000 experiment, Waltman says he was so moved by the stories—kids rummaging through trash cans—that he cried in Armstrong’s office. As Waltman remembers it, the Coinbase CEO told him, “Hey, I don’t normally deal with people crying in work meetings, so this is a new thing for me, too.”

Since then, GiveCrypto has conducted gobs of experiments in crypto-charity—often in Venezuela—which Waltman meticulously chronicles each month with a mix of candor and humility. “GiveCrypto is a charity that gives cryptocurrency to people in need,” he writes, noting that “I have no charity or crypto experience.” A sample post from October shows a spreadsheet of locations (Uganda, Lebanon, Brazil, Argentina, Venezuela), amount dispersed ($13,926), and the number of people impacted (854).

After nearly half-a-year under his belt (that’s nearly a decade, in blockchain years) Waltman opens up to BREAKER about the nuts-and-bolts challenges of giving bitcoin, refugee camps in Bangladesh, and how crypto-charity can (maybe) disrupt traditional philanthropy. This interview has been condensed and edited.

Why give crypto, and not just cash?
There’s a philosophical battle going on with charity. And not to insult your intelligence, as you probably know about this…

Pretend I don’t! [Note: I don’t.]
There’s a belief that most traditional charities are somewhat paternalistic, and very prescriptive. They’ll give you a goat or they’ll build you a school or they’ll deliver clean water. These are all great things. But more recently, there’s been an increasing amount of evidence suggesting that, no no no, don’t pretend that you know what these people need. They know what they need. Just give them money. And they’ll buy it.

Not only is that a lot more efficient, but it gives some agency to the recipient. They might not need a goat. They might need something else. So we’re certainly disciples of that philosophy—of direct cash transfers.

But you’re not giving cash…
We’re giving crypto, and that’s super freaking weird right now. In almost every single place on the world, it would certainly be more helpful to give cash.

So why crypto?
We’re thinking a little bit farther down the road. As the UX challenges get solved, as smartphone adoption increases, and blah, blah, blah, we see a scenario—I don’t know if it’s three years, five years, or 10 years from now—where it’s far less difficult to transfer, to hold, and to use crypto. So we want to build the infrastructure that’s ready for the hopefully not-too-distant future.

Got it. So what are you doing now, in the short term?
We’re focusing on places where cryptocurrencies have the highest likelihood of being used, and are maybe even more useful than fiat. Our hypothesis—and we’re very much still testing this—is that those places are either where money is broken, like in Venezuela, or where government is broken. There are many examples of broken government, but we’re most interested in the refugee situation.

Are there really examples of where crypto is more useful than fiat?
We’re connected with this charity called FreeFrom, which helps the survivors of domestic abuse. There’s some crazy number that 90 percent of the time, the victims go back to their abusers for economic reasons. FreeFrom helps the victims set up some entrepreneurial activity, so they have economic independence once they’ve moved away from their abusers and found a new place to live. But they’re also contacted by victims still living with their abusers, and they can’t really help, because if the abusers find out, then that might put the victim in danger. So they’re careful about minimizing contact, and they haven’t been able to help in these “active” situations.

Do you mean that, for example, it’d be risky to send a physical check to the victim, because it might be seen by the abuser, who lives in the same house?
Exactly. If you open a bank account, stupid fricking Bank of America will send you a bunch of marketing materials. You [the victim] can’t set up a separate bank account, because they’ll ask, ‘What’s the home address?’ And then the bank is going to send you a bunch of shit, which is probably going to raise the eyebrows of the abuser. I didn’t come up with this idea, but we said, ‘Oh my gosh, we can actually get crypto into the hands of these people, in a way that the abuser wouldn’t be able to find out.’

If we set up a crypto wallet, we can give these people a non-trivial amount of money. We’re giving nine different active victims of domestic abuse $2,500 worth of bitcoin. Which is not a huge amount of money, but enough for somebody to make a deposit on a new apartment—to get out of the situation.

Amazing. It seems like a lot of your work is in Venezuela?
We’ve done about 10 experiments in Venezuela. From a broken-money perspective, nowhere is even close to Venezuela. Their 12-month inflation is like 500,000 percent. Country number two is like 80 percent. If you’re really thinking about inflation, there’s no reason to even think about anywhere else. (Iran, Congo, and Argentina are creeping up there, but they’re four or five orders of magnitude from Venezuela.)

What have you done there?
It’s all over the map. In one case, we put up a sign in a butcher shop that happens to accept crypto, and said, “Download this wallet, and take a picture of yourself in front of the sign, and upload a picture to Instagram with the right tags, and we’ll send you 10 bucks.”

