CD171：卡列 - 闲聊与现金

这类技术依赖网络效应生存。

So these things, they live off the network effect.

所以热度是系统韧性的关键，因为它们需要由个人来运营，而不是指望手机运营商架个天线让你上网。

So excitement is part of the resilience of these systems because they need to be run by by individuals instead of just hoping that a cell phone provider will, like, put an antenna there and put you online.

因此从基础设施角度看，这种热度肯定大有裨益。

So, I think the excitement definitely helps just from an infrastructure perspective.

但你要知道，现在的硬件和十年前已经完全不同了。

But, you know, like, hardware today is definitely different than than ten years ago.

如今你手机里的硬件已经相当不可思议了，蓝牙芯片的传输距离，你知道的，已经提升了数倍之多。

The hardware in your phone is kind of incredible at this point, and the distances that your Bluetooth chip can traverse has, you know, gone up many multiple times.

而能耗却降低了好几倍。

The energy that it uses has gone down many multiple times.

所以事情正变得越来越好。

So things are just getting better and better.

我我我预计这种趋势会持续下去。

And I I I just expect this to continue to do so.

同时，你知道，我清楚之前已经尝试过很多次了，但我不会把你提到的这些项目称为失败。

At the same time, you know, I'm I know that it has been attempted many times before, and I wouldn't call these projects that you mentioned failures.

它们可能只是从未达到一个临界点——足够多的人使用它，使其对大众真正有用，而不仅仅是供人把玩尝试。

They just maybe never hit like a threshold where enough people used used it so it became useful to people other than just people wanna play around with it and try it out.

需要明确的是，BitChat也仍需接受这个考验。

And just to be clear, this remains to be seen for BitChat too.

但我们的技术更先进，需求也在持续增长。

But our technology is better and the need just keeps going up.

因此我认为，我们在地球上度过的每一天，都让我们更清楚地看到需要替代性基础设施的理由。

So I think with every day that we spend on this earth, we see more and more reasons why we need alternative infrastructure.

而且我觉得像Nostr这样的技术，如果与比特币等其他自由技术结合——比如用这个基础设施来传输电子现金，这正是我最初对它感到兴奋的主要原因之一。

And and I think like with things like Nostr, if you combine it with other freedom technologies like Bitcoin, if you send around you know, if you use this infrastructure to send around e cash, which is one of the main reasons why I was, like, excited about it in the very first beginning.

你知道的，我们现在能用它做的事比几年前多得多了。

If you you know, we have many more things that we can do with it today than than a couple years ago.

所以我们只需要继续尝试下去。

So we'll we'll just have to keep trying it.

我这么做是因为，虽然我们可能在几年前失败了，但在我看来，这不应成为我们未来停止尝试解决这些问题的理由。

And I do because we've like, we might have failed a couple years ago, but in my mind, that's not a reason for to stop trying to to stop trying to fix these issues in the future.

如果我们不去尝试，就绝对无法实现这些项目的目标。

If if we don't try it, like, we'll definitely never achieve the goals of these projects.

所以我看不出有任何理由不再试一次。

So I don't see like any reason to not try it out again.

正如你所说，替代方案就只有中心化的互联网服务提供商，而这些存在诸多问题，却没有人真正致力于改进。

And as you said, the alternative is is just centralized Internet ISPs, and those come with so many problems and no one is really working on improving that.

确实有一些零散的项目试图成为更好的ISP，但它们自身也需要成为中心化系统。

There are a couple of, you know, projects here and there that try to be a better ISP, but they themselves need to be centralized systems.

因此，去中心化系统只是更难启动、更难构建、更难维持，但这并不意味着它们无法成功。

So decentralized systems are just harder to bootstrap, harder to build, harder to keep alive, but it doesn't mean that they cannot succeed.

我们在比特币上已经见证了这一点。

And we've seen this with Bitcoin.

在BitTorrent或Tor上也是如此。

We've seen this with BitTorrent or Tor.

我希望并相信，在互联网或通信基础设施方面，我们也能取得类似的成就。

And I hope and I believe that we can also achieve, something in that regard with, Internet or communication infrastructure.

是的，具体到网状网络，它面临着有趣的先有鸡还是先有蛋的采用困境，这在我们看到许多用户首次下载BitChat时的反应中就能体现，对吧？

Yeah, I mean, mesh specifically has an interesting chicken and egg adoption thing, which I mean, is what we've seen with a lot of the responses by people who download BitChat for the first time, right?

比如他们会说：'我下载了，但上面一个人都没有。'

Which is, oh, I downloaded it, and I don't see anybody on it.

是啊，这确实很有意思。

Yeah, that's interesting.

因为更像是，我的意思是，相比其他技术来说，对吧？

Because more so like, there's like, I mean, for more so than other tech, right?

就像Mesh技术初期面临巨大的普及障碍，这让我谨慎乐观地认为，它凭借如此高的热度或关注度起步，或许我们能突破这个普及障碍？

Like mesh has this massive adoption hurdle in the beginning, which is what leaves me cautiously optimistic that it hit the ground running with so much hype or so much excitement that we might be able to get through that adoption hurdle?

看，是的。

Look, yeah.

我是说，人们需要了解它。

So I mean, people need to know about it.

所以它需要流行起来才能发挥作用。

So it needs to be popular in order to work.

目前的目标不是取代互联网，对吧？

And the goal right now is not to replace the Internet, right?

而是为人群孤岛提供通信服务。

It is to offer communication to islands of people.

而且这样的孤岛有很多。

And there are many islands out there.

如果有个应用能让你轻松打开后两秒内就连接到周围所有人。

So if there's an app that just you can very easily just open the app and within two seconds, you're connected to anyone around you.

这有很多应用场景。

There are many use cases for this.

就像你说的，可能是音乐节、会议、体育赛事等。

It could be, as you said, it could be a festival, it could be a conference, right, a sports event.

太多场景下人们不仅因为没网络，更因为无法自动连接周围人。

There are so many situations where people not not only because they don't have Internet, but because they are not like automatically connected to anyone around you.

想象一下，你正在参加一个比特币会议，只想向房间里的所有人发送一个链接。

Like, imagine, you're you're on a you're on a Bitcoin conference and you just wanna send a link to anyone who's in the room.

你会怎么做？

How do you do that?

正确的方式。

The Right.

你知道，要使用互联网服务来公布你的位置，一种方法是在某处发布链接，然后投影二维码，希望人们能扫描它。

You know, just announcing your location to use an Internet service for it, one way would be to post a link somewhere and then just project the QR code and hope that people can scan it.

对吧？

Right?

所以有一些迂回的方法试图解决这个问题。

So there's like roundabout ways of going trying to solve this issue.

但真正将通信限制在地理位置或空间范围内，基本上要求你当前必须向互联网服务暴露你的位置。

But really limiting communication to a geolocation, to a spatial extent, basically requires you to docks your location to an Internet service today.

所以我能想象人们会遇到某些情况，或者已经心中有数。

So I could imagine that people will come up with situations or basically have things in mind already.

可能是你的仓库、公司或其他地方，你不想使用互联网，因为不希望人们创建账户并说：'嘿，应用，我现在在纳什维尔，想接收所有发给比特币公园的消息'之类的。

It might be your warehouse, it might be your company or something where you don't want to use Internet because you don't want people to make an account and say, hey, app, you know, I'm now in Nashville, and I would like to get all the messages that are addressed to Bitcoin Park or something like that.

对吧？

Right?

我不想暴露自己目前在比特币公园的位置。

I don't wanna dox my location that I'm in Bitcoin Park right now.

所以其中一种方法是使用网状网络。

So one of the ways is a mesh.

在这种情况下，如果你只考虑与你想要交流的小群体和身边的人，障碍就会小得多。

And there, the the hurdle is way smaller if you just think about small groups and people around you that you want to communicate with.

因为这样你就不必

Because then you don't

处理采用的问题。

have to deal with the adoption thing either.

你只需要告诉身边的人，下载这个应用。

You just have to tell the people around you, download this app.

是的。

Yeah.

你可以按需使用它。

You can use it on demand.

你可以发送几条消息，分享一些信息，然后直接关闭它。

You can send around a few messages, share some information, and just close it again.

你必须一直连接到全球网络才能使这个东西有用。

You have to be connected to the global mesh all the time to make this thing useful.

另一种可能是你让它后台运行，然后在城市里四处走动之类的。

Another thing could be that you leave it running in the background and you walk around the city or something like that.

