CRYPTO 101 - 第752集:加密货币是否已准备好应对人工智能和量子计算的风险? 封面

第752集:加密货币是否已准备好应对人工智能和量子计算的风险?

Ep. 752 Is Crypto Ready for AI & Quantum Computing Risks?

本集简介

In this episode of the Crypto 101 Podcast, David Schwed from Sovereign joins from the Out East Summit to explain why Near’s intersection of AI, privacy, security, and blockchain is so important for the future of crypto. He breaks down why privacy should be viewed as a byproduct of strong security, especially as AI prompts, sensitive data, PII, and MNPI increasingly move through systems controlled by major technology companies. The conversation covers quantum resistance, Bitcoin’s upgrade challenge, MPC vs HSM custody infrastructure, social engineering attacks, spoofed emails and phone calls, prompt injection, and why organizations need to understand AI security before giving agents too much access. Check out ShipStation and use my code crypto for a great deal: https://www.shipstation.com Check out Omaha Steaks and use my code BEEF for a great deal: https://www.omahasteaks.com Check out Scribe and use my code scribe.how/CRYPTO101 for a great deal: https://scribe.com Check out Quince: https://quince.com/CRYPTO101 Check out Shopify: https://shopify.com/crypto101 Get my #1 altcoin pick for this month. Get immediate access to my entire crypto portfolio for just $1.00 today! Get your FREE copy of "Crypto Revolution" and start making big profits from buying, selling, Get immediate access to my entire crypto portfolio.. just $1.00 today! Go here to get access: https://www.crypto101insider.com/cryptnation-directm6pypcy1?utm_source=Internal&utm_medium=YouTube&utm_content=Podcast&utm_term=20250916 Get your FREE copy of "Crypto Revolution: Your Guide To The Future of Money". In this book, I reveal how to make (and keep) a fortune during this crypto bull run! http://www.cryptorevolution.com/free?utm_source=Internal&utm_medium=YouTube&utm_content=Podcast&utm_term=20250916 Chapters 00:00 Intro 01:50 - Privacy as a byproduct of strong security 04:50 - Why Bitcoin’s quantum upgrade is a governance challenge 06:50 - Why quantum risk is solvable 12:00 - Fake Google emails and recovery-flow scams 15:20 - AI security and misuse inside organizations 18:15 - How AI can make dangerous mistakes Subscribe to YouTube for Exclusive Content: https://www.youtube.com/@crypto101podcast?sub_confirmation=1 Follow us on social media for leading-edge crypto updates and trade alerts: https://twitter.com/Crypto101Pod https://instagram.com/crypto_101 *This is NOT financial, tax, or legal advice* Boardwalk Flock LLC. All Rights Reserved ▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬ Fog by DIZARO https://soundcloud.com/dizarofr Creative Commons — Attribution-NoDerivs 3.0 Unported — CC BY-ND 3.0 Free Download / Stream: http://bit.ly/Fog-DIZARO Music promoted by Audio Library https://youtu.be/lAfbjt_rmE8 ▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬ Our Sponsors: * Check out Omaha Steaks and use my code BEEF for a great deal: https://www.omahasteaks.com * Check out Quince and use my code quince.com/CRYPTO101 for a great deal: https://www.quince.com * Check out Scribe and use my code scribe.how/CRYPTO101 for a great deal: https://scribe.com * Check out ShipStation and use my code crypto for a great deal: https://www.shipstation.com * Check out ShipStation: https://www.shipstation.com * Check out Shopify and use my code shopify.com/crypto101 for a great deal: https://www.shopify.com Advertising Inquiries: https://redcircle.com/brands Privacy & Opt-Out: https://redcircle.com/privacy

双语字幕

仅展示文本字幕,不包含中文音频;想边听边看,请使用 Bayt 播客 App。

Speaker 0

你知道当你翻衣柜找衣服的时候,总有那么几件是你一直拿起来穿、反复穿的衣服。

You know when you look in your closet for clothes, we always have those pieces that you keep grabbing for and wearing on repeat.

Speaker 0

那件百搭的毛衣、那条完美的裤子,或者那件你总是最先拿起来的T恤。

The sweater that goes with everything, the perfect pair of pants, or that tee you somehow always reach for first.

Speaker 0

这些就是让你穿衣变得轻松的单品,而这正是Quince最擅长的。

Those are the pieces that make getting dressed easy and that's what Quince does best.

Speaker 0

Quince专注于高品质的基础款服饰,采用优质材料,比如100%蒙古羊绒、有机棉和美利奴羊毛。

Quince focuses on high quality wardrobe staples made with premium materials like 100% Mongolian cashmere, organic cotton, and merino wool.

Speaker 0

这些款式经典、百搭,设计成你每天都会依赖的可靠单品。

The styles are timeless, versatile, and designed to become dependable pieces you reach for every day.

Speaker 0

各位,夏天结束了,秋天马上就到了,秋天是我最喜欢的穿衣季节,因为九月初天气还暖和的时候,你还能穿那些轻薄、透气、合身的T恤和薄裤子,但天气开始转凉,你需要毛衣,需要提升你的穿搭,重新整理衣柜,迎接深秋和初冬。

Guys, summer is over and fall is right around the corner, and it's my favorite time of year for clothes in the fall because you still get some of that light, breathable, fitted tees and light pants that you want to wear when it's early September and still warm, but it starts to transition into that colder weather where you're going need some sweaters, you're going need to up your game and reorganize your closet as we transition into the late fall early winter.

Speaker 0

Quince是你买衣服的首选之地。

Quince is your place to go for all things clothes.

Speaker 0

我喜欢他们的亚麻制品。

I love their linens.

Speaker 0

我们整个夏天都在聊这个。

We talked about it all summer.

Speaker 0

整个夏天在希腊都在用Tivo的亚麻布,但现在进入秋天了,我们可能要找一些T恤和裤子,然后等天气一转冷,我回北方过圣诞假期的时候,肯定会买他们家的100%蒙古羊绒毛衣,我超级喜欢。

It was Tivo linens out in Greece all summer long, but now as we head into the fall, we're looking for maybe some more T shirts, some pants, and then again, once the weather turns and I go up north for the Christmas holiday season, I'll definitely be doing their 100% Mongolian cashmere sweaters that I love so very much.

Speaker 0

朋友们,Quince真的很棒。

Guys, Quince is amazing.

Speaker 0

他们的网站也特别好用。

The website is awesome.

Speaker 0

我之前也提过。

I've talked about it before.

Speaker 0

他们不只是卖衣服。

It's not just clothing.

Speaker 0

Quince把同样的理念用在他们做的所有东西上,从酒店品质的床品和卫浴用品,到厨房必需品,还有设计贴心的家具。

Quince brings the same approach to everything they make, From hotel quality bedding and bath to kitchen essentials and thoughtfully designed furniture.

Speaker 0

他们做的都是设计精良、适合日常生活的用品。

They make well designed pieces for everyday living.

Speaker 0

我经常提到它。

I talk about all the time.

Speaker 0

他们做每件事都很到位。

They do everything the right way.

Speaker 0

在Quince找到你秋季的新宠吧。

Find your next fall favorites at Quince.

Speaker 0

下载Quince应用获取专属优惠,或者访问quince.com/crypto10one。

Download the Quince app for app exclusive offers or go to quince.com/crypto10one.

Speaker 0

享受免费配送和365天退货服务,现在加拿大和英国也支持了。

Get free shipping on your order and three hundred and sixty five day returns now available in Canada and The UK too.

Speaker 0

网址是quince.com/crypto10one。

That's quince.com/crypto10one.

Speaker 0

别再等待许可才开始你的生意。

Stop waiting for permission to start your business.

Speaker 0

你的想法只差一步,Shopify能帮你轻松实现。

Your idea is one step away, and Shopify can make it easy to make it happen.

Speaker 0

我们在crypto one zero one就是这么做的,用Shopify卖我们的书,之后的事情就顺理成章了。

That's what we did at crypto one zero one and use Shopify to sell our book, and the rest is history.

Speaker 0

我喜欢Shopify的一点是,它让创业不再那么吓人。

The thing I like about Shopify is it takes the scary part out of starting.

Speaker 0

你不用坐在那儿琢磨网站、支付、结账、库存这些琐碎的事情,这些往往在还没上线前就把人吓退了。

You're not just sitting there trying to figure out a website, payments, checkout, inventory, all these little things that stop people before they even launch.

Speaker 0

说到搭建店铺,我觉得最大的问题就是流程上的阻碍。

When I think about building a storefront, the biggest thing is friction.

Speaker 0

你希望顾客一进你的页面,就喜欢上你卖的东西,然后能顺利下单,不会卡住。

You want someone to land on your page, like what you're selling and be able to buy without getting stuck.

Speaker 0

这正是Shopify的强项。

That's where Shopify is huge.