In another, we found a church and a minister just over the border from Brazil. The minister would help us figure out who needs the help, and get them set up with the appropriate [crypto] wallet. We would then transfer some funds to the minister, and he would help distribute. Then he went one step further, and said, “I’ll set up a store inside my church, and I’ll accept crypto as a payment, so these people can buy food with the crypto.”

This is some minister.
He’s amazing. He did the targeting, he helped with the redemption, and helped them cash out. And we think there’s something with that model—this ambassador concept. Where we have an informal, distributed group who wants to help, and we’re just giving them the resources.

How does this ambassador model work?
They can identify recipients who need help. They can help recruit vendors who will accept crypto. Maybe they can either serve as a cash-out option, or recruit other sorts of businesses to be cash-out options. I’m sure I’m being naive and overly optimistic, but I think there’s an interesting opportunity to have this distributed network of ambassadors, auditors, and vendors for cash-out partners.

Do they do this for free?
Of course they’re all making a little bit of money—we’re not asking for their time for free, but I’m convinced this will be far more efficient than your traditional charity.

You mentioned that you’re also focused on areas where “government is broken.” What’s an example?
The Rohingya refugee camp in Bangladesh. I had three or four false starts with potential partners, trying to find some way to get inside the camp. I finally found a guy about a month ago, who was willing to sneak into the camp. (It’s much less James Bond than it sounds.) He went into the camp, recruited 21 widows, and then found a vendor—inside the camp—who will accept payments in crypto.

And you send them crypto directly?
We’re using a system called CoinText that allows people to receive, hold, and send cryptocurrency by SMS. Each Monday I send the local agent crypto, and he fans it out to the 21 widows, and they go to the vendor to buy food. It’s very messy and very manual, but you can probably see some early seeds of what a scalable system would look like, where you could automate some of these transactions. We very much view this as a start-up. And most start-ups are doing it all with duct-tape and bubble gum, and they figure out what works, and then implement that solution.

So, is more of your time spent on the “government-broken” (like with refugee camps) or “money-broken” (like in Venezuela)?
Money-broken. The government-broken category of work is definitely more challenging. We’re excited about it, but in some of these places, the rules are vague. For example, it is unclear what we are allowed to do in the Kutapalong refugee camp in Bangladesh. And, frankly, for some of the initial experiments, we have taken an “ask forgiveness rather than permission” approach.

It’s one of the worst refugee situations in the world. From what I’ve heard, these refugees aren’t even allowed to buy or sell anything, much less get a job, much less move freely. They’re supposed to only be able to receive the aid they get from NGOs in the camp. It’s bad.

What do you mean exactly?
They have to walk a 12- to 15-day “trail of tears” from where they’re being genocided out of Myanmar, by the military, just over the border into Bangladesh, and they’re greeted with a horrible refugee camp.

Any estimates of how many people you’ve been able to get money to into that camp?
Very small. 21 widows, or 21 families—so maybe 21 times four or five, if you count the entire family.

Is it fair to say that the “ambassador model” is the one you found most effective?
Yes. That seems to the most effective approach, for various reasons. “Direct” is challenging. We just don’t know what’s going on in these places. We’re not going to be an organization like U.S. Aid where we’ve got boots in the ground everywhere. We’re a software organization. We’re hopefully going to build tools, systems, and incentive programs that encourage people to help us [help others], either out of philanthropic need or economic motivation.

This sounds almost like a mission statement?
If you zoom out, all charities do three things. They target: Who gets the help? They distribute: How do I get the help in the hands of people? And they track: Did the people get the help, and did the help help?

Most charities do this in a pretty centralized manner, by privileged Ivy Leaguers sitting in San Francisco or New York, making these decisions. We think there’s an interesting opportunity, using a combination of software and cryptocurrency, to push as much of this decision-making as possible out to the edge, in the local communities. To the people who are on the ground, who understand the situation, who have a stronger desire to help, frankly, but don’t have the resources. If we can build systems to allow them to make as many of these decisions as possible, that’s super powerful. We think the first iteration of that—the V1—is an ambassador program.

So you’re looking for a way to scale this ambassador program. Are you building an app?
Bingo. We’re already developing that app right now.

How far along are you?
It’s already internally launched. We’re using it ourselves. It’s unclear when it will be open to the public.

You’ve kind of touched upon this already, but to address it more squarely: Isn’t there something weird about trying to give cryptocurrency to people who might be starving, and might not have a smartphone, and don’t care about things like “decentralized” principles?
Yeah, yeah. I hear you.

How do you guys grapple with that?
I don’t care about educating refugees about cryptocurrencies. In some of the experiments, we never even used the word crypto to the recipients. This is just money, or credits, or coupons. Because the recipients don’t give a shit what form the help comes in. They need to get some food in their children’s mouths. In the world that you and I are familiar with—or probably you; you probably deal with more crypto lunatics—

Oh, I know those lunatics—
Who say, ‘You need to use x currency! And you need to train them on the elegance of the blockchain and hash-codes and blah, blah, blah…’ No you don’t. Get out of your fucking bubble.