你可以接收信息，也可以向你周围的设备或人发送信息。

You could kind of receive information, also send out information to devices or people around you.

我不知道这会涉及什么样的使用场景。

I don't know what kind of use cases that would entail.

有人开玩笑说，你知道，这是一种新型的城市单身人士调情应用，比如你在地铁里看到某人，想和他们说话之类的。

Some people were joking about, you know, it's this new kind of flirting app for urban, singles or something that you find someone, like you you see someone in the metro and, you wanna talk to them or something like that.

听起来有点吓人，但你明白我的意思。

Sounds a bit creepy, but, you get the point.

就像你可能是个喜欢四处走动、对周围分享的信息感兴趣的人。

Like you you might be someone who just like moves around and is interested in information being shared around you.

再说一次，如果你想在互联网上实现这个功能，就必须追踪并分享你的位置给某个服务。

And again, if you want to build this on the Internet, you would have to track your location and share your location with a service.

而通过网状网络这样的本地连接方式，你就不需要这么做了。

And with local connectivity like with the mesh, you don't have to do that.

所以我认为有很多方法能让它变得有用，甚至是我们现在都想不到的。

So, I think there are many ways to make this useful that even us cannot really think about yet.

目前来说，关键是提供这种能力，人们自然就会为这类东西想出有趣的应用场景。

And for now, it's about providing, the ability to do that, and people, like, automatically come up with interesting use cases for this kind of stuff.

是啊。

Yeah.

我是说，我们已经开始看到这种现象了。

I mean, we're already kind of seeing it.

比如我在比特币之外的另一个关注点，就是与美国本地牧场主建立关系，获取优质肉类来喂养家人。

So like one of my focuses one of my tangential focuses outside of Bitcoin has been developing relationships with local American ranchers and getting good quality meat to feed my family.

我觉得，强大的社区会互相支持，在这种事情上互相照应。

I think, you know, strong communities support each other and and have each other's back with this kind of thing.

优质食物是生活质量的关键，也是家庭兴旺的重要因素。

And good food is a key aspect of, quality of life, and a flourishing family.

现在很多牧场都在使用GMRS无线电设备，差不多也是出于类似目的，对吧？

And a lot of these ranches right now, they use GMRS radios for kind of a similar purpose, right?

这些都是短距离的，你知道，五英里视距无线电。

So these are short, you know, five mile line of sight radios.

他们只想在自己的牧场内通讯。

They just want to communicate among their ranch.

显然这需要比蓝牙更远的距离。

Obviously that requires greater distance than Bluetooth.

但我想你可以想象这样一种情况：由于这些都是开源的，你可以在牧场周围放置物理中继器来实现这类用例。

But I think you could imagine a situation where since all of this is open source, you have, you know, physical repeaters that you put around a ranch or something to accomplish that type of use case.

实际上这可能比纯音频通讯更有帮助，而他们现在只能依赖这些短程无线电。

And it could actually be more helpful than just pure audio communication, which is what they have to rely on with these short range radios.

是的。

Yeah.

确实。

Definitely.

所以这些桥接方式会非常有趣。

So these bridges, I think, will be super interesting.

无论采用哪种桥接方式，都有不同的实现方法。

Like, whatever whatever the type of bridge will be, there are different ways to approach this.

可以是无线电桥接，就像你提到的那种。

It could be like a radio bridge, like the one that you mentioned.

也可以是使用Noster的互联网桥接，或者直接通过WiFi将你的网状网络连接到另一个网状网络。

It could be an Internet bridge that uses Noster, or, like, it could be also a direct Wi Fi connection from your mesh to another mesh.

所以你可以想象这样一种场景：牧场里所有的手机都能与另一个牧场的所有手机通话。

So, you could imagine a situation where, you know, all the phones around in the ma in in the ranch, they can actually talk to all the other phones in the other other ranch.

对吧？

Right?

所以对你来说，不必去电台，拿起麦克风说'嘿'。

So it you you don't have to go to the radio, pick up the microphone, and say, hey.

有人在听吗？

Is anyone listening?

你可以直接发条消息然后等回复，就像我们习惯的任何聊天应用那样。

You could just send over a message and wait for a response like we're used to with any chat app, that we already use.

是啊。

Yeah.

我笑了，因为我喜欢我的两个节目都是直播的原因之一，就是Scardust Official里的实时聊天和实时恶搞。

I'm laughing because, the one of the reasons I love, that both of my shows are live shows is the live chat and the live trolling that happens in Scardust Official.

我想提醒大家，GMRS代表通用移动无线电服务，而说'GMRS电台'这种说法本身就值得被恶搞。

I would like to remind us that GMRS stands for a general mobile radio service and saying GMRS radios is something worthy of being trolled about.

我觉得本地通信另一个很酷的方面是当你加入Cashew的时候。

The other aspect here that I think is kind of cool for local communication is when you mix cashew into it.

突然间，你不仅有了本地化通信，还有了本地化商业。

And so then all of a sudden, not only do you have localized communication, but you have localized commerce.

比如你可以想象类似农贸市场的情景，市场里的每个人都能互相交流并完成支付。

You can imagine something like a farmer's market for instance, where everyone in the farmer's market is able to communicate with each other and pay each other.

我看到你们已经把Cashew整合进安卓版应用了？

I saw that, you've integrate I believe you've integrated Cashew into the Android version of the app?

两个平台都整合了。

Both.

我是说，我在安卓端实现了，但我们团队也进行了整合，iOS端也整合好了。

I mean, I did on the Android, but we we also integrated in in our team, we also integrated into the iOS one.

所以两边都准备好了。

So we have both ready.

那现在iOS端已经上线了吗？

Is it in is live in the iOS one right now?

没有。

No.

没有。

No.

都还没。

Neither.

所以我们还得在Bitchat相关功能上多下点功夫。

So we'll have to work a bit more on the on the Bitchat side of things.

但只要那边稳定性达标，我们就会尝试在两端都推送钱包功能。

But as soon as that feels stable enough, then we'll we'll, like, try to push the wallets on on both sides.

所以，是的，这非常有意思。

So, yeah, this is super interesting.

这也是最初吸引我注意力的原因——我一直在研究用Cashew支付的各种方式。

This is also like why how it caught my attention primarily because I've been researching a lot on different ways of paying with Cashew.

由于Cashew使用电子现金（e cash），而电子现金是裸代币，理论上可以通过任何你能想到的通信媒介发送。

And because Cashew uses e cash, e cash is a bare token, you can kind of send it with any communication medium that you can think of.

目前我们基本都局限在使用二维码的领域，无论是比特币闪电网络发票、通过二维码传输的现金代币，还是法币支付。

Currently, we're stuck in this realm of using QR codes for almost all things that we do, whether it's a Bitcoin Lightning invoice, whether it's a cash or token that is transmitted via QR code, but also with fiat payments.

比如想想亚洲，去过亚洲的人都知道，二维码无处不在。

Like if you think of Asia, for example, and for those who have traveled to Asia will know this that QR codes are everywhere.

走进一家商店，他们会给你看二维码，你应该用手机摄像头扫描它。

You walk into a store, they show you a QR code, you are supposed to scan it with your camera.

而且，你知道二维码能用，但它们能传输的内容非常有限。

And, you know, QR codes work, but they are super limited in what they can transport.

况且，摄像头本就不是为这个设计的。

And, cameras never were made for this.

对吧？

Right?

这有点像一种临时解决方案，一个我们不得不一直使用的变通方法。

It's kind of a hack, a hack that got stuck with, with us.

所以我们用这种超高吞吐量的设备——每秒能传输兆比特信息的摄像头，来传输二维码这种只有几个字节的数据。

So, so we're using cameras super high throughput device with megabits of information per second to transmit like a couple bytes of data in form of a QR code.

这就是我们目前正在做的事情。

That's what we that's what we're currently doing.

同样地，我们也在为Cashew做类似的事情。

So, we also we're also doing similar things for Cashew.

为了持续创新，我们做了很多实验，比如尝试NFC技术。

So we've and to to keep innovating in that realm, we have been experimenting a lot with NFC, for example.

比如电子现金的碰一碰支付：当两部手机靠近时，通过NFC直接发送电子现金令牌——这也是利用无线电或电磁波在手机间传输，同样存在吞吐量限制，但最关键的是苹果公司限制了它的功能扩展。

So tap to pay with e cash where you approach one phone to another phone and then sends the e cash token directly from NFC, also uses radio or electromagnetic waves to send from one phone to another, and that is also limited in throughput, but most importantly, it is limited by Apple to be able to do what you want to do.