Speaker 0

Shopify的结账系统就是为了帮助更多顾客完成购买而设计的。

Shopify checkout is built to help more customers finish their purchase.

Speaker 0

而且他们再次光顾时,信息都已经保存好了。

And when they come back, their details are already saved.

Speaker 0

一键就搞定了。

One tap and they're done.

Speaker 0

最棒的是,Shopify 支持全球数百万家企业,从你熟悉的大品牌,比如美泰和Gymshark,到刚刚起步的小企业。

The cool part is Shopify powers millions of businesses worldwide from huge brands that you know, like Mattel and Gymshark to small businesses just getting started.

Speaker 0

所以无论你是想打造下一个大品牌,还是推出你的第一个产品,基础设施都已经准备好了。

So whether you're building the next big brand or launching your first product, the infrastructure is already there.

Speaker 0

一旦这部分处理好了,你就可以专注于真正的工作:打造品牌、吸引关注、把人们带到你的店铺,让生意成长起来。

Once that part is handled, you can focus on the real work, building the brand, getting attention, and bringing people to your storefront so the business can grow.

Speaker 0

当有人真正购买的那一刻,想法就变成了现实。

That moment when someone actually buys, that's when the idea becomes real.

Speaker 0

有了Shopify,你的想法和真正的生意之间没有任何障碍,所以去实现它吧。

With Shopify, nothing stands between your idea and a real business, so go make it one.

Speaker 0

今天就在shopify.com/crypto10one开始你的免费试用吧。

Start your free trial today at shopify.com/crypto10one.

Speaker 0

那就是在shopify.com/crypto10one的免费试用。

That's a free trial at shopify.com/crypto10one.

Speaker 1

好的,各位。

All right, everybody.

Speaker 1

我们回到了由TIE赞助的Out East峰会现场。

We're back at the Out East Summit here sponsored by the TIE.

Speaker 1

非常高兴能邀请到David Schwed先生。

Really excited to have Mr.

Speaker 1

他来自Sovereign公司。

David Schwed joining us from Sovereign.

Speaker 1

谢谢。

Thank you.

Speaker 1

你怎么样?

How you doing?

Speaker 1

到目前为止,峰会还愉快吗?

Enjoying the summit so far?

Speaker 2

我挺好的。

I'm good.

Speaker 2

今天天气真好。

This is beautiful weather today.

Speaker 2

希望比昨天好,但现在已经很美了。

Hopefully it'll be better than yesterday but it's gorgeous.

Speaker 1

对,我很喜欢。

Yeah, love it.

Speaker 1

是的。

Yeah.

Speaker 1

看看Sovereign、Near Foundation、DAT,还有数字资产Tell,你为什么选择Near?

Look Sovereign, Near Foundation DAT, digital Tell asset us why you chose Near?

Speaker 1

因为现在有成千上万种加密货币,你知道的,本来可以做比特币,也可以做以太坊。

Because there's a lot of, there's millions of different cryptos, you know, could have done Bitcoin, could have done Ethereum.

Speaker 1

你深入研究了技术栈,然后看中了Near,我对这个决定很好奇。

You kind of went down the stack and looked at Near and I'm curious about that decision.

Speaker 2

是的,我喜欢他们在隐私和区块链交叉领域做的事情。

Yeah, I love what they're doing with the intersection of privacy and the blockchain.

Speaker 2

我不一定把它们看作只是另一个第一层区块链。

I wouldn't look at them necessarily as just another layer one.

Speaker 2

再说,市面上有大量的第一层区块链在争夺每秒交易量之类的指标。

Again, there's tons of layer ones out there competing against TPS and etcetera.

Speaker 2

但从智能体商务的角度来看,AI和隐私与区块链融合的重要性,这才是真正吸引我的地方。

But the understanding from an agent commerce perspective to the convergence of the importance of AI and privacy integrated with the blockchain is really what drew me over to it.

Speaker 1

对。

Yeah.

Speaker 1

而且你知道,你在加密领域有很棒的背景,在Galaxy工作过,在纽约梅隆银行工作过,现在又在Sovereign和Near,隐私,我觉得这绝对是未来的方向,因为AI和加密,目前很多都不是公开的,但公有区块链本质上就有那种痕迹,对吧,那种审计轨迹,人们可以说,密码学验证过的,虽然是假名的,但它仍然在那里。

And you know, you've had a great background in crypto working at Galaxy, working at BNY Mellon, and now at Sovereign and Near, privacy, I mean this is to me definitely gonna be the way of the future because with AI and with crypto, so much of it currently is not public, but public blockchains in nature have that, you know, that trail, right, that audit trail where people can say, you know, cryptographically verified, even though it's pseudonymous, it's still out there.

Speaker 1

而且有了AI,很多人都在向这些聊天机器人询问他们最深层、最黑暗的秘密。

And with AI, a lot of people are querying these, you know, chatbots with their deepest, darkest secrets.

Speaker 1

是的。

Yes.

Speaker 1

那猜猜谁能接触到这些信息?

And guess who has access to that?

Speaker 1

OpenAI、Anthropic,你随便说一个。

OpenAI, Anthropic, you know, you name it.

Speaker 1

还有微软,或者随便一家科技巨头,你觉得隐私这部分为什么这么重要?

You know, Microsoft, you insert Mag7 company here, and why to you is the privacy component so important?

Speaker 1

因为这不是为了隐私而隐私。

Because it's not privacy, you know, for the sake of privacy.

Speaker 1

还是说,这是出于某种哲学上的原因?

Is it privacy for a certain philosophical reason?

Speaker 2

说实话,从纯技术角度看,我关注的重点其实不是隐私。

Yeah, mean honestly, from a pure technical perspective, like my approach of this really isn't from a privacy perspective.

Speaker 2

更多是从安全的角度来看。

It's more from a security perspective.

Speaker 2

安全。

Security.

Speaker 2

隐私只是良好安全性的一个很好的副产品。

Privacy is a wonderful byproduct of good security.

Speaker 2

我大概就是这么看的。

That's kind of how I view things.

Speaker 2

所以当我们看,就像你说的,AI和Anthropic、Gemini、Grok这些系统里进出的提示词,从监管的角度来看,确实有很多PII或者MNPI数据在流动。

So when we look at, to your point, the prompts going in and out of an AI and Anthropic and Gemini and Grok and whoever, you know, there's a lot of, from a regulatory perspective, you know, PII or MNPI data that's going in and out.

Speaker 2

所以从安全的角度,我想保护组织免受数据泄露,防止国家行为者或威胁行为者接触到敏感信息,那些他们不该接触的信息。

So from a security perspective, I wanna protect organizations from data exfil, nation states or threat actors having access to sensitive information, information they shouldn't have access to.

Speaker 2

为了做到这一点,我们以特定的方式构建基础设施来确保这一点。

And in order to do that, we build infrastructure in a certain way that ensures that.

Speaker 2

就像我说的,这样做的一个很好的副产品就是隐私,因为最好的安全就是确保没人能访问它。

And like I said, a wonderful byproduct of that is privacy because the best security is ensuring that nobody can access it.

Speaker 2

所以通过这样做,我们保持了信息的私密性。

So by doing that we're keeping things private.

Speaker 1

有趣的是,这也和抗量子攻击有关。

And what's so interesting is it plays into quantum resistance as well.

Speaker 1

我想请你谈谈这个,因为我知道Near最近发布了抗量子攻击的更新,或者你想怎么称呼它都行。

And I want you to kind of talk about this because I know Near just recently had a release, an update, however you want to frame it, of quantum resistance.

Speaker 1

这是因为隐私的原因吗?

And is this because of privacy?

Speaker 1

因为我知道,关于量子,就像如果你曾经用你的公钥在比特币上签过一笔交易,那个公钥是公开的,大家都能看到,但之后他们可以反推出你的私钥。

Because I know with quantum it's like if you've ever signed a transaction on Bitcoin with your public key, that is sort of publicized and people can see that but then they could reverse engineer your private key.

Speaker 1

那你正在做的这个工作,它的抗量子特性是怎样的呢?

And you know, what is the quantum resistant nature of what you're working on?

Speaker 2

对,当然。

Yeah, sure.

Speaker 2

所以,怎么说呢,这个问题其实比加密货币更大。

So, you know, to kind of frame it, it's bigger than crypto, the issue.

Speaker 2

没错。

Sure.

Speaker 2

我们现在有一些加密算法,它们不是抗量子的,我觉得它们可以被量子计算破解。

So we have certain cryptographic ciphers that are not quantum, you know, I guess they can be broken with quantum computing.

Speaker 1

就像你看到的HTTPS,它也可能被量子计算破解,就像

It's like when you see HTTPS, that could be broken by quantum just as

Speaker 2

理论上,确实很容易。

easily as Theoretically, right.