So you’re saying that, generally, your goal is not to proselytize crypto?

Absolutely not. Hopefully, at some point down the road, crypto’s inherent features will shine through, so that it will be obvious that this is the right method to transfer, and to hold, and to transact value, but right now, that’s very difficult, because of the technical challenges and smartphone adoption rates. But we think that down the road, those [barriers] will start to go down, and it will become more compelling, and we want to make sure that we have this platform up and running.

Blockchain is often touted as a way to bring transparency to social impact, by tracing the actual dollars. Are you guys doing that?
Certainly, to the extent that you can. You can track amounts and track addresses very well on the blockchain. It’s more difficult to track identities, and what the money is used for. Part of our platform is going to use block interrogation scripts and tools.

To give one example: We’ll rely upon a lot of self-reporting from our ambassadors about what they did with the money, and who got it. But we’re also monitoring the blockchain, and corroborating what was self-reported versus what actually happened on the blockchain. This is more from a fraud protection point of view. The idea is to trust but verify. So we’ll probably start ambassadors with a small amount of money, and ask them to tell us they’re doing with it. And if it seems that they’re they’re not doing what they said, then they’ll probably get cut off.

Have you started that?
We’re doing it on a manual basis right now.

How often do you detect variances between self-reporting and what you see on the blockchain?
I’m not sure we’ve got enough data yet to to to answer that question. One experiment in Uganda just went horribly wrong, and we had almost 100 percent fraud rate.

What happened?
We gave them a relatively large amount of money, and asked them to give it to other people in need, and they all basically used it for themselves. Because they were genuinely in need. They were lower-middle-income Ugandans, and they need food, they need clothes and what have you. Almost 100 percent of them didn’t comply with the spirit of the experiment. But in other cases, we’ve had surprisingly low fraud rates—single digits.

Wait, how did the blockchain work in this Uganda case, exactly?
You send the money—in this case, I think $50—to the recipient, and you can see it sent to a specific address. We asked them to give it to five other people. But clearly they’re not doing what we asked them to do [because the funds didn’t get transferred from that one address.] Afterwards we went and surveyed them, and they didn’t even pretend that they gave it to five people. They said things like, I bought food for my children, a new uniform for my job. They were saying, this is good stuff I’m doing with the money, don’t worry, it’s being well-used.

Getting back to GiveCrypto overall, how much have you raised so far?
It changes every day, thanks to crypto volatility. [Laughs.] If you used time of donation, we raised $4.1 million. A lot of that was raised in early to mid-2018; the ETHs and the bitcoins have decreased quite a bit in value. So right now we have about $2 million. Ninety percent of which is held in crypto, 10 percent in U.S. dollars.

How has the bear market affected what you’re doing?
It would probably be a little more difficult to raise money, which we’re not trying to do.

Is it fair to say that most of your funds came at the very beginning?
We haven’t tried to get additional donations. Basically, the only fundraising we’ve done to date, is Brian hosted a dinner in San Francisco and invited most of the crypto-celebrities. Thankfully a lot of them ponied up money. You can see who gave. [Editor’s note: The list of donors, includes individuals like Zooko Wilcox, Chris Larsen, Brock Pierce, Roger Ver, and founder Brian Armstrong.]

I’m guessing that 90 percent of the current treasury comes from those high-dollar donations, and not people chipping in $10 in ETH here and there?
I’ve since been educated by more experienced charity people that there are two ways to raise money, and they’re mutually exclusive: You either get a couple of big checks, from a few people, or a bunch of small checks, from a whole bunch of people. Most charities do the former.

Huh, I wouldn’t think they’d be mutually exclusive.
Yeah. The naive, optimistic entrepreneur in me wants to say, ‘we can tackle both!’, but the more experienced people say, ‘No, no, no, it’s a totally different fundraising mindset,’ to do one versus the other.

What are your plans for 2019?
We’re shifting from super early-stage experimentation into more of a product development phase. The focus is very much on building the first layers of this platform, to do this at scale.

Where do you see this in five years?
It’s like that saying with technology—things seem to move very slowly in the short-term, but very quickly in the long-term. I’m butchering that, but you know what I mean. I think that in five years we’re going to look back and say, ‘You were giving to people in need, and you gave them a goat? Or, you gave them physical cash? Why wouldn’t you give them digital currency, as it has so many advantages?’

Love that enthusiasm. No more goats. Best of luck to you and the team.

Photos courtesy Joe Waltman/GiveCrypto.