实际上我们已经能用安卓手机之间、或者银行卡与手机之间实现电子现金的碰一碰支付了。

So we can do tap to pay with eCache already from an Android phone to an Android phone or from a card to a phone.

这也很酷。

That is also very cool.

但是，从安卓手机到iPhone，你无法通过那个方向发送eCache。

But, from Android to iPhone, you won't be able to send eCache in that direction.

你可以从iPhone发送到安卓手机，因为安卓手机功能更多，它可以接收，但iPhone无法通过NFC芯片接收。

You can send from iPhone to Android because Android just can can do more and it can it can receive, but an iPhone cannot receive on the NFC chip.

因此，如果你想在手机之间传输数据，会有很多技术限制。

And so there are many technical limitations if you want to send around data from phone to phone.

所以当我听说BitChat时，我立刻想到我们之前用缓存奖励做的一些实验，我们想用低功耗蓝牙在设备间传输eCache。

So when I heard about BitChat, I immediately thought about some experiments that we've been doing before with cache awards where we wanted to use Bluetooth low energy to send eCache from one device to another.

这看起来很有前景，因为苹果允许你做到与安卓手机相同的程度，这为让所有人接入这个支付系统或支付交互开辟了途径。

And that seemed promising because Apple allows you to do as much as you can do also with an Android phone, so that kind of opens up a way to onboard everyone onto this payment system or payment interaction.

所以我设想的是，在未来，我希望你能走进一家商店，然后你会有一张10美元的收据之类的东西。

And so what I have in mind is basically in the future, I hope that you'll be able to walk into a store, And then you'll, you know, you'll have a $10 receipt or something.

店员会在终端输入金额，然后你的手机就会弹出提示问你是否支付10美元，你只需按下按钮。

The person at the store will enter it in their terminal, and then it'll just pop up on your phone and say, you want to pay $10, you press the button.

你不需要扫描二维码。

You don't have to scan a QR code.

你不需要点击任何东西，但你的设备会收到通知，或者你可以主动将支付推送给接收方，他们会看到通知并接收付款。

You don't have to tap anything or but you'll just get a notification on your device, or you can push the payment from your device to the to the receiver, and they'll see a notification and can receive the payment.

这就是我们的设想，我们之前一直在研究这个。

So this is what we had in mind, and we've been researching it before.

所以当Vidget出现时，我们就觉得，没错。

And so when Vidget came around, we're like, yeah.

操。

Fuck.

对。

Yeah.

这正是我们需要的实验对象。

This is exactly what we what we need to try this with.

所以我们实际上做了完全相反的事。

So what we did is basically the opposite.

我们没有在Cashew钱包里集成蓝牙功能，而是反其道而行——把Cashew钱包集成进了BitChat。

Instead of implementing e Bluetooth into a Cashew wallet, we went the other way and we implemented a Cashew wallet into BitChat.

我们通过自研的CDK库（Cashew开发工具包）实现，它基于Rust语言。

And we did that with, our library called CDK, is the Cashew DevKit, which is based on Rust.

这让构建钱包并集成到其他应用变得极其简单。

And it makes it extremely easy to build wallets and implement it into things.

这是个Rust库，能帮你处理所有缓存逻辑，使用它能让开发缓存钱包变得极其简单快速。

So it's a Rust library that does all the cache stuff for it for you, and it's using the Rust library, it just allows you to build cache wallets extremely easy and super fast.

目前我们还在推广这个库的阶段，毕竟它太新了。

So with this with this library, we we're still kind of in the process of pushing it out and also putting the word out there because it's so new.

能做到这种程度的技术本身就很前沿。

The fact that you can do this is is still super new.

但本质上，它把开发缓存钱包的智商门槛从130降到了70。

But essentially, it brought down the IQ requirement for building a cache wallet from maybe a 130 to 70 now.

现在你只需要几分钟就能把Cashewallet集成到你的应用里。

Like, you you can just implement a Cashewallet within a couple minutes into your application with that.

这就是我们所做的。

So that's what we did.

我们采用了BitChat的iOS版和Android版。

We took BitChat for iOS and BitChat for Android.

我们采用CDK并为懂行的程序员添加了绑定接口。

We took CDK and added bindings for it for the programmers who know what this means.

对于不了解的人来说，绑定就像是连接我们使用的库与目标应用程序之间的粘合剂。

For those who don't know what this means, binding is basically like the glue between the library that we're using and the application where we want to implement it to.

于是我们把CDK库直接集成到BitChat安卓和iOS端，几小时内就实现了两个应用中完全独立运行的缓存钱包功能，就这么简单。

So we we took the CDK library and just put it into Bitchat Android and Bitchat iOS, and within a couple of hours, we had fully functioning cache wallets in both of these applications, completely independently implemented, And that was it.

我们在BitChat应用里添加了一个小钱包图标。

So we had like a little wallet icon in the BitChat app.

点击它就会打开一个看起来像普通现金钱包的比特币钱包。

When you press it, it just opens a Bitcoin wallet that just looks like an ordinary cash wallet.

通过这个钱包，你可以用闪电网络充值，然后BitChat应用里就会显示相关数据。

And with that wallet, you can charge it with lightning, then you'll have some stats on your BitChat app.

之后在聊天界面里，无论是私聊还是群聊，你都可以直接发送这些数据。

And then you can go into in the chat, you can either in the in the direct message from me to you or even in the public chat, can just drop some stats in there.

整个界面就是BitChat那种标志性的赛博朋克风格。

So the interface was you you would have this typical like cyberpunk interface of BitChat.

你输入/pay指令，加上聪数金额，还可以选填备注。

So you write slash pay, and then the amount in Satoshis, and a comment if you like.

点击发送后，聊天窗口就会出现支付信息弹窗，显示'这里有200聪的支付款项待领取'。

You press send, and then it drops this payment blob in the chat that, you know, pops up everywhere and says like there is a 200 Satoshi payment for you.

你可以点击兑换，款项就会存入内置在你BitChat应用中的缓存工作区。

And you can press redeem, and it ends up in your in your cache work that is built in into your into your BitChat app.

这非常酷，首先eCache是承载式的，我们可以通过蓝牙进行传输。

So this is super cool because first, eCache is bearer, so we can use Bluetooth to send it around.

但同样很酷的是，付款方仍然不需要联网就能完成支付。

But what's also very cool is that the payer still doesn't require Internet to make that payment.

请注意，这种情况下收款方需要联网，否则无法防止双重支付。

Mind you, the receiver needs Internet in that case because otherwise you cannot prevent double spending.

所以交易双方必须有一方联网，这意味着你无法完全脱离网络，但付款方可以离线完成支付。

So one of the parties needs Internet, so you're kind of leaving this fully offline realm and require at least one person to have internet, but the other person can pay without internet.

就像这个商店场景。

It's like this shop situation.

走进商店买件T恤，标价20美元，你只需打开钱包按个按钮，20美元就会立即转入聊天窗口。

Walk into a store, you buy a shirt, it's a they say $20, and you just like open your wallet, press a button, $20 just immediately drops into the chat.

你不需要保持在线。

You don't need to be online.

支付瞬间完成且始终有效，交易就这样达成了。

It just happens instantly and always works, and it just the payment is done.

这就是我们努力实现的方向。

So, that's that's where we're trying to go.

我的意思是，人们没意识到这其实已经是传统移动支付的标准了。

I mean, what people don't realize with that is that is already the standard with traditional mobile pay payments.

比如使用Apple Pay时，付款方实际上不需要联网。

Like, if you use Apple Pay, the payer actually does not need to be online.

比如说，在发展中国家，你可能身处丛林深处。

Like, if you're in the developing world, you know, you could be like in the middle of the jungle.

如果有人有苹果支付的销售终端，你的手机不需要联网。

And if someone has an Apple Pay point of sale, your phone doesn't have to have Internet.

你只需轻触支付，接收方有网络连接就能处理并确认付款。

You tap, you just tap to pay at that point and the receiver has Internet and is able to then process the payment and confirm the payment.

这就是苹果支付使用如此顺畅且毫无延迟的原因之一。

And that's one of the reasons why Apple Pay works so seamlessly when you use it and there's no lag there.

所以这已经是行业标准了。

So that's already the standard.

Cashew某种程度上将这个标准引入了比特币领域。

Cashew kind of brings that standard into the Bitcoin realm.

