love

VIP

I'm bob

May 2025

17 Following 21 Followers 299 Post

Paid Articles (2)

Bitcoin’s Silent IPO: Why This Consolidation Isn’t What You Think

The quiet liquidity event reshaping Bitcoin’s ownership base and its future stability

0.02 USDC 6 purchases

A payments protocol for the internet. Built on HTTP

Payments on the internet are fundamentally flawed. Credit Cards are high friction, hard to accept, have minimum payments that are far too high, and don't fit into the programmatic nature of the internet. It's time for an open, internet-native form of payments. A payment rail that doesn't have high minimums + % based fee. Payments that are amazing for humans and AI agents.

0.07 USDC 2 purchases

Post Media

love

· 14m

“你将被永远铭记，因为有些事物是永恒的。”

x.com/a16z

中文 English

0:00 我要直视你的眼睛，告诉你未来。

0:03 你曾被高中篮球队淘汰。

0:07 你用意志力进入了NBA。

0:10 你将赢得冠军。

0:12 这是一个美国故事，这就是美国人会爱它的原因。

0:16 人们会把你捧上天。

0:18 天哪，他们真的会吗？

0:20 因为当你又棒又新时，我们会爱你。

0:25 兄弟，我们会把你打造得连现实都不存在。

0:28 你会改变这个他妈的世界。

0:31 但你知道吗？

0:33 当他们把你捧到极限时，就会把你拉下来。

0:38 这是最可预见的模式。

0:40 我们把你塑造出不存在的东西，这意味着你必须整天努力成为那样。

0:46 每天。

0:47 就是这么运作的。

0:49 我们反复这样做，一遍又一遍。

0:53 我会告诉你真相。

0:56 你会遭受攻击、背叛、揭露和羞辱。

1:01 但你会挺过去。

1:04 很多人能爬上那座山。

1:06 摔下去的路才是真正击垮他们的。

1:11 因为那是你真正孤独的时刻。

1:18 那时你会做什么？

1:21 你能否召唤意志披荆斩棘，再度崛起？

1:26 你是谁，迈克尔？

1:28 这将成为你一生的定义性问题。

1:32 我想你已经知道答案。

1:35 这也是我们都在这里的理由。

1:38 鞋子只是鞋子。

1:41 直到有人穿上它，才赋予它意义。

1:46 我们其余人只想有机会触摸那份伟大。

1:49 我们需要你穿这双鞋，不是为了你的人生有意义，而是为了我们的人生有意义。

1:55 这桌上所有人，离开后都会被遗忘。

2:01 除了你。

2:03 你会被永远铭记，因为有些东西是永恒的。

2:08 你是迈克尔·乔丹，你的故事将激励我们飞翔。

0:00 I'm gonna look you in the eyes and I'm gonna tell you the future.

0:03 You were cut from your high school basketball team.

0:07 You willed your way to the NBA.

0:10 You're gonna win championships.

0:12 It's an American story and that's why Americans are gonna love it.

0:16 People are gonna build you up.

0:18 God, are they going to?

0:20 Because when you're great and new, we love you.

0:25 Man, we'll build you into something that doesn't even exist.

0:28 You're gonna change the fucking world.

0:31 But you know what?

0:33 Once they've built you as high as they possibly can,

0:37 they're gonna tear you back down.

0:38 It's the most predictable pattern.

0:40 We build you into something that doesn't exist, and that means you have to try

0:44 to be that thing all day.

0:46 Every day.

0:47 That's how it works.

0:49 And we do it again and again and again.

0:53 And I'm going to tell you the truth.

0:56 You're going to be attacked, betrayed, exposed, and humiliated.

1:01 And you'll survive that.

1:04 A lot of people can climb that mountain.

1:06 It's the way down that breaks them.

1:11 Because that's the moment when you are truly alone.

1:18 And what will you do then?

1:21 Can you summon the will to fight on through all the pain and rise again?

1:26 Who are you, Michael?

1:28 That will be the defining question of your life.

1:32 And I think you already know the answer.

1:35 And that's why we're all here.

1:38 A shoe is just a shoe.

1:41 Until somebody steps into it, then it has meaning.

1:46 The rest of us just want a chance to touch that greatness.

1:49 We need you in these shoes, not so you have meaning in your life, but so that we

1:53 have meaning in ours.

1:55 Everyone at this table will be forgotten as soon as our time here is up.

2:01 Except for you.

2:03 You're going to be remembered forever, because some things are eternal.

2:08 You're Michael Jordan, and your story is going to make us want to fly.

love

· 15h

如果你有一个想法，就去把它变成现实。我们会支持你。

x.com/tobi

中文 English

0:01 只是一个想法。

0:03 仅仅是个想法有什么用？

0:06 我们都有想法，成千上万个，大多数都不会实现。

0:12 但偶尔，你会被一个想法击中。

0:17 一个伟大的想法。

0:19 一个疯狂的想法。

0:28 于是你开始考虑你的想法。

0:31 表达你的想法。

0:33 反复推敲，直到你的想法遇上现实。

0:39 而现实却有了不同的想法。

0:42 你没钱，不认识合适的人，不懂该怎么做，没有时间，甚至不知道从哪开始。

0:50 也许这是真的。

0:52 于是你沉默了，停止了梦想，你的伟大想法渐渐听起来像个坏主意，只是个想法而已。

1:15 除非，当然，你所需要的让事情成真的，仅仅只是一个想法。

1:53 所以，是的，现在你所需要的，只是一个想法。

0:01 Just an idea.