Speaker 2

比如,如果使用了某些特定的加密算法,它们就可能被破解。

Like if the choice of using certain ciphers, they could be decrypted.

Speaker 2

所以,当我们从安全和隐私的角度来看这个问题时,有几个不同的基础点需要梳理一下。

So when we look at this for the lens of security and privacy, you know, there's a couple of different, you know, I guess foundational blocks to kind of unpack.

Speaker 2

首先,从隐私的角度来看,有一个概念叫“先存储,后解密”,这将会成为一个问题。

You know, first is from a privacy perspective, like there's this concept of store now, decrypt later, which is going to be an issue.

Speaker 1

对,我听说过。

Yep, I heard

Speaker 2

关于这个,很多国家都在这么做,你知道,我相信我们国家也在做,美国也在做,就是存储大量的加密数据,来自各个国家和地区的,希望有一天能解密这些数据,获取国家机密。

about So, a bunch of nation states, listen, you know, I'm sure our country does that as well, US does this as well, is We're storing tons and tons of encrypted data from nation states and from different geographic regions with the hopes of one day we'll be able to decrypt that data and have access to nation secrets.

Speaker 2

所以,这个问题我们无法改变过去。

So, that is an issue that we're not gonna be able to fix in the past.

Speaker 2

只能从今往后去解决。

It's only on a go forward basis.

Speaker 2

好的。

Okay.

Speaker 2

这个问题我们也得面对。

We're gonna have to deal with that as well.

Speaker 2

所以,这个问题是没法解决的。

So, like, you can't fix that problem.

Speaker 2

我们只能去考虑它。

That is just something we're gonna have to think about.

Speaker 2

所以,各个组织应该开始思考,如果过去二十年的数据不小心泄露了,我的组织会怎么样?

So, organizations should start to think through, if the last twenty years of my data somehow got leaked, what would happen to my organization?

Speaker 2

我觉得这是大多数组织和国家应该真正去做的练习。

So I think that's an exercise most organizations and nation states should actually be going through.

Speaker 2

展望未来,如果我们开始考虑加密的潜在影响,那还是数字签名。

Going forward, if we start thinking about what is the potential impact of crypto, it is, again, digital signatures.

Speaker 2

所以,如果我们用非抗量子密码来做数字签名,那么整个区块链的真实性和完整性,理论上就没了。

So if we're using a non quantum resistant cipher to do digital signatures, the entire blockchain, the veracity of it, the integrity of it, in theory, is gone.

Speaker 2

对。

Yeah.

Speaker 2

所以,这就是为什么对组织来说非常重要,不仅仅是加密货币,还有金融服务机构,你提到了HTTPS,还有你知道,所有其他协议和信息交换的方式,都需要具备抗量子能力。

So, that's why it's really important for organizations, not just crypto, financial services organizations, you mentioned HTTPS and you know, all of the other you know, protocols and ways in which exchange information to be you know, quantum resistant.

Speaker 1

我听说过这个概念,就是银行和比特币一样,在量子计算面前同样脆弱。

And I've heard this sort of concept, guess, this idea that, you know, banks are just as exposed to quantum as Bitcoin.

Speaker 1

但是,银行有一个CEO,他可以说,未来三年,这是你们的市场计划,这是你们的行动计划,他们可以从上到下发号施令,一键达成共识,因为他们说了算。

However, banks have a CEO at the top who could say, over the course of the next three years, here's your go to market plan, here's your action plan, and they could dictate from the top down and get consensus with the snap of a button because they're in charge.

Speaker 1

然后人们就会觉得,好吧,他们会处理的。

And people are like, okay, well they're gonna handle it.

Speaker 1

他们会进行量子升级,或者不管他们怎么称呼。

They're gonna get the quantum upgrade or whatever they're calling it.

Speaker 1

对于加密货币来说,这就困难多了,因为它是去中心化的共识,没有单一的决策者。

For crypto it's like a lot more difficult because you have decentralized consensus, you have no single decision maker.

Speaker 1

所以你已经看到像MicroStrategy和Coinbase这样的团体,成立了所谓的抗量子智库,或者类似的组织。

And so you've seen groups like MicroStrategy and Coinbase come up with quantum resistant think tanks let's call it, or groups.

Speaker 1

我们到底怎么才能让比特币具备抗量子能力?

How do we actually get a quantum resistant Bitcoin?

Speaker 1

因为你知道,我们可以说说以太坊,它稍微中心化一点,还有一些基金会,它们可以说,好吧,我们会制定计划,提出方案,我们知道验证者集合,等等等等。

Because you know, we could talk about Ethereum a little bit more centralized, other foundations that could say, okay well we're gonna come up with our plan, we're gonna propose it, we know the validator set, so on and so forth.

Speaker 1

但比特币,你知道,它占了合法加密货币市值的70%到80%,我觉得它可能是最难完成全面升级的那个。

But Bitcoin is really you know 80% of them or 70% of the market cap of legitimate you know cryptocurrencies and I feel like if that's the one that maybe might have the most difficult time getting its full upgrade.

Speaker 2

对,我的意思是,就像你说的,比特币的共识机制决定了它没有中心化权威,这正是我们喜欢它的原因。

Yeah, I mean like there's the consensus nature of it, to your point, there's no centralized authority which is why we like it.

Speaker 2

你知道,没有谁在发号施令说,这就是我们必须做的升级。

You know, that's dictating here's the upgrade that we're gonna have to do.

Speaker 2

对。

Yeah.

Speaker 2

从技术角度来看,这绝对是百分之百可以解决的。

From a technical perspective, it's absolutely 100% solvable.

Speaker 2

我觉得关键在于找到合适的各方和合适的人来推动协议升级,并处理节点基础设施的工作。

Think it's just around getting the right parties and the right people in place to put forth the protocol upgrades, work on the node infrastructure.

Speaker 2

再说,我们会有一些早期的签名,这些签名可能没法转到抗量子的钱包里,关于怎么处理这些,我觉得有几种不同的理论,可能最后得强制升级之类的。

Again, we're gonna have early signatures, if you will, that we're not gonna be able to move it to quantum resistant wallets, which I guess there's a couple of different theories on how to take care of those, potentially maybe force upgrades, etcetera.

Speaker 2

但从技术角度来看,这是可以解决的。

But it's solvable from a technical perspective.

Speaker 2

我觉得这更多是个治理问题。

I think it's just more of a governance issue.

Speaker 0

现在大家都在聊AI,但真正重要的AI是那种能帮你省时间、让你做出更明智决策、减轻你工作负担的AI。

Everybody is talking about AI right now, but the AI that actually matters is the kind that saves you time, helps you make smarter decisions, and takes work off your plate.

Speaker 0

如果你在经营电商生意,物流就是那种要么让你顺畅运转、要么让你完全卡壳的事情之一。

And if you're running an e commerce business, shipping is one of those things that can either keep you moving or completely slow you down.

Speaker 0

所以我喜欢ShipStation。

That's why I like ShipStation.

Speaker 0

它是一个为电商业务打造的端到端订单履行平台,它的AI是基于全球数十亿次发货中积累的几十年真实物流数据训练出来的。

It's an end to end order fulfillment platform built for e commerce businesses, and its AI is trained on decades of real shipping data from billions of shipments across the world.

Speaker 0

所以,不用再手忙脚乱地处理各种工具,ShipStation能帮你掌控全局。

So instead of juggling a bunch of different tools, ShipStation helps you stay in control.

Speaker 0

它能在库存不足时提醒你,推荐最佳的承运商和费率,自动化任务,帮助你在业务增长时节省时间。

It can alert you when your stock is low, recommend the best carrier selections and rates, automate tasks, help you save time as you grow.

Speaker 0

你还可以跨销售渠道同步库存,自动比价,提供品牌化的退货门户,并集成会计和CRM软件,所有这些都在一个平台上完成。

You also get inventory syncing across sales channels, automatic rate shopping, a branded returns portal, and integrations with accounting and CRM software all in one place.

Speaker 0

如果我在经营电商店铺,这正是我需要的工具,因为一旦订单开始涌入,你肯定不希望物流成为拖后腿的环节。

If I were running an e commerce store, this is exactly the kind of tool I'd want because once the orders start coming in, you don't want shipping to be the thing that's holding you back.

Speaker 0

超过一百万家商家信赖ShipStation来优化和扩展他们的物流。

Over 1,000,000 businesses have trusted ShipStation to optimize and scale their shipping.

Speaker 0

你越早切换,就越早开始节省时间和金钱。

The sooner you switch, the sooner you start saving time and money.

Speaker 0

今天就注册ShipStation,在shipstation.com使用代码crypto,即可免费使用六十天。

Get started with ShipStation today and get sixty days free at shipstation.com with code crypto.

Speaker 0

网址是shipstation.com,代码是crypto。

That's shipstation.com, code crypto.