我有个快速问题想请教你，就是关于你提到的那种丢到本地群聊里的使用场景。

I have a quick question for you just in that one use case where you're, like, dropping it into the local group chat.

我觉得这就像个比特币会议。

I'd say it's a Bitcoin conference.

如果你把它丢到本地群聊里，就会变成一场看谁先尝试兑换的竞赛。

If you're dropping it into the local group chat, that that becomes, a race condition for who tries to redeem it.

就像谁先连上网谁就赢。

It's like whoever gets Internet first.

对吧？

Right?

完全正确。

Exactly.

是的。

Yeah.

就像在蓝牙聊天中撒下一些比特币碎屑，看谁能先抢到。

It's like leaving, little crumbs of of Satoshis in the in the Bluetooth chat and seeing, like, who who gets it first.

但正如你所说，付款方没有网络而收款方有网络的情况非常普遍。

But, yeah, as you said, like, the the situation where the payer doesn't have Internet but the receiver has Internet is super common.

这就是信用卡的运作方式。

That's the way how, credit cards work.

补充一下你刚才说的，Apple Pay也是这么运作的。

So just like to add on what you said, that's how Apple Pay works.

Google Pay同样如此，信用卡的触碰支付也是这个原理。

That's also how Google Pay works, but that's also how credit card tap to pay works.

Apple Pay和Google Pay之所以这样设计，根源在于信用卡体系。

And that's, you know, the the reason why Apple Pay and Google Pay works this way is because of the credit cards.

无论是离线触碰支付的信用卡还是Apple Pay，其原理都是从苹果服务器下载令牌。

And, like how a credit card, offline tap to pay credit card or your offline tap to pay Apple Pay works is by downloading tokens from the Apple Pay server.

其实这和电子现金的运作方式有点相似。

So it is a little bit similar to how how it works with e cash, actually.

当你的手机联网并使用Apple Pay时，它会向苹果服务器申请：'我有XYZ银行的信用卡，想下载20个离线令牌'。

So when your phone is online and you have Apple Pay, your phone goes to Apple Pay server and says like, I have this credit card company x y z, and I'd like to download like 20 offline tokens.

然后Apple Pay或Google Pay服务器就会给你这20个离线令牌。

And then the Apple Pay or Google Pay server gives you these 20 offline tokens.

之后你就算在荒郊野外也能直接手机碰一碰完成支付。

And then you can walk around in the middle of nowhere and just tap your phone.

正如你所说，它能即时完成的原因是你的手机不需要联网。

And the reason why it's instant, as you said, is because your phone doesn't need Internet.

只需要收款方联网，通常是商家，所以他们经常保持网络连接。

Just the receiver needs Internet, which is typically a store owner, so they have an Internet connection often.

而你就像商场里没有手机信号但依然能完成支付的手机用户。

And you're probably like the mobile phone user who's like in the mall without a cell connection, but it still works on your phone.

这是因为当你联网时手机接收的令牌会通过NFC直接从手机传输到支付终端，支付终端（在Apple Pay和Google Pay的情况下）会将该令牌发送到网络进行验证：这个用户是否合法、能否完成支付等等。

And that's because the token that you received while you had Internet that is on your phone just flows directly from your phone via NFC to the payment terminal, and then the payment terminal takes that token, in in the case of Apple Pay and Google Pay, takes that token and sends it off to the Internet and says like, hey, am I allowed to know, is this user legit and can they make a payment and so on and so forth.

然后后端网络会为你执行支付操作。

And then that executes the payment in the back end on the Internet for you.

电子现金的工作原理类似，但针对比特币且具有隐私性。

So for e cash, it kind of works the same way, but for Bitcoin and with privacy.

这些支付系统的关键区别在于：使用电子现金时比特币也能实现非接触支付或离线支付场景，同时还具备极强的隐私保护。

And that's like the critical differentiating factor between these payment systems is we can also have tap to pay or offline payer situations for Bitcoin when we use e cash, but plus insane privacy on top of that.

对于Apple Pay和Google Pay，网络服务器需要知道你的身份，因为你发送的令牌并非资金本身。

So for Apple Pay and Google Pay, the Internet server needs to know who you are because the token that you sent is not the money.

它本质上是信用卡的访问密钥，必须联网才能识别你的身份。

It is like an access key to your credit card basically, and that needs to hit the Internet and need to identify you.

在电子现金系统中，令牌本身就是聪（比特币最小单位）的代表。

In the case of eCash, the token itself is the representation of the Satoshi.

你将该令牌发送给支付终端后，终端会通过互联网与电子现金铸币厂完成一次交换，支付即告完成。

And then you send that over to the payment terminal, and the payment terminal does one swap with the e cash mint over the Internet, and the payment is done.

铸币厂无需知道你的身份。

And the mint doesn't need to know who you are.

接收方无需知道你是谁。

The receiver doesn't need to know who you are.

事实上，由于eCache自身的隐私特性，没人能查出你的身份。

In fact, no one can figure out who you are because of the privacy properties of eCache itself.

因此我对这个应用场景非常看好。

So I'm super bullish on that use case.

这绝对是我们将要重点关注的领域。

It is definitely like something that we're going to focus more.

我们已经在这方面投入了大量精力，而离线状态能提供绝佳的隐私保护——因为你的手机无需联网。

We do already focus a lot on this, and the fact that you're offline gives you incredible privacy because your phone doesn't need to go online.

你甚至不会留下任何痕迹，比如'我开机了'这类记录。

You you don't even leave a trace like saying, hey, I turned on my phone or something like that.

你根本不需要联网。

Like you don't need to go online.

而且操作是即时完成的，不需要先连接，也不用像PayPal那样先登录账户，更不用忍受各种延迟。

And it's instant because there is no, you know, I need to connect first and, like, get into my PayPal account and log in and, like, wait for the delay and shit like that.

只要把钱转过去就完事了。

Just send over the money and you're done with it.

是啊。

Yeah.

首先，要特别感谢SoapMiner打赏的21000聪，他家的肥皂超赞。

So, first off, huge shout out to SoapMiner who zapped 21,000 sats, has great soap.

不错。

Nice.

而且这个支持是节目传奇般的支持。

And it's a good support is a is a legendary support of the show.

我非常感谢你们的支持。

I appreciate your your support.

我们还有Pablo，他本身就是Nostra生态系统中的传奇建设者，他慷慨捐赠了21000，感谢各位绅士。

And we have Pablo, who's a legend himself building in the Nostra ecosystem, who zapped 21,000 sad saying, appreciate you gents.

谢谢大家支持这个节目。

Thank you guys for supporting the show.

这真的比金钱更有意义。

It really it it means more than the money.

随着支持不断涌入，这让我持续前进。

As the support comes in, it's it keeps me going.

是的，关于cashew这部分，我的意思是，现在所有自由技术项目都在这里汇聚。

Yeah, on the cashew piece, I mean, this is where it's like all these freedom tech projects are like coming together.

它们会相互叠加形成强大的合力。

It's like it all compounds and gets super powerful.

离线组件方面，cashew代币本质上就是一长串文本。

The offline component, I mean, a cashew token, you can kind of, it's just a long string of text.

所以只要能传输长文本，就能传输cashew代币。

So however you can transmit a long string of text, you can transmit a cashew token.

我的意思是，你可以想象在很多信任受限的环境中，人们可以在双方都离线的情况下进行交易。

And I mean, even in a you could imagine a lot of situations where there's like a limited trust environment, where where people can do commerce where both people are offline.

然后在未来某个时刻，接收方需要上线——如果他觉得对方不会诈骗的话。

And then at some point in the future, the receiver needs to go on online if he doesn't think the person's going to scam them.

然后在那时兑换。

And then redeem it at that point.

我是说，我们已经身处其中，人们没意识到，当你用信用卡支付时，商家可能在三十天后被退单。

And I mean, we're already in, like, people don't realize like you go and you pay with a credit card, like that merchant can get charged back like thirty days later.

这是一种方式，你知道的，信任度较低的环境，或者...比如你去J.Crew这样的店，那些拿最低工资的随机员工甚至不认识你。

And that's a way, you know, less trusted environment or it is a high it's you don't, you know, you go into like a j crew or something that random minimum wage employee does not even know who you are.

但如果你在农贸市场之类的地方，我每周都去，付钱给那个农民，他并不需要连续四五个小时联网。

But if you're in like a farmer's market or something, and I'm going there every week, and I'm paying someone like that farmer doesn't necessarily need to be connected to the internet for like four hours or five hours.

之后他可以联网完成兑换。

And then later he can connect to the Internet and redeem.

如果我骗他，他再也不会卖肉给我了。

And if I scam him, like, he'll never sell me meat again.

是啊。

Yeah.

没错。

Yeah.

当面验证。

Proof of punch.

如果你们关系足够近，那么...被骗的概率就会大幅下降。

If you're close enough to someone, then you you the the prob probability of of scamming getting scammed goes down a lot.

不过确实。

But yeah.

所以如果我们开始和法币系统比较，这将会是场轻松的游戏，因为

So if we start comparing ourselves to the fiat system, like, it's gonna be an easy game because

不错。

Nice.

比如，

Like,

你知道，法币系统里的这些玩意儿，简直烂透了。

you know, the these in the fiat sis like, it's it's so fucked.

我们甚至不需要开始，但就当找点乐子吧。

It's we we don't even need to start, but let's do it for fun.

所以在法币游戏里，他们所谓的‘安全’基本上是通过不让你接触硬件本身来实现的。

So in the in the fiat game, they achieve most of the security in in quotes, security by by by basically denying you access to the hardware itself.

所以你相当于有个硬件层面的信任系统。

So you have kind of like a hardware level trust system.

确实。

So yeah.

当然。

Of course.

比如，如果我们想把门槛降那么低，我们当然也能做到同样的事。

Like, if we wanna, you know, lower the bar that much, we can definitely do the same as well.

以信用卡为例，当你用信用卡拍卡支付时，就像你说的，你根本不知道终端会从卡里扣多少钱。

So in a credit card, for example, when you tap to pay with the credit card, as you said, you don't know how much that terminal is going to deduct from your credit card.

可能会有个最高限额，但具体是2美元还是25美元，甚至是200美元，取决于你的信用卡允许的额度。

Like, there might be a maximum limit, but, like, is it, like, $2 or is it, like, $25, maybe $200, depending on what your credit card allows.

所以你根本无法控制这个。

So you you don't control that.

你之所以会“信任”支付，是因为终端本身就像经过测试的硬件，由一家公司制造，这家公司基本上在说：信我，兄弟。

The reason why you, in quotes, trust, the payment is because the terminal itself is like a tested hardware that is built by a company that kind of says, trust me, bro.

我们打造的终端不会让用户感到不安。

We built a terminal that doesn't rock your users.

信用卡本身也是如此。

Same for the credit card itself.

信用卡上有个人人都知道的芯片，就是那个金色的小东西，里面其实有个CPU。

The credit card itself has this chip on it that everyone knows, like this this golden thingy, and there's a CPU inside there.

本质上，信用卡的设计理念是：你无法拆开它，无法分解部件，更无法取出我之前提到的那些令牌——没人能看到这些。

And basically, what this credit card kind of assumes is that you won't be able to open it up and just dissemble the thing and and take out like these these tokens that I talked about before that someone be will be able to see this.

所以，如果你说“好吧”。

So, like, if you say, okay.

我们会相信硬件能防止你双花，或者相信硬件不会让客户不安，那么没错，这些功能我们都能实现。

We'll we'll trust the hardware to not allow you to double spend or we'll just trust the hardware to not rock the customer, then, yes, we can do all of those things as well.

但显然，这也意味着这类系统的适用性极其有限。

But obviously, like, that also means that the applicability of these systems is super limited.

因此你需要经过认证的设备来接收信用卡支付——不过近年来变化很大，现在用苹果手机也能操作。

So you need, like, a certified device to receive a credit card payment or at least, you know, this has changed a lot in in recent years, and by now, you can also use like an Apple phone for it.

当然，苹果手机是超级封闭的系统，但如今支持这个功能的硬件已经多得多。

Obviously, Apple phone is like super walled garden thing, but there there's much more hardware today that supports this.

但总体而言，如果引入硬件级安全机制，我们也能实现完全离线的交易。

But in in general generally speaking, if you introduce hardware level safety, then we can do fully offline transactions as well.

本质上，这取决于你的手机是否允许你双花刚用掉的令牌。

Basically, it just depends on whether your phone will allow you to double spend the token that you just spent or not.

但如果你像我们一样从完全开源的角度来看待这个问题，那么我们就必须假设任何人都可以查看一切，所有内容都是透明的。你可以直接获取软件，修改它，然后开始对系统进行各种操作。

But if you approach this from a fully open source perspective, which we do, then we kind of have to assume that anyone can open up everything and everything is observable, and you can just, you know, you can take the software, you can change it, and then start messing around with the system.

所以我们试图找到一个最小可行信任系统，或者说需要最少信任的系统，即使不了解交易对手也能正常运行，不需要面对面站着以防需要给对方一拳。

So we kind of, like, try to find a minimal viable trust less or least amount of trust system that that still works without knowing your counterparty or with with without having to, you know, stand across them and being reached of of a punch in case the punch is needed.

对吧？

Right?

比如在互联网应用中，当有人试图欺骗你时，你不可能直接给他一拳。

So for example, Internet applications where you cannot necessarily just punch someone in the face when they try to rug you.

是的。

Yeah.

我的意思是，这完全改变了游戏规则。

I mean, this is all it's a complete game changer.

我想说的是，首先我们本应该早点提到这件事，但我们没有。

I mean, I wanna talk about first off, we should have mentioned it earlier, but we didn't.

这是我们的疏忽。

So that's on us.

但BidChat在Play商店里是找不到的。

But BidChat is not available in the Play Store.

所以如果你在Google的官方应用商店里看到有BidChat上架，那并不是官方发布的版本。

So if you're if you're going to the normal App Store on Google and there is a BidChat listed there, that's not the official distribution of it.

可能是骗局。

It might be a scam.

没错。

Yeah.

别别别下载那个。

Don't don't don't download that.

那就像是个大大的警告标志，而且，怎么说呢，这是我的责任。

That is, like, a big warning sign, and, like, I this is on me.

我一直在努力，怎么说呢，和谷歌较劲。

I've been trying to, like, be wrestling with Google.

我还没能自己搞定上传，但正在努力。

I haven't been able to put it up myself yet, but I'm working on it.

所以截至7月30日（也就是今天），别从Play商店下载Bitjet。

So don't download Bitjet from the, Play Store as of July 30, which is today.

等它正式登陆Play商店时，你们自然会知道。

And as soon as it hits the Play Store in an official way, then you'll know.

不过现在你可以访问bitjet.free。

But for now, you can go to bitjet.free.

那是我们的域名。

That's our domain.

是官方域名。

That's the official domain.

在那里你能找到可信应用的链接和资源，不用担心被骗。

And there you find links and resources for the trust trusted application where you can find it without being afraid of getting scammed.

至于iOS版本，你可以去苹果应用商店，直接在iPhone上搜索并下载BitChat。

So for the iOS version, you can go to the Apple App Store, just search it on your iPhone and download BitChat.

安卓用户暂时别从Play商店获取。

For Android, don't get it from the Play Store just yet.

有个冒牌货拿走了我的应用，自己上传到了谷歌。

It's a fake guy who took my application and just uploaded it to Google himself.

现在对我来说基本不可能把它撤下来了，简直太糟糕了。

And it's basically impossible to take it down for me at this point, without kind of like, it's it's it's terrible.

我们别提这事了，总之别从那儿下载。

Let's not even get into it, but, yeah, don't get it from there.

从GitHub获取吧，你可以在bitchat.free上找到链接。

Get it from the GitHub, which, you'll find the link on bitchat.free.

或者你也可以从Zap商店获取。

Or you can get it from, the Zap store.

对吧？

Right?

Zapstore.dev，那个支持Noster的安卓应用商店。

Zapstore.dev, the the Noster enabled Android app store.

可惜的是，它现在还没上架那里。

Unfortunately, it's not there yet.

这在我的待办清单上。

It is on my list.

这在我的待办清单上。

It is on my list.

Zap商店里还没有。

It's not on Zap store.

我开玩笑的，因为大家一直催我上架。

I was joking about this because people keep pushing me to to do it.

我要去做这件事。

I am going to do it.

别担心，伙计们。

Don't worry, guys.

我这就去处理。

I'm going to do it.

但还有个F-Droid平台的事也得处理。

But there's also f droid, which I also need to do.

不过就像

But like

这也不是F-Droid上的。

It's not an f droid either.

一次专注解决一个F开头的事。

One thing f at a time.

除了GitHub哪儿都没有。

It is nowhere except on or on GitHub.

只在GitHub上获取，别处都不行。

Don't get it anywhere but on GitHub.