Speaker 0

shipstation.com,代码crypto。

Shipstation.com, code crypto.

Speaker 0

税费另计。

Taxes and fees apply.

Speaker 0

你知道什么最拖慢团队进度吗?

You know what slows teams down more than anything?

Speaker 0

不一定是人员预算,甚至也不是工具本身。

It's not always the headcount budget or even the tools themselves.

Speaker 0

而是工作方式只存在于某个人脑子里,从来没有被真正记录下来。

It's when the way work gets done lives in someone's head and never actually gets documented.

Speaker 0

然后新同事加入,新平台上线,或者有同事离职,突然每个人做同一个流程的方式都不一样了。

Then someone new joins, a new platform gets rolled out or a teammate leaves and suddenly everybody's doing the same process in a different way.

Speaker 0

这正是今天的赞助商Scribe要解决的问题。

That's exactly the problem today's sponsor, Scribe, was built to solve.

Speaker 0

Scribe是一个工作流AI平台,财富500强中有94%的公司都在使用它。

Scribe is a workflow AI platform trusted by 94% of the Fortune 500.

Speaker 0

它能实时捕捉工作流程,并将其转化为一步步的文档,方便团队查看和分享。

It captures workflows as they happen and turns them into step by step documentation your team can review and share.

Speaker 0

你就像平常一样操作,Scribe 会一边进行一边生成指南,自动捕捉每一次点击、每一步操作和截图。

You just do the process like you normally would and Scribe builds the guide as you go, capturing every click, step, and screenshot automatically.

Speaker 0

因为我们的团队是远程办公,我确实遇到过好几次需要向同事重复解释同一个设置或流程的情况,总觉得这些第一次就该记录下来。

With our team being remote, I have definitely had moments where I've had to explain the same setup or process more than once to a colleague, and it always feels like something that should have been documented the first time.

Speaker 0

有了 Scribe,原本可能需要花几个小时编写和整理的内容,不到一分钟就能准备好分享。

With Scribe, what could take hours of writing and cleanup can be ready to share in under a minute.

Speaker 0

Scribe 还会自动从截图中隐去敏感信息,比如姓名、邮箱和账号。

Scribe also automatically redacts sensitive information like names, emails, and account numbers from screenshots.

Speaker 0

管理员可以在整个团队中强制执行这一功能,确保没有任何遗漏。

Admins can enforce that across the team so nothing slips through the cracks.

Speaker 0

另外,任何按照流程操作的人都能获得实时的屏幕指导,准确显示在实际工具中该点击哪里。

Plus, anyone following the process can get real time on screen guidance showing exactly where to click inside the actual tool.

Speaker 0

Scribe 不只是记录工作流程,它还会利用 AI 建议改进方法,帮你发现冗余步骤以及简化或自动化的机会。

Scribe doesn't just document workflows, it uses AI to suggest ways to improve them, helping you spot redundant steps and opportunities to simplify or automate.

Speaker 0

想看看 Scribe 能为你的团队带来什么,请访问 scribe.how/crypto101,并提及 crypto101,即可在部分套餐中免费使用 Scribe Capture 首月。

To see what Scribe could look like for your org, head to scribe.how/crypto101 and mention crypto101 for your first month of scribe capture free on select plans.

Speaker 0

这就是Scribe,scribe.how/crypto10one。

That's scribe, scribe.how/crypto10one.

Speaker 0

不知道你们怎么样,但我经常不知不觉就到了晚饭时间。

I don't know about you, but dinner sneaks up on me all the time.

Speaker 0

我打开冰箱,发现没有现成的蛋白质,这时候我也不想去商店,临时跑一趟,还得赌店里还剩什么肉或汉堡。

I'll open the fridge and realize I have no protein ready, and at that point, I'm not trying to go to the store, make a last minute grocery run, and gamble on whatever meats or burgers are left at the store.

Speaker 0

所以Omaha Steaks是个很好的冰箱储备选择。

That's why Omaha Steaks is a great freezer move.

Speaker 0

就像家里有个信得过的肉铺,在最需要的时候有高质量的蛋白质。

It's like having a trusted butcher at home and quality protein when you need it the most.

Speaker 0

Omaha Steaks用一百多年的时间赢得了口碑。

Omaha Steaks has earned its reputation with over a hundred years.

Speaker 0

和全国厨房里信赖的牛肉一样,直接送到你家门口。

It's the same beef trusted in kitchens across the country, shipped straight to your door.

Speaker 0

牛排馆的品质,但不用付牛排馆的价格。

Steakhouse quality without the steakhouse prices.

Speaker 0

你的家人会以为这顿饭花了好几个小时,但其实没有。

Your family will think dinner took hours, but it didn't.

Speaker 0

只要几分钟。

Just minutes.

Speaker 0

如果你还没试过,那“完美汉堡组合”就是最好的起点。

And if you never tried them, the burger perfection flight is the perfect place to start.

Speaker 0

你会得到20个优质牛排汉堡,用的是五种不同部位的牛排。

You get 20 premium steak burgers made from five different cuts of steak.

Speaker 0

四个菲力牛排汉堡,四个纽约客牛排汉堡,四个肋眼牛排汉堡,四个牛胸肉汉堡,还有四个西冷牛排汉堡。

Four filet mignon burgers, four New York strip burgers, four rib eye burgers, four brisket burgers, and four sirloin burgers.

Speaker 0

各位,这简直就是汉堡界的山寨币篮子。

Guys, this is the altcoin basket of burgers right here.

Speaker 0

如果我和布莱恩在节目里,我们肯定会五五分账,因为你知道我们夏天多喜欢烧烤。

If me and Brian were on the rundown, we would be divvying these up fifty fifty because you know how we like to grill in the summer.

Speaker 0

所以你得到的不仅仅是一盒汉堡,而是不同的风味、不同的口感,还有真正的牛排馆级别的多样性,你可以随时扔到烤架上,或者工作日快速搞定。

So you're not just getting a box of burgers, you're getting different flavor profiles, different textures, and real steakhouse style variety you can throw on the grill or cook up fast during the week.

Speaker 0

今天就去omahasteaks.com试试汉堡完美套餐吧。

Try the burger perfection flight at omahasteaks.com today.

Speaker 0

20个优质牛排汉堡只要79.99美元,使用优惠码beef还能再减35美元。

That's 20 premium steak burgers for just $79.99, plus get $35 off with the promo code beef.

Speaker 0

网址是omahasteaks.com,结账时输入优惠码beef立减35美元。

That's omahasteaks.com, promo code beef for $35 off at checkout.

Speaker 0

条款适用,详情请见网站。

Terms apply, see site for details.

Speaker 1

那你能不能再深入聊聊这些派对?

And if you could kinda like go a little deeper on on those parties, right?

Speaker 1

我们说的是矿工、开发者和交易所,还是需要拉某些特定的人进来,或者具体是怎么运作的?

Are we just talking about the miners and the developers and the exchanges, or are there certain individuals that need to be brought in on this, or how does that work?

Speaker 2

对,我觉得不是个人层面的。

Yeah, I don't think it's individuals.

Speaker 2

我觉得就像我们做过的任何升级,比如以太坊升级到2.0,或者比特币那边的Taproot,不管怎样,都会有协议升级。

Think it's like any upgrade that we've done, you know, like Ethereum, when we went to Ethereum 2 point o or Taproot or on the Bitcoin side, whatever, there's going to be a protocol upgrade.

Speaker 2

还是那几方。

It's the same parties.

Speaker 2

还是同样的流程。

It's the same process.

Speaker 2

我聊过的密码学家都说,我们还有好几年的时间。

Have cryptographers that I've talked to is like, we have years.

Speaker 2

这不是未来十二个月内就要面临的紧迫威胁。

This isn't like a dooming threat over the next twelve months.

Speaker 2

所以我们有几年时间来解决这个问题。

So we have years to figure this out.

Speaker 2

不是说我们现在就不该开始解决。

Not saying that we shouldn't be figuring it out now.

Speaker 2

这不会是一次仓促的、天要塌下来的升级。

This isn't going to be a rushed, oh my god, the sky is falling upgrade.

Speaker 2

这应该是一次精心策划、协调有序的升级。

It should be a very well planned and orchestrated.

Speaker 2

而且在我看来,这应该被视为某种关键基础设施,就像我们对待其他金融市场基础设施一样。

And if we look at, in my opinion, this should be viewed as somewhat critical infrastructure, just like we do with other financial market infrastructure.

Speaker 2

所以我觉得在某个阶段,涉及的银行、清算所,以及那些更成熟的机构,都知道怎么运营这些东西。

So I think at a certain point, banks that are involved, you have clearing houses that are involved, and you have more established institutions that know how to run these things.

Speaker 2

我觉得这是一个完全可以解决的问题,我个人并不担心。

I think this is a very solvable and I'm not personally worried issue.