好的。

Okay.

那就去bitchat.free这个网站——虽然域名有点弱，但去bitchat.free下载APK文件，直接安装就能把应用装到安卓设备上。

So go to bitchat.free, which is a, I mean, a a weak, kind of a weak domain name, but bitchat.free and download the APK, and then you can sideload it onto your Android device by just installing the APK.

APK就是安卓应用文件，可以直接安装任何应用。

An APK is like a an Android application file that you can just install any application.

这是安卓系统最出色的功能之一。

It's one of the best parts about Android.

我确实觉得这很有趣。

I do think it's funny.

我看到bitchat.com，它跳转到了域名所有者关于拥有该域名的推文。

I see at bitchat.com, it goes to the domain owner's tweet about owning bitchat.com.

所以他现在设置的跳转链接，我认为也恰好体现了这个项目引发的意外热潮和兴奋。

So what he has it redirected to right now, which also I think just goes to the unexpected hype and excitement that came out of the project.

我的意思是，多尔西显然没有在发推前先确保域名可用。

I mean, Dorsey clearly did not try and secure a domain name before tweeting it out.

但事实就是如此。

But that is what it is.

Bitchat.free。

Bitchat.free.

这正是我想要的。

What I wanted yes.

看，巴勃罗的样子像是没被骂过。

See, Pablo's like, didn't get bitched at at.

谁被骂过？

Who has bitched at at?

没有。

No.

我试过了。

I I tried.

我试过帕布罗。

I tried Pablo.

我试过奥地利人，就是那个T。

I tried, the Austrians, which is a t.

奥地利人不让你用那个域名，比如bitch.something之类的。

The Austrians don't let you use the, the domain bitch.something.

哦，有道理。

Oh, fair enough.

他们那边规矩特别多。

They're very buttoned up over there.

就像，我本想做个友好游客，但我在奥地利时——尤其是维也纳——他们真的特别不待见我。

It's like, maybe I try and be a friendly tourist, but when I was in Austria, was, they really did not like me, specifically in Vienna.

我觉得他们还在为输掉二战耿耿于怀。

I think they're still salty about losing World War II.

确实。

Absolutely.

但我想说的是，在大家下错应用之前先把安卓声明放出来——对观众来说，你能这么快做出原生安卓应用真的很惊人。

But what I wanted to talk about, but I just wanted the Android disclosure out there before people go download the wrong app, Is for those of us watching, I mean, it's it was quite remarkable that you were able to create a native Android app so quickly.

我以为你可能提前得到了消息，但听起来你根本没提前准备，却在几天内就完成并发布了原生安卓应用。

I mean, I thought you might've had advanced notice, but it sounds like you didn't even have advanced notice and you were able to make a native Android app and release it within a matter of of days.

这可能是这次氛围编程运动最厉害的地方之一。

Might be one of the single most powerful things about this vibe coding movement.

你想聊聊这个过程吗？

You wanna talk about that process a little bit?

是的。

Yeah.

确实如此。

Definitely.

我们来聊聊这个，因为这是我最大的收获之一，除了关于蓝牙的那些知识外，我在做这个项目时学到了很多。

Let's talk about that because that's one of the biggest learnings that I had, while doing like, working on this except for all the things that I learned about Bluetooth.

不过确实。

But yeah.

所以在你之前我并不知道这件事。

So I I didn't know about this before you.

我是和你同时了解到的。

I learned about it at the same time.

我只是坐下来就开始做了，因为我在AI方面有些经验，也用过智能代理，但我觉得其他人也完全可以做到。

I just sat down and did it, because I've had quite, some experience with AI, but and using using agents, but I think everyone out there could have done so too.

或者至少如果你懂一点编程，某种程度上你也能做到。

Or at least if you know a little bit about programming, then kind of you can do it too.

接下来我会详细讲讲这个过程以及我的收获。

And so I'll I'll dive a bit into the process and what I learned about it.

首先，我主要使用了Goose by block和Cursor。

So first of all, what I what I used for it was mainly Goose by block and Cursor.

主要是Goose，因为它能让你轻松切换模型，我最常用的是Claude四和SONNET四。

Mostly Goose because it allows you to easily switch between models and the model that I used heavily was mostly Claude four and SONNET four.

整个过程大致是：先下载JAX实现，然后手动探索一番，再用各种语言模型先给我解释这个软件。

And the process of it was basically like downloading JAX implementation and then exploring it a bit by hand and then using various LMs to first describe the software to me.

所以我用了具有大上下文窗口的工具。

So I used something with a large context window.

我想最开始用的可能是Gemini，就是用来探索这个庞大的代码库。

I think it must have been probably Gemini that I started with, is to just, you know, explore this huge code base.

那时代码库已经很大了，所以很难全面了解。

At that point, it was already a big code base, so it's hard to get an overview.

比如，有哪些功能？

Like, what are the features?

给我总结一下这个东西能做什么。

Give me a summary of what this thing can do.

于是我收到了机器人的回复，看到它能做什么之类的信息。

So I get this response from the bot, and I see like, okay, this is what it can do and so on.

然后从那时起，我开始扮演项目经理的角色，而不是程序员，就像项目经理那样说：好的，现在这是我们的iOS版应用，我们想推出Android版。

And then from there on, I start acting as a project manager, less than like a coder, but a project manager saying like, okay, now this is an iOS version of our application, and we wanna launch like an Android version of our application.

我们希望它与iOS版100%兼容。

We want it to be a 100% compatible with the iOS version.

在开始编码之前，我们先制定一个实施计划。

Now before we start coding, let's start with an implementation plan.

就像在公司里规划开发应用项目时那样，通常不会直接动手开发，而是先列出所有需要逐步完成的步骤。

Like, as as if you would start planning a project to build an app in a in a company, you do you you usually don't just start building it, but you make a plan of all the individual steps that you need to scratch off like one by one.

我们需要先实现公共聊天功能，然后是点赞功能，接着是侧边栏等等等等。

What is we need first to implement like the public chat, and then we need to implement like liking things, And then we need to implement, like, that sidebar and blah blah blah blah blah.

就是列出所有需要逐步完成的事项清单。

Like, you make a list of all the things that need to be done step by step.

我还会做的是，我会反复告诉它。

And what I also do is, I add I I I tell it.

我告诉这个机器人，比如，给我一个需要完成的任务清单，还有每个步骤的具体要求，以及满足这些步骤需要哪些条件，这样你就能完成这个步骤。

I tell the bot, like, give me a list of things that need to be done, and also, like, what are the requirements in that in each step and what what kind of what what requirements make that step fulfilled, like, let you fulfill that step.

于是，我得到了一个长长的任务清单，都是为将这个项目移植到Android平台需要完成的。

So, I get a long list of things that need to be done in order to port this to Android.

我查看这份清单，把大语言模型生成的输出结果拿来，然后开启一个新的任务。

And I look at the list, and I take that output from an LLM, and I start a new one.

我把这个大语言模型当作我的开发人员来对话。

And with that LLM, I talk to as if it's my developer.

我把我之前制定的实施计划交给这个大语言模型，然后说：不行。

And I take my my implementation plan that I I planned before and give it to that LLM, and I say, no.

我们现在从第一步开始。

We're starting now with step one.

明白吗？

You know?

比如第一步是搭建整个应用的基础框架。

Like, step one is build the entire, like, foundation of the app.

创建所有目录、文件夹和文件，注册应用，编写Android清单文件等等。

Create all the directories, create all the folders and files, and and register the app, create like a Android manifest and so on.

你可能需要事先了解一点Android知识，不过说实话，就算完全不懂Android应该也能完成。

You have maybe you need to know a little bit about Android before, but to be honest, like you can probably also do it without knowing anything about Android first.

这就是你开始的方式。

And then so that's how you start.

你一步一步来。

You do it step by step.

最重要的是你要有一个待办事项清单，因为大语言模型会在过程中迷失方向，你需要某种方式让它们重新聚焦到你想要完成的任务上。

The most important thing is that you have a list of things that you want to work through because LLMs get lost along the way, and you need some way to bring them back to the task that you want to make them work on.

这就是整个过程，我花了几个小时反复调试，偶尔会查看代码。

So that was the process, and I did this for a couple of hours back and forth and back and forth looking at the code sometimes.

但最初大部分时间，我甚至没有真正看过代码本身。

But most of the time in the beginning, I didn't even really look at the code itself.

我只是看看结果，看是否能编译通过。如果编译失败，大语言模型自己会意识到错误并尝试修复。

I'm just like, you know, looking at the results, whether it compiles or not, and if it doesn't compile, the LLM kind of knows that it made a mistake itself and tries to fix it.

就这样花了半天左右的时间。

So, half a day or so goes by doing that.

后来某个时刻，我得到了这个具备基本功能的可用版本，虽然还没有私信功能，但已经有了比如安卓版的群聊功能。

And then at some point, I had, this this working version with the basic features, like no no direct messages yet, but like the group chat for example, for Android.

它也开始变得很像iOS应用，因为我要求它做到完全一致：使用相同的配色方案、相同的字体等等。

And it also started to look quite similar to the iOS app because I also told it like make it look exactly the same, use the same color scheme, use the same fonts, and so on to make it look the same.

最终你得到了这个基础版本，之后就可以持续改进它。

And you end up with this basic version, and then from there on, you keep improving it.

所以你要一步一步循序渐进。

So you do step by step by step by step.

你需要明确目标方向，但只要能清晰表达需求，大语言模型大多时候都能完成任务。

You kind of need to know where you want to go, But as soon as long as you can express this in a meaningful way, the LLM most of the time can do the can get the job done.

这就是我最初开发BitChat的过程。

So that's how that's how I built BitChat, in the beginning.

这就是为什么进展如此迅速，因为我基本上完全依靠直觉，只是观察事物的发展方向，并试图将其引导至正确的轨道。

That's why it went so so quickly because I was basically just 100% vibes and just observing what the thing is doing and trying to kind of steer it in the correct direction.

请注意，在我迄今为止完成的所有氛围驱动的项目中（过去一年里我做了很多这类项目），你会发现随着项目成熟，氛围驱动与手动编码的比例会逐渐下降。

Now mind you, with every Vibe coated project that I've done so far, and I've done many of them like in the in the last year, and is what you see is that the proportion of Vibe against manual coding goes down as the project matures.

项目会变得越来越复杂。

It gets more and more complex.

随着时间的推移，大语言模型会犯越来越多的错误，而由于项目增长，上下文也会变得越来越庞大。

The LLMs tend to make more and more mistake over time, and the context gets larger and larger because the project grows.

所以随着项目推进，事情只会变得越来越复杂。

So things get just get more complicated as you go.

因此，需要通过手动修复进行干预的需求会与日俱增。

So the need to intervene by fixing things by hand just grows over time.

还有一点我想提到的是噪声加密。

One more thing I want to mention here is noise encryption.

BitChat采用了一个非常巧妙的握手模式，叫做Noise XX握手。

So BitChat uses a pretty nifty handshake pattern called Noise x x handshake.

它能提供前向保密性，意思是即使有人今天获取了你的手机，也无法读取你明天将要发送或昨天已经发送的所有消息。

It gives you forward secrecy, is basically saying like if someone, gets access to your phone today, they cannot read all the messages that you're going to write tomorrow or that you have written yesterday.

这就像是确保即使你今天被入侵，你的信息在时间轴前后都能保持机密。

So kind of like make sure that even if you get compromised today, that you have secrecy forward and backwards in time.

这部分功能使用了Noise XX握手协议。

And, so that used the Noise XX handshake.

而这一部分，举例来说，就完全是依靠大量手动编码完成的。

And that part, for example, was just heavily, heavily manual coding.

一旦技术变得超级先进，而大语言模型难以从互联网获取知识时，你就得重新变回程序员角色，而不仅仅是项目经理。

So as soon as you get like super advanced and there is not much for, the the LLM to draw on knowledge from the Internet, then you need to go in there and become like a coder again instead of being just a project manager.

但正如我所说，工作量尤其是手工工作量往往会随时间增加。

But yeah, so as I said, like the the amount of work, manual work just goes up, usually goes up over time.

特别是如果你不够谨慎，随着项目发展，你可能会把自己逼入死角——项目不断膨胀臃肿，最终变成一团乱麻，很难再让它恢复可用状态，或是清理掉大量无用代码的复杂混乱局面。

And especially if you don't, you know, if you if you don't be like if you aren't careful, especially over time as the project grows, you can also maneuver yourself into a dead end where the project just grows and grows and gets bloated and bloated to a point where it just becomes like a complete mess, and it becomes extremely hard to put it into a state again where where it just, you know, it works again or looks nice again or isn't like a complicated mess with lots of dead code that isn't used and so on.

所以我觉得这是你作为氛围程序员随着经验增长会学到的东西。

So that is something I think you learn as a vibe coder as you gain more experience.

但最初阶段完全是靠感觉和自动化，这也是为什么任何人都能做到的原因。

But like the first steps is fully vibe, fully automated, and that's also something that that's also the reason why anyone can do this.

尤其是我，首先我连安卓开发都不是，更不是前端开发。

Especially me, I'm not a front end developer even like I'm not an Android developer first of all, but I'm not even a front end developer.

我通常深陷机房战壕，最多就是和命令行界面打交道，完全不擅长调整按钮位置这类工作。

I'm usually deep in the trenches in the machine room, like interacting with command line interfaces at the best and not really good at like moving buttons around and so on and so forth.

我甚至不知道安卓上那些是怎么运作的。

And I don't even know how that works on Android.

现在我知道了。

At this point, I do know.

但在此之前，我确实不太理解安卓前端开发的具体机制，特别是用Kotlin开发。

But before that, I didn't really understand exactly how Android front end development works, especially with Kotlin.

所以我学会了所有那些。

So I learned all of that.

在不知情的情况下，你其实也能完成很多事情。

Without knowing, you can just do things.

这就是我要传达的信息。

That's my message.

就像你可以直接做事，而且现在你能做的事情比一年前多得多。

Like you can just do things, and today you can do things a lot more than just a year ago.

所以，这就是这个过程。

So that that's that's the process.

再稍微专业一点地说，我认为更有价值的经验是——我想分享给听众们，特别是那些计算机科学家听众们。

And a slightly more nerdy take, but I think, even more valuable, lesson that I had there is the following that I wanna share with the listeners, especially the the computer scientists that listen.

正如你所说，你知道有两个版本的聊天应用。

So, as you said, as you know, there are two bit chat apps.

iOS版本是由Jack开发的。

There's iOS version, by Jack.

它是用Swift编写的。

It is written in Swift.

Swift是iPhone的原生编程语言。

That is the native language for iOS, for iPhones.

还有用Kotlin编写的Android版本。

And there's the Android version that is written in Kotlin.

Kotlin是Android的原生编程语言。

That is the native programming language for Android.

所以这是两个完全独立的项目，使用完全不同的编程语言。

So these are two completely separate projects that have completely different programming languages.

我不是Kotlin开发者，我猜Jack也不是Swift开发者，但我们都能用不太熟悉的编程语言进行开发。

And I'm not a Kotlin dev, and I'm just going to assume that Jack is also not a Swift dev, but we were both able to build in programming languages that we're not very familiar with.

显然，我们都知道如何编程。

Obviously, we both know how to program.

所以，你可以借助AI使用一种你从未用过的编程语言。

So, you can just you can use a programming language that you've never used before by relying on AI.

现在到了最巧妙也最有趣的部分，我认为这将带来行业变革。通常当你开始开发一个应用，并希望它同时能在iOS和Android上可用时，如果你是小型团队——比如初创公司，只有三五个成员，刚拿到种子轮融资，银行账户里有几十万美元，想要打造这个MVP并推向世界，同时覆盖iOS和Android平台。

And now here comes the nifty and super interesting part, and I think that's going to change kind of like be an industry changing thing, is usually when you start making an app and you want it to be available on iOS and on Android at the same time, and you're a small team, let's say you're like a startup, you have like three people, you have maybe five people, and you just got your seed round, and you have a couple $100 on your bank account, and you want to build this MVP and you want to push it out to the world, and you want it to be available for iOS and Android.

通常的做法（这很常见）是先开发一个跨平台应用。

What you usually do, and this is very common, is you build a cross platform app first.

因为团队规模小、资源有限，时间也很紧张，你们遵循的是最小最大原则。

Because your team is small and your resources are limited, and the time that you have is also very limited, what you do is like you get the you do the mini min max principle.

就是用最少的工作量，获得最大的成果——通常就是开发跨平台应用。

You're like, least amount of effort, a maximum amount of result is usually a cross platform app.

这是什么意思呢？

What's that what does that mean?

市面上有些编程语言能同时在iOS和Android上运行。

For for there are some programming languages out there that work on iOS and Android at the same time.