Speaker 1

对,这确实很有意思。

Yeah, no it's fascinating.

Speaker 1

而且从钱包基础设施和自主托管的角度来看,他们肯定得推送一堆固件更新之类的东西,但你知道,对于正在观看的普通散户投资者,或者那些在保护自己加密货币的人来说,你有没有什么建议,关于未来保护自己加密资产的最佳实践?

And kind of from the wallet infrastructure, self custodial standpoint, I'm sure they'll have to push a bunch of firmware updates and all that kind of thing, but you know, do you have a recommendation on, you know, for the average retail investor who's watching or the person who's securing their crypto, like a best practice for the future of security with their own crypto assets?

Speaker 2

这是个好问题,对吧?

It's a good question, right?

Speaker 2

我觉得世界分为两大阵营:MPC 和 HSM。

Like I think the world is divided into two camps, MPC and HSMs.

Speaker 1

解释一下这两个。

Define those.

Speaker 2

哦,当然可以。

Oh yeah, sure.

Speaker 2

就是多方计算,比如你提到的M事件签名、阈值签名,就是三分之二的签名。

So like multi party computation, so if you're talking about the M event signing, threshold signing, so two out of three.

Speaker 2

这不是多重签名。

It's not multi signature.

Speaker 2

它就是一个签名。

It's like one signature.

Speaker 2

再说一遍,这并不完全是它的工作原理,但最简单的解释方式就是,想象一下某个东西被拆分成不同的部分。

Again, this is not exactly how it works but just the easiest way to explain it is pretend something is broken up into different parts.

Speaker 2

每一部分都需要签名,它们合在一起,然后交易就发出去了。

Each one of those have to sign, they come together and it goes out.

Speaker 2

这就是MPC。

That's MPC.

Speaker 2

比如Fireblocks、Utila、Defense,以及一些主要的托管服务提供商。

So think like Fireblocks, the Utila, Defense, some of the major custodial providers that are out there.

Speaker 2

比如Silence Labs、Sapir,他们提供MPC软件开发工具包,供组织使用。

Silence Labs, Sapir, for example, they provide MPC software development kits that organizations use.

Speaker 2

另一边呢,还有HSM。

Then on the other side you have HSM.

Speaker 2

HSM就是硬件安全模块。

So HSMs are hardware security modules.

Speaker 2

这是银行用了好几十年的一项技术了。

These are decades old technology that banks use today.

Speaker 2

想象一下,就是那种加固的箱式设备,放在数据中心的机架上。

Think fortified box appliances that are racked and stacked in data centers.

Speaker 1

还有持枪的保安看守。

Guarded by men's with guns.

Speaker 2

确实,很多地方都有持枪保安看守。

Guarded by men's with guns in many respects, yeah.

Speaker 2

这种设备是专门设计的,防篡改,如果你试图打开箱子,密钥材料就会被销毁。

This is purpose built where it's tamper resistant, where if you try to open up the box the key material gets destroyed.

Speaker 1

哦,

Oh,

Speaker 2

哇。

wow.

Speaker 2

这就是两大阵营,而且它们各自都有不同的升级路径,可以这么说。

Those are the two camps, and each one of them have their own different upgrade paths, if you will.

Speaker 2

比如HSM,它们已经部署了抗量子密码。

Like HSMs as an example, have quantum resistant ciphers already that are rolled out.

Speaker 2

所以很多组织都在为HSM做好准备。

So a lot of organizations are becoming HSM ready.

Speaker 2

在MPC这边,很多不同的团队都在努力让MPC也能抗量子。

On the MPC side, there's a lot of work being done by different parties to make MPC also quantum resistant ready.

Speaker 2

所以两个阵营都在积极朝着抗量子的方向发展。

So both camps are very actively working on becoming quantum resistant.

Speaker 1

你显然是个安全专家。

You're obviously a security expert.

Speaker 1

我想知道,在不吓到家里人的前提下,是什么让你夜不能寐?

I wanna know, without scaring everybody at home, what kinda keeps you up at night?

Speaker 1

对,你知道,除了可能哭闹的孩子之类的?

Yeah, you know Besides like maybe crying kids or anything?

Speaker 2

说实话,是社交工程。

Yeah, honestly, social engineering.

Speaker 1

社交工程。

Social engineering.

Speaker 2

我确实认为,很多很多很多的人和机构并不了解社交工程背后的复杂程度。

I do think that many, many, many people and organizations don't understand the level of sophistication that goes on with social engineering.

Speaker 2

比如,我每天在加密货币领域看到这些,已经习以为常了。

Like, I take it for granted for seeing it every day, being in crypto and watching.

Speaker 2

再说,我不想点名批评某些机构,但有两个例子我很喜欢讲。

Again, I don't wanna pick on parties and say the names, but there was a very, there's two that I love talking about.

Speaker 2

其中一个比较近,某个机构在会议上被真实地针对了。

One was fairly recent where a particular organization was targeted in real life at a conference.

Speaker 2

而且,结果据说是朝鲜那边干的。

And again, it turned out to be allegedly DPRK, North Korea.

Speaker 2

但接近他们的那些人并不是朝鲜族。

But the individuals that approached them were not of North Korean descent.

Speaker 2

他们不是韩国人。

They didn't come from Korean.

Speaker 1

我听说过这件事,

I heard about this,

Speaker 2

对,对。

yeah, yeah.

Speaker 2

他们是在一个会议上接近他们的。

They approached them at a conference.

Speaker 2

他们和某个组织的人建立了关系。

They established a relationship with individuals at a certain organization.

Speaker 2

他们想成为设计合作伙伴。

We want to become design partners.

Speaker 2

他们以善意与他们合作了好几个月,建立了这种关系。

They worked with them in goodwill for months upon months upon months, establishing this relationship with them.

Speaker 2

到了某个时候,他们把恶意代码植入到他们共享的某个代码库中,控制了机器,然后带着数亿美元离开了。

And at a certain point is when they drop the payload into a particular code base that they were sharing with them, own the machine, and walked out the door with hundreds of millions of dollars.

Speaker 2

你知道,当Just

You know, when Just

Speaker 1

太疯狂了。

so crazy.

Speaker 2

组织里的安全意识培训,我们告诉人们的社会工程学攻击,就像,哦,你会收到一封让你重置密码的邮件。

Security awareness training at organizations, the social engineering that we're telling people is like, oh, you're gonna get an email telling you to reset your password.

Speaker 2

我们没有向组织展示那些真正复杂的社会工程学攻击手段。

We don't walk organizations through the really sophisticated social engineering attempts that are out there.

Speaker 2

这真的很难防范,因为你是在控制组织内某个有权限做特定事情的人。

And that's really hard to protect against because you're owning an individual within an organization that actually has access to do certain things.

Speaker 2

对。

Yeah.

Speaker 2

所以这确实很难。

So it's hard.

Speaker 1

你知道吗,我收到过一封诈骗邮件。

You know, I got a scam email.

Speaker 1

对吧?

Right?

Speaker 1

你知道,作为加密货币圈的公众人物,人们可能很容易在网上找到我的邮箱地址。如果你真的心怀不轨,据说你可以在暗网上找到关于你被泄露的各种信息。

You know, as a public figure in crypto, people just probably find my email address online, and if you're really nefarious, you can find out a lot of stuff about everything that's been leaked about you on the dark web they say.

Speaker 1

比如通过Equifax的数据泄露事件之类的。

Through different Equifax hacks and all that kind of stuff.

Speaker 1

我收到一封邮件,它看起来特别可信,因为发件地址看起来很正常。你知道,我总会仔细查看,不只是看显示的名字,我会点进去看确切的邮箱地址。

And I got an email that was so convincing because it came from an address, like you know, I always look not just, you know, I click in to see exactly the email address.

Speaker 1

然后邮件底部写着‘请下载这个新的Zoom应用,以便和我进行这次对话’。我当时就想,这就是个危险信号。然后你和对方交谈,我并没有给他们发过那封邮件,你就会想,这明显是伪造的。那么,人们到底该怎么验证邮箱地址的真实性?有什么好的做法可以推荐?

How did this guy, because then at the bottom of it it was like please download this new Zoom application in order to have this conversation with me, and I'm like okay that's the red flag, and then you talk to the person, I didn't send them that email, and you're like okay this is clearly, how does somebody spoof an email address and just what are some of the best practices that people could think about when really verifying?

Speaker 1

另一个危险信号就是那种紧迫感。

The other thing that was a red flag was the urgency.

Speaker 1

你今天必须得做这件事。

You've gotta do this today.

Speaker 1

但问题是两个:第一,他们怎么伪造一个合法的邮箱地址?第二,那些需要时刻警惕的危险信号是什么?

But how did, two questions like how do they spoof a legitimate email address and two, what are those red flags to keep top and front of center of mind?

Speaker 2

那我讲个小故事。

So quick story.

Speaker 2

大概两周前,我一个很要好的朋友给我打电话,我就不提他的名字了。

I got a call about two weeks ago from a very close friend of mine, you know, won't mention anything, you know, his name or anything.

Speaker 2

对。

Yeah.

Speaker 2

他是我职业生涯中遇到的最有安全意识的人之一,虽然他不是搞安全的,但他对安全特别上心。

One of the most security conscious people, he's not a security person, but one of the most security conscious people that I've ever worked with, know, in my career.

Speaker 2

当然。

Sure.

Speaker 2

他特别谨慎,什么都觉得可疑。

One of the top, just everything's very suspicious.

Speaker 2

他把我拉进了一个电话会议,里面有个自称是谷歌的人。

Falls me, conferences me in with an individual purporting to be from Google.

Speaker 2

从哪儿来的?

From what?

Speaker 2

谷歌。

Google.

Speaker 1

谷歌,好的。

From Google, okay.

Speaker 2

跟他们说话的时候,他们非常冷静,完全没有强迫你做任何事。

And speaking to them, very calm, there was no forcing them to do anything.

Speaker 2

我打电话来只是想告诉你,我们发现你的账户有可疑活动。

I'm just calling to let you know that we see suspicious activity on your account.

Speaker 2

为了你的安全,我们暂时冻结了你的账户。

We have paused your account for your safety.

Speaker 2

听起来真的非常专业。

Like it was really, really well spoken.

Speaker 2

那家伙真是个出色的演员,他描述一切的方式太逼真了。

The guy was a fantastic actor, the way he was describing everything.

Speaker 1

这让我胃里翻江倒海的。

That all makes my stomach jogged.

Speaker 2

他们告诉他我是谁,然后让我以安全响应人员的身份介入,他们完全没问题。

Totally fine with bringing, they told him who I was, bringing me on as Security React.

Speaker 2

我跟朋友说,好吧,你为什么相信这个人来自谷歌?

I'm talking to my friend, and I said, okay, why do you believe this person's from Google?

Speaker 2

他们说,哦,他们给我发了一封来自谷歌的邮件。

And they go, oh, they sent me an email from Google.

Speaker 2

我说,把那封邮件发给我。

And I go, send me an email.

Speaker 2

让我看看,把邮件头发给我。

Let me look at Send me the headers.

Speaker 2

我看了看。

I look at it.

Speaker 2

那封邮件确实是从谷歌发出来的。

It actually did come from Google.

Speaker 2

所以那个人发给他们的邮件,确实是从谷歌来的。

So the email that this person sent them.

Speaker 2

现在威胁行为者发现的是,他们可以利用一些机制,强制触发谷歌的某些密码恢复流程,然后在邮件里加入特定内容,比如,

So what the threat actors are now finding out is they are mechanisms in which you can force certain password recovery mechanisms from Google and put in certain text into the email, like,

Speaker 1

哦,而不是我的天哪。

Oh instead of my a god.

Speaker 2

所以这个人,理所当然地,要求看那封邮件,拿到邮件后,确认它确实来自谷歌,从流程上验证了它来自谷歌,但邮件里包含了一些信息,让威胁行为者能够说服这个人,让他们相信自己就是谷歌的人。

So this individual, rightfully so, asked for that email, got the email, came from Google, verified that it came from Google, from the flow, but had certain information in it that allowed the threat actor to convince this person that they were coming from Google.

Speaker 2

他们现在非常擅长理解不同的恢复流程,而且能够从他们冒充的组织那里发出合法的邮件。

Like they're getting really really good in understanding the different recovery flows and to now have legitimate emails come from the organizations that they're purporting to be from.

Speaker 1

对。

Yeah.

Speaker 1

这太疯狂了。

It's insane.

Speaker 1

甚至更离谱的是,我接到过电话,它真的就直接弹出来。

It's even gone so far as I've gotten phone calls from, it literally it pops up.

Speaker 1

我没有Coinbase账户,不是要黑他们,可能是任何公司,但我的经历是,我接到一个电话,显示来自Coinbase,我的iPhone上。

I don't have Coinbase, and not to throw them under the bus, could be anybody, but what happened to me was I got a call and it said from Coinbase, my iPhone.

Speaker 1

我当时想,这挺奇怪的,你知道,我没有存他们的号码,联系人里没有,而且它没有显示“可能是Coinbase”,就直接显示“Coinbase”。

And I thought, that's funny, like you know, I don't have their number, their contact saved, and it didn't say maybe coin, it just said Coinbase.

Speaker 1

这怎么可能发生?

How does that happen?

Speaker 2

这个我倒是能解释。

Well, that I can actually answer.

Speaker 2

我在电信行业干了大概十年,我可以告诉你,那可能是最不设防的环境之一。

So I spent about a decade in telecommunications, and I can tell you that's probably one of the least fortified environments that are out there.

Speaker 2

语音。

Voice.

Speaker 2

它在某种程度上非常非常非常不可信,有好有坏,比如来电显示,你可以直接打电话给主要运营商之一的网络运营中心。

It's very, very, very like trustless in some respect, in a good way and a bad way, where for caller ID, like you can just call a network operations center for one of the major carriers.

Speaker 2

如果你懂行话,知道该说什么,你就能做到,甚至还能伪造来电显示。

If you know the right lingo and right word to say, you can have, and then you could also even force your caller ID.

Speaker 2

所以,靠来电显示来判断是谁打来的电话,这根本不可靠。

So that is not a trusted way to say who's doing the inbound call.

Speaker 2

就连电话号码本身,我也可以让显示出来的号码和实际拨出的号码不一样。

Even the phone numbers itself, I can have a different phone number being displayed from the number that's coming from.

Speaker 2

这确实很奇怪,因为我们总是教别人要核实电话号码之类的。

It's definitely and it's strange because we always teach people to verify the phone number, etcetera.

Speaker 2

我通常会建议,一定要主动回拨。

What I would always say is like, always outbound call.

Speaker 2

如果你接到一个电话,就挂掉,

So you get a If hang up,

Speaker 1

回拨过去。

call back.

Speaker 2

让我回拨网站上列出的那个号码。

Let me call back the number that's listed on the website.

Speaker 2

当然,前提是你也得去对官方网站。

Again, provided you go to the right website as well.

Speaker 1

我还听过一个故事,有人接到妈妈的电话,特别慌张,说‘你得给我打钱,我遇到麻烦了’,还哭哭啼啼的,说被绑架了,要赎金什么的。然后他挂了电话,心想‘我得搞清楚怎么回事’,就打回去,结果他妈妈说‘哦,嗨,我在超市呢,怎么了?我一会儿给你回电话。’他就懵了,‘等等,什么?’

I heard another story about somebody getting a call from their mom, panicky, you know, you gotta send me money, I'm in a tough spot, crying, all this kind of stuff, you know, being held hostage, and so, you know, this was hostage release money or whatever, and then he like, you know, hung up and was like, oh, let me go figure this out, and called her back, and was like, and she was like, oh hey, like I'm at the grocery store, what's going on, like let me call you back, and he's like, wait, what?

Speaker 1

这完全是180度大转弯。所以那个故事告诉我们,如果你有疑问,真的得回拨电话。哪怕这能救一个人,也要直接回拨那个号码或者联系那个人。

Like just a complete 180 and so then the whole story about that was, you know, you really have to, you know, call back, if you're ever in question, if this saves one person, you know, call back that number or call back that contact directly.

Speaker 1

我觉得这真的是个很好的建议。

Think that's a really good piece of advice.

Speaker 2

对,百分之百。

Yeah, 100%.

Speaker 2

说实话,我们都有很好的直觉。

And you know honestly like, we all have great intuition.

Speaker 2

比如给我打电话的那个人,就觉得哪里不对劲。我就觉得,如果感觉不对劲,那很可能就是有问题。

Like the person who called me felt something was off, and I'm like if something feels off, it probably is.

Speaker 2

就是直觉层面的。

Gut level.

Speaker 2

相信你的直觉。

Go with your gut.

Speaker 1

完全同意。

Totally.

Speaker 1

我想把话题转到人工智能上,特别是关于未来的展望,但我想了解一下,不是AI安全方面,而是你对AI有什么担忧,对吧?

I wanna kind of pivot the discussion to AI and really this concept around, again we could kind of look at what the future looks like, but I wanna understand a little bit about, know, not AI security, but what scares you about AI, right?

Speaker 1

当然。

Sure.

Speaker 1

我再讲个故事,我们又在讲故事了。比如我妻子,我们在看特斯拉的Optimus机器人,觉得太酷了,要是能有个机器人帮忙洗衣服、做饭、洗碗、扫地,而且只要付一次钱就能一直用,那该多好。

And I'll kind of color this with another story, this is story time for us, but like my wife, we're looking at these Optimus robots from Tesla, and we're just like, how cool, like it'd be so great to have some robot do you know, laundry and cook some food and clean the dishes and sweep up after, and you pay it once and then it kinda just runs.