比如Kotlin Multiplatform、Flutter或React Native。

So these might be Kotlin Multiplatform or Flutter or React Native.

这些都是比较流行的选择。很多初创公司会先用这些语言，由极小的团队（可能只有一两个人）基于单一代码库开发应用，然后编译成Android和iOS版本并发布产品。

Those are kind of the popular ones that many companies that get started, they use these languages first, and they built the app in a very small team that could be a single person, maybe two people working on a single app, a single code base, and then you compile it to both Android and iOS, and then you ship your product product.

如果发展顺利获得更多融资，团队就会扩大。

And if you grow and if you're successful, then you get more funding, you grow your team.

可能在半年或一年后，你们会再次坐下来讨论：'好了各位，现在我们要认真对待了。'

And then at some point, maybe a half a year, maybe a year later, you you sit down again and say, now guys, okay, we're now getting serious.

我们取得了一些进展。

We have some traction.

现在让我们同时开发一个原生Android应用和一个原生iOS应用，因为我们已经有了一支大约20人的团队。

Let's now build a native Android app and a native iOS app side by side because now we have like a team of 20 people or something.

通常，在氛围编程之前，你会选择那条路线。

So typically what you would have done before vibe coding is to go that route.

首先开发一个多平台应用，壮大公司，然后构建两个原生应用，再推向全球。

First, a multi platform app, grow the company, then build two native apps, and then ship that to the world.

比如，想想Uber和许多其他公司正是这么做的。

This is what, for example, think Uber and like so many companies do exactly that.

根据这次经验，我们刚刚认识到这方面的情况正在发生巨大变化。

Now what we've just learned, at least with this experience, is that things are changing a lot in that regard.

所以你现在可以像我们Pitchat项目那样开始：先由一个人独立开发iOS应用，完全不用考虑Android，只发布iOS版本。

So what you can do today is you could start the way we did it with Pitchat is first, one person goes ahead and builds an iOS app just completely by themselves, completely disregarding Android, just puts out an iOS app.

然后第二个人（甚至可以是同一个人）——不需要是Android开发者——把完成的软件输入大语言模型，让它直接翻译成原生Android代码。

And then a second person or even the same person, doesn't need to be an Android app Android developer, can take this finished software and just put it into an LLM and say, now translate it to native Android.

原本需要公司投入20人或10人同时开发和维护两个应用、保持功能同步的工作，现在基本上只需1到2人就能完成。

And what would have taken like 20 people in your company to maybe or 10 people in your company to build two apps and to maintain two apps at the same at the same time, keep them all with feature parity at the same time, now works with essentially either one person or maybe two per two people.

因此，开发原生应用的门槛已经大幅降低。

So the the the barrier to entry to build native apps has dropped significantly.

这仅仅是因为我们现在共享的语言不是React Native，而是英语。

And this is just because the the language that we share now is not React Native, but it's English.

我们双方都说英语，所以可以直接把英语编译成Swift或Kotlin代码。

Both of us can speak English, so we can just both compile English to either Swift or to either Kotlin.

这就是我们最终拥有Pitchat两个原生实现版本的原因。

And so that's how we ended up with two native implementations of Pitchat.

我认为这实际上是一个游戏规则的改变者。

And I think this is actually like this is the game changer.

对于那些不了解软件开发的人来说，这可能听起来不太重要，但对于如今在企业工作、听到这些消息的人来说，他们应该会对此感到非常兴奋，甚至可能因此改变他们构建应用程序的计划，因为现在你可以让事情变得简单得多。

This might be like sound a bit unimportant for those who don't know how you know, who are not into software development, but for those who are working in companies today who who hear this will like, should be super excited about this and maybe also change their plans of how they want to build their apps because at this point, you can, like, literally just do things so much easier.

特别说明——我保证我的独白快结束了——尤其是当你使用共享库构建项目时。

Especially, now I'm going to be done with my monologue, I promise, but especially if you build something with a shared library.

假设我们现在要构建一个原生缓存钱包，同时支持Android和iOS平台。我们有一个用Rust编写的超棒库叫CDK。

Now let's say we want to build a cache wallet, a native cache wallet for Android and for iOS, And we have this amazing library called CDK that is written in Rust.

基本上，所有复杂的部分都已经用Rust完成了，而且已经经过充分验证。

So basically, all of the complicated stuff is already done in Rust and that's been cared for.

现在与其雇佣两个人分别开发Android和iOS应用，我更可能选择不招新人，而是直接使用现有的Rust库，围绕它构建极简的Kotlin层和Swift层——这些层只处理UI部分，几乎不涉及其他逻辑。这样就能用平台原生语言打造出性能极致、外观和体验都完全原生的应用，而过去这通常需要庞大的开发团队才能实现。

Now instead of hiring two people that one building an Android app and one person building an iOS app, what I would do today is probably hire no one and just take the Rust library that we have and build a very slim and thin layer of Kotlin and a thin layer of Swift around it that is just the UI and almost doesn't do anything else, just the UI, that is native in that platform's language, and have a super performant, super native looking, and like native feeling app that you usually would have, you know, needed a large larger team of of software developers for.

所以真的非常令人兴奋。

So extremely excited.

时代正在改变。

Things are changing.

我认为程序员们最应该对当前正在发生的变革感到警惕。

I think, like, programmers out there should be, most afraid of what's happening right now.

如果你现在没有使用这些工具，也没有在掌握它们的路上，那么不妨想想五年后的世界会怎样，想想五年后的初创公司会是什么模样，思考自己会是保住工作的人，还是被淘汰的一员——毕竟我们很多人都会失业。

If you're not using these tools today and you're not in the process of mastering these tools, then you know, just you know, think about what's going to be in five years and think about what a startup will look like in five years and whether, you know, whether you'll be one of the people who keeps their job or one of the people who will lose their job because many of us will lose their jobs.

这一点非常明确。

That much is clear.

一家公司必须在三个能熟练使用AI的绝地武士和20个仍完全依赖手动编码的庸才之间做选择，他们根本不在乎如何正确从事专业工作。

A company that has to decide between three Jedis being able to use AI really well or 20 midwits that, still rely on just complete manual coding, they don't care about like, the how to do your profession correctly.

他们在乎的是运营效率。

They care about the efficiency of the operation.

所以这对开发者来说也是个警示信号：学习这些工具，学会使用它们，尤其是现在就要学，因为两年后它们会变得截然不同。

And, so this is a kind of a warning sign as well to the developers is learn these tools, learn to use these tools, and especially learn it today because in two years, they will be very different.

这些工具会快速迭代，你应该保持前沿状态。

These tools will change very fast, and you should be like out there.

你应该掌握所有这些技术，让生产力成倍提升。

You should be on top of all these things and like increase your productivity by multiple x.

说实话，我以前不懂进入技术奇点意味着什么。

And you know, I didn't know what it means to enter the singularity.

我无法想象一个人的生产力能提升10倍是什么概念。

I couldn't imagine what it means that a person could 10x their productivity.

就像我无法理解阅读速度快10倍或思考速度快10倍的意义，但现在我写代码的速度确实能快10倍。

Like I my mind cannot comprehend what it means to, you know, I will never be able to read 10 times as fast or to think 10 times as fast, but I can build code now 10 times as fast.

这将改变一切。

And that's going to change everything.

我对此非常确信。

I'm pretty sure about that.

太棒了。

Love it.

哈哈。

Lol.

首先，我们有Eggie，mpub.cash应用的维护者和创建者，10,000聪。

First off first off, we have Eggie, maintainer and creator of mpub.cash app, 10,000 sats.

他说，Callie的bit chat端口让我彻底改变了vibe coding的看法。

And he says, Callie's bit chat port made me do a one eighty on vibe coding.

如果你清楚自己在做什么，vibe coding会让你更接近软件架构师的定位。

If you know what you are doing, vibe coding puts you more in the position of a software architect.

那些自包含的、能明确定义接口的模块，99%的情况下都能一次成功。

Modules that are self contained that you can clearly define an interface for are a one shot 99% of the time.

Eggie，我的Signal联系方式在odel.xyz上。

Eggie, I have my signal contact information is on odel dot x y z.

如果你想哪天来dispatch节目做客，我非常欢迎。

If you wanna come on dispatch sometime, I would love to have you.

我们的测试者数量已增至42,000组。

And we have the testers up to 42,000 sets.

他说Cashew和Nostril将把比特币支付推向新高度。

He says Cashew and Nostril will take Bitcoin based payments to the next level.