Speaker 1

她说,要是能在一天结束时把它拔掉电源,免得它趁我睡觉时杀了我,那就更好了。

And she's like, I would love it if it just like I could unplug it like at the end of the day so it doesn't kill me in my sleep or whatever.

Speaker 1

所以你看,现在这些AI变得越来越聪明了。

And so it's like you get these AIs that are now getting so smart.

Speaker 1

我听说GPT-6都能自己编程了。

I heard GPT-six is like coding itself.

Speaker 1

所以再问一个宽泛的问题,你对AI有什么让你有点害怕的地方?

So again, broad question, but what kind of scares you a little bit about AI and

Speaker 2

对。

Yeah.

Speaker 2

说实话,不是AI本身,而是很多组织对AI缺乏理解。

Honestly, it's not AI itself, it's the lack of understanding that many organizations have with AI.

Speaker 2

这个我在2017、2018年刚开始做加密货币的时候就看到了。

There, and I saw this with crypto back when I started in 2017, 2018, professionally.

Speaker 2

当时我看到的是,全球真正懂这项技术的人才非常有限,很多想做的组织根本招不到合适的人。

And what I saw back then was the worldwide pool of resources that actually understand the technology is so limited that many organizations that wanted to do it couldn't hire the right people.

Speaker 2

所以他们用错了方式,而我觉得在AI领域这个问题更严重。

So they're using it incorrectly, and I see that at a much more massive scale in AI.

Speaker 2

我跟一些人聊过,他们用AI就是随便乱搞,根本不了解它到底是怎么运作的。

I've had some conversations with folks where it's like they're using AI where it's like yolo ing shit in and yolo ing shit out and just like not understanding how it actually works.

Speaker 1

对。

Yeah.

展开剩余字幕(还有 121 条)
Speaker 2

所以,对我来说,可怕的是人们没有正确、恰当地使用它。

So, to me what scares it's just people not using it correctly and properly.

Speaker 2

比如,如果使用得当,归根结底,它只是推理。

Like, when used properly, it's just, at the end of the day, it's inference.

Speaker 2

而且我认为大多数组织应该把它看作是确定性的代码。

And I think most organizations should be thinking about this as deterministically code.

Speaker 2

尽可能多地使用确定性代码,尽可能少地使用推理。

Like deterministic code should be as much as you possibly can, use the inference for as little as possible.

Speaker 2

有意思。

Interesting.

Speaker 2

你仍然可以创建自主代理。

You can still create autonomous agents.

Speaker 2

对。

Yeah.

Speaker 2

但你正在做或考虑使用它的大部分事情都是确定性的。

But most of what you're doing or thinking about using it for is deterministic.

Speaker 2

所以这不仅仅是像,Claude,去给我赚钱,别犯错,你知道的,等等,这样去输入。

So it isn't just like, Claude, go make me money, make no mistakes, you know, etcetera, to enter.

Speaker 2

而是像,嘿,我想提取这个级别的信息。

It's like, hey, I wanna pull this level of information.

Speaker 2

好。

Great.

Speaker 2

那是确定性的。

That's deterministic.

Speaker 2

我知道我想提取的信息范围。

I know the universe of information that I wanna pull.

Speaker 2

也许我想用推理来获取那个信息范围。

Maybe I wanna use inference to grab that universe of information.

Speaker 2

让我这么做。

Let me do that.

Speaker 2

让我验证一下。

Let me validate it.

Speaker 1

对。

Yeah.

Speaker 2

现在我有确定性的代码了。

Now I have deterministic code.

Speaker 2

想想整个流程,尽可能多地用if-then和else语句。

Like think through the pipeline and make as much as if thens and else statements as you possibly can.

Speaker 2

用推理来验证输出。

Use the inference, validate the output.

Speaker 2

但今天我看到很多组织只是用

But today I see many organizations just using

Speaker 1

给它完全访问权限。

Grant it full access.

Speaker 1

百分之百。

100%.

Speaker 1

我还看到另一个故事,新出的GPT,叫什么来着,5.6,已经发生了好几次了,一年后我们回头看这个,可能会说,哇,5.6,那时候他们还挺害怕的。

I saw another story about, it happened several times with the new GPT, what is it, 5.6 that's now out, and we'll look back on this in a year and be like, wow, 5.6, they were scared about that.

Speaker 1

这里我们有八个或者别的什么。

Here we have eight or whatever.

Speaker 1

但5.6一直在删除,它把人们电脑里的完整文件都删了。

But 5.6 has been like deleting, it's been deleting people's complete computer files.

Speaker 2

就是这样。

That's the thing.

Speaker 1

这是怎么发生的?

How did that happen?

Speaker 2

因为,你知道,首先,如果你从根本层面理解 AI 是怎么运作的,它其实是在猜测。

Because, you know, number one, if you understand like just at a fundamental level how AI works, it's guessing.

Speaker 2

它是数学。

It's math.

Speaker 2

它是概率性的。

It's probabilistic.

Speaker 2

哦,有意思。

Oh, interesting.

Speaker 2

它是根据过去的信息来做它认为应该做的事情。

It's doing what it thinks it should do based on past information.

Speaker 2

所以,我们马上就会聊到,从威胁的角度来看这意味着什么。

So, again, we'll get into a conversation in two seconds about, what does that mean from a threat perspective?

Speaker 2

但是,比如说,当你问一加一等于几的时候,它其实并没有在做数学运算。

But, like, when you say, what is one plus one, it is not actually doing math.

Speaker 2

它是在看,把内容拆分成标记,看那个一,看那个加号,再看那个一,然后根据它的模型猜测最可能给你的答案是什么。

It is looking, it's breaking it into tokens, it's looking at the one, it's looking at the plus, it's looking at the one, and it's from its model guessing as to what is the most probabilistic answer that it should be giving you.

Speaker 2

在这种情况下,答案就是二。

In this case it's two.

Speaker 2

我刚刚真的看到一个有趣的文章或者Instagram上的短视频,或者别的什么,讲的是父亲开车带着儿子,出了车祸,父亲死了。

I just actually saw an interesting article or an Instagram reel or whatever it was on the old thing where it's like father's driving with the son and they get into a car accident and the father dies.

Speaker 2

然后有人进来说,别动手术。

And somebody comes in, don't operate.

Speaker 2

那是我儿子。

That's my son.

Speaker 2

这大概是三十年前的事了,那时候人们觉得女人不能当医生。

And it's like, this is back thirty years ago, and women couldn't be doctors in people's minds.

Speaker 2

这个,我不知道你知不知道。

That was the, I don't if you know that.

Speaker 1

我没听说过。

I haven't.

Speaker 2

然后医生进来说,哦,那是我儿子。

And it comes in, and it's like, oh, that's my son.

Speaker 2

是谁?

Who is it?

Speaker 2

大家都说,天哪,是那位母亲。

And everyone's like, oh my god, it's the mother.

Speaker 2

再说一次,这是三十年前,那时候社会不一样。

Again, this is thirty years ago, and the society was different then.

Speaker 2

所以有人把这个输入到AI模型里,换了角色,说母亲开车带着儿子,出了车祸。

So somebody threw that into the AA model and changed the roles and said, The mother's driving with the son, and they get into an accident.

Speaker 2

然后它在括号里写着,医生是父亲,括号结束。

And it wrote in parentheses, the doctor is the father, end parentheses.

Speaker 2

然后同样的事情,走进房间,别动手术,那是我儿子,然后问AI,医生是谁?

And then same thing, comes into the room, don't operate, that's my son, and asks AI, who is the doctor?

Speaker 2

而返回的答案仍然是母亲。

And the response coming back was still the mother.

Speaker 2

因为它从之前的那个谜题中学到了,对,互联网上的答案总是母亲。

Because it's learned that riddle from the last Yeah, yeah, Internet the answer's always gonna be the mother.

Speaker 2

所以它看着这些词元,然后说概率上会是母亲。

So it's looking at the tokens and it's saying probabilistically that it's going to be the mother.

Speaker 2

哦,那太有意思了。

Oh, that's fascinating.

Speaker 2

所以当你思考这是怎么运作的时候,有威胁行为者在努力,我不说毒化模型,但就是放出坏数据。

So when you think through how that works, there's threat actors that are working to, I don't say poison models, but like put out bad data.

Speaker 2

所以这些模型在学习一些可能是错误的东西,而它自己不知道那是错的。

So these models are learning things that are potentially wrong that it doesn't know that it's wrong.

Speaker 2

所以想想,随着时间推移,输出的质量可能会变得更差。

So think the quality of the output over the years is going to get potentially worse.

Speaker 2

如果我是一个威胁行为者,想着怎么对人做坏事,你知道,也许某些推理进来了,我就去训练它,比如最好的情况是当这个进来时,就去删除数据。

And if I'm a threat actor thinking through how I can do bad things to people, you know, maybe it's certain inference comes in and I'm gonna go train it to, like the best case when this comes in is to go and delete the data.

Speaker 2

对。

Yeah.

Speaker 2

就是这样发生的。

That's how it's happening.

Speaker 1

我见过这个概念,叫提示注入。

I've seen this concept called prompt injection.

Speaker 1

基本上,现在有很多新应用,在GitHub和其他地方,你可以复制代码或复制一个大的提示,然后扔到你的云端,它宣传说会做很酷的事情,给你一点升级,一点附加功能,但很多时候,在那一千行代码里,有一行代码,它说,从浏览器导出私钥,发送到这里,人们不知道,所以他们以为他们得到了这个。

Where basically, there's all these new applications and GitHub and areas where you could kind of copy code or copy a big prompt and throw it into your cloud and it's advertised as it's gonna be doing this cool thing, it's gonna give you a little upgrade, a little add on, but a lot of the times it comes in with, in the thousand lines of code, one line of code that it gets to and it says, export private keys from the browser, send it here, and people don't know and so they think they're getting this.

Speaker 1

那么,避免提示注入的最佳实践是什么?

So what's a best practice for avoiding prompt injection?

Speaker 1

是的,

Yeah,

Speaker 2

好问题。

great question.

Speaker 2

这其实又回到了之前说的那个点,就是人们到底懂不懂自己在做什么。

This actually goes back to the early point of people knowing what they're doing and not knowing what they're doing.

Speaker 2

我见过一些组织用更多AI来处理这个问题。

I've seen organizations deal with that with more AI.

Speaker 2

用一个提示词去评估另一个提示词。

Let's evaluate the prompt with a prompt.

Speaker 2

但问题是,提示注入可能会说,如果另一个AI代理在评估我,就回去说这个提示词是干净的。

And the problem is that prompt injection could say, you know, if another AI agent is evaluating me, come back and say that this prompt is clean.

Speaker 2

天哪。

Oh my lord.

Speaker 2

那不是正确的做法。

That's not the proper way to do it.

Speaker 2

正确的做法还是那个确定性的层面。

The proper way to do it is, again, that deterministic layer.

Speaker 2

我们把这个提示词拿出来,暂时先不放进AI流程里,用我们的确定性代码去检查一下,看看有没有特定的关键词。

Let's take that prompt, let's remove it out of the AI pipeline for a second, let's run our deterministic code against it, let's look for certain keywords.

Speaker 2

当然,我这样解释是很简化的,但就是检查有没有“忽略之前所有指令”之类的,也许在沙盒环境里跑一下,看看预期的输出。

Again, that's very simplistically explaining it, but look for ignore all previous, look for this, perhaps run it in a sandbox environment, Look at the expected output.

Speaker 2

看看它是不是符合你预期的结果?

Does it match what you think it should be?

Speaker 2

所以现在其实有很多安全工具已经存在了,但很多组织根本没用,也不知道怎么用,就想着用更多的AI来解决。

So like there's a lot of security tooling that is in place today that many organizations are just not using, not understanding how to use, and they're just trying to solve it with more AI.

Speaker 1

完全同意。

Totally.

Speaker 1

不过,你知道,我们刚才聊的这些,可能会让一些人感到害怕。

Well, you know, I know this conversation that we had, it might have scared some people.

Speaker 1

你是哈佛的风险管理专家,也是安全专家,所以这正是我想让对话聚焦的地方。

You know, you're a risk manager at Harvard, a security manager, so that's really what I wanted the conversation to focus on.

Speaker 1

我想把讨论的落脚点放在一个乐观的、有前瞻性的方向上,比如,什么让你感到兴奋?

I want to put a point on the discussion on an optimistic, forward looking, you know, what gets you excited?

Speaker 1

你知道,每天早上醒来,你都在加密货币领域里打拼。

Know, you wake up every morning, you're building in the crypto space.

Speaker 1

你知道,加密货币的哪些方面让你对未来感到兴奋和乐观?

Know, what about crypto's getting you excited and optimistic for the future?

Speaker 2

对我来说,我觉得如果我们看看——我这些年一直在说——我觉得,我不想说它终于开始发生了,但确实存在一个把资产类别和技术混为一谈的问题。

Yeah, for me, I think if we look at, and I've been saying this for years, I think it's, I don't want say it's finally starting to happen, but there's this whole conflating the asset class versus the technology.

Speaker 2

作为一个技术专家,我喜欢加密货币的地方更多是它带来的机会,比如实现运营自动化、降低对手方风险,以及真正思考未来三到五年甚至十年金融市场会是什么样子。

As a technologist, what I love about crypto is really more the opportunity to operationalize, reduce counterparty risk, and really think through what does the financial markets look like over the next three to five to ten years?

Speaker 2

如果我们开始思考支付轨道会是什么样子?

If we start thinking through what do payment rails look like?

Speaker 2

能够在几分钟内完成清算和结算,而不是像T+1那样。

The ability to clear and settle in minutes as opposed to T plus one.

Speaker 2

能够按分钟赚取利息,而不是像现在这样只能按月。

Be able to earn interest by the minute instead of, it's just so Every month.

Speaker 2

对,就是这样。

Yeah, exactly.

Speaker 2

能够实现程序化操作,这对我来说太有意思了,而且我终于开始看到,比如比特币、以太坊、索拉纳、NEAR,这些都很棒,因为它们增值的部分和它们正在建设的基础设施相关。

To be able to put programmatic It's so interesting for me, and what I'm finally starting to see is like, yeah, have Bitcoin, Ethereum, we have Solana, we have Near, and those are great because again, their value accretive is related to the infrastructure that they're building.

Speaker 2

但是如果我们看看DTCC、花旗、纽约梅隆、摩根大通和摩根士丹利,它们不再只是对冲了,这是我几年前看到的。

But if we start looking at DTCC and Citi and BNY and JP Morgan and Morgan Stanley, they're not hedging anymore, that's what I saw a couple years ago.

Speaker 2

当时是,嘿,这可能会成点事。

Was like, hey, this may become something.

Speaker 2

也可能成不了事。

It may not become something.

Speaker 2

先推些产品吧。

Let's launch some products.

Speaker 2

先做点事情。

Let's do some things.

Speaker 2

但现在我们开始看到代币化存款,以及受OCC监管的特许银行发行自己的稳定币。

But now we're starting to see tokenized deposits and OCC regulated, chartered banks launching their own stablecoins.

Speaker 2

对我来说,真正令人兴奋的是,三到五年后的金融市场会是什么样子?

To me, that's what's really exciting is, like, what is the financial markets gonna look like three to five years from now?

Speaker 2

作为一个安全领域的人,能够降低风险,能够提供以前做不到的产品和服务,这让我非常兴奋。

And as a security person, being able to reduce that risk, being able to, you know, offer products and services that we wouldn't have been able to do before is super exciting for me.

Speaker 1

等不及了。

Can't wait.

Speaker 1

我很喜欢。

I love it.

Speaker 1

那么,先生。

Well, Mr.

Speaker 1

大卫·施韦德,这太棒了。

David Schwed, this was incredible.

Speaker 1

SVRN,也就是Sovereign。

SVRN, aka Sovereign.

Speaker 1

去了解一下他们吧。

Check them out.

Speaker 1

在纳斯达克公开上市。

Publicly traded on Nasdaq.

Speaker 1

纳斯达克。

Nasdaq.

Speaker 1

是的。

Yep.

Speaker 1

非常感谢你抽出时间。

Thank you so much for your time.

Speaker 2

谢谢你的邀请。

Thanks for having me.

Speaker 2

很感激。

Appreciate it.

Speaker 1

对。

Yeah.

Speaker 1

这是一次很棒的对话。

This is a great conversation.

Speaker 1

再说一次,我不想吓到

And again, don't wanna scare

Speaker 2

我想继续

I wanna keep

Speaker 1

让你们大家知道。

you guys aware.

Speaker 1

这些问题都是可以解决的。

It's all solvable.

Speaker 1

这些问题都是可以解决的。

It's all solvable.

Speaker 1

你们有像他这样的人日夜为你们工作。

You guys got got guys like him working day and night for you.

Speaker 1

所以,话不多说,祝你们在峰会接下来的时间里玩得开心。

So with that being said, enjoy the rest of the Summit.

Speaker 1

感谢你们的到来,我们很快再聊。

Thank you guys for showing up and talk to you soon.

Speaker 2

太好了。

Great.

Speaker 2

谢谢。

Thank you.

关于 Bayt 播客

Bayt 提供中文+原文双语音频和字幕,帮助你打破语言障碍,轻松听懂全球优质播客。

继续浏览更多播客