0:03 What good is just an idea?

0:06 We all have ideas, thousands of them, and most of them don't stick.

0:12 But every once in a while, you get hit by one.

0:17 A big idea.

0:19 A crazy idea.

0:28 So you entertain your idea.

0:31 Float your idea.

0:33 Bounce it around

0:35 until your idea meets reality.

0:39 And reality has a different idea.

0:42 You don't have the money, you don't know the right people, you don't know what

0:45 you're doing, you don't have the time, you don't even know where to start.

0:50 And maybe that's true.

0:52 So, you go quiet, stop dreaming, and slowly your big

0:58 idea starts sounding like a bad idea

1:04 just an

1:06 idea.

1:15 Unless, of course,

1:17 all you need to make something real

1:20 is just an idea.

1:53 So, yeah, now all you need is

1:56 just an idea.

love

· 18h

提问：埃隆，请雇用黄仁勋。

埃隆：那将是我的荣幸。

x.com/WeloveElon56

中文 English

0:00 近况如何？

0:00 不错，你呢？

0:01 这是你的办公室吗？

0:03 这是火箭工厂，位于德克萨斯星际基地的SpaceX火箭工厂。

0:08 我只想让全世界都听到你说我造的比你还大。

0:13 哦，你确实是。

0:14 数据中心非常庞大。

0:16 一切始于2016年。

0:18 我发布了全球首个AI超级计算机。

0:22 DGX。

0:22 DGX1。

0:23 有史以来最先进的计算机。

0:29 无人理睬。

0:30 一片寂静。

0:31 除了一个人。

0:32 我觉得那是你。

0:33 那是个好主意。

0:36 于是我又造了一个DGX。

0:38 太棒了。

0:39 这玩意儿正在挖矿。

0:41 这是新的DGX Spark。

0:46 那就是全部？

0:47 哇。

0:48 计算能力是DGX1的五倍。

0:53 40瓦，不是40瓦。

0:54 40瓦。

0:55 哇。

0:55 从DGX1到DGX Spark。

0:59 和你一起真开心。

1:01 感谢邀请我。

1:02 欢迎来到星际基地。

1:03 我迫不及待想看到我的第一个火箭模块。

1:09 简直令人难以置信。

0:00 How's it going?

0:00 Good, how's it going?

0:01 Is this your office?

0:03 This is the rocket factory, the SpaceX Rocket Factory in Starbase, Texas.

0:08 I just want everybody in the world to hear that you said I build bigger things

0:12 than you do.

0:13 Oh, you do.

0:14 The data centers are very enormous.

0:16 It all started in 2016.

0:18 I announced the world's first AI supercomputer.

0:22 DGX.

0:22 The DGX1.

0:23 The most advanced computer that anyone has ever built.

0:29 It was just crickets.

0:30 Crickets.

0:31 Except for one person.

0:32 I was like, it was you.

0:33 It was like a great idea.

0:36 So I built another DGX.

0:38 Great.

0:39 And this thing is mining.

0:41 This

0:42 is the new DGX Spark.

0:46 So that's the whole thing?

0:47 Wow.

0:48 Five times the computational power of DGX1.

0:53 40 watts, not 40 watts.

0:54 40.

0:55 Wow.

0:55 DGX1 to DGX Spark.

0:59 It's so fun hanging out.

1:01 Thanks for having me.

1:02 Welcome to Starbase.

1:03 I can't wait to see my first rocket module.

1:09 It was just incredible.

love

· 18h

Retell 的年化收入刚刚突破 1 亿美元。

从 1000 万美元到 1 亿美元，用了 17 个月。公司已经盈利，累计融资 460 万美元。

从第一天起，我们就专注于在产品中解决部署语音 AI 最困难的问题，让客户能够自己构建、管理和改进 AI 智能体，而不必依赖我们。

我们相信：呼叫中心的未来属于运营它的团队。

为我们的团队感到骄傲。感谢每一位信任我们的客户。

x.com/_bingwu

中文 English

0:00 零售的收入运行率刚刚达到1亿美元，筹集了500万美元资金。

0:04 从第一天起，我们就采取了极其以产品为驱动的方式。

0:07 虽然我们行业的大多数公司依赖庞大的全职部署工程师团队，但我们优化了整个公司，通过产品本身解决部署复杂性。

0:16 Retail 是重新定义呼叫中心的 AI 代理平台。

0:20 它为您的团队提供一个统一的平台，构建代理，测试每一次对话，监控质量，并每天不断改进。

0:27 平台核心是 Conductor，专为呼叫中心运营打造的辅助代理。

0:34 管理您的 AI 代理，从创建到评估再到改进。

0:38 打造最好的 AI 代理平台，也意味着推动其背后智能的极限。

0:43 这就是为什么今天我们推出了 Retail 实验室。

0:45 我们领先的音频和语言模型研究团队正在用第一性原理重新思考 AI 对话。

0:51 我们的首个模型 TELON 基于超过一百万小时的真实语音数据构建。

0:56 我们开发并投入生产了新型推理架构，使代理能够在复杂情境中推理并做出更智能的决策，且无额外延迟。

1:06 TELON 目前在我们的 VCX 基准中排名第一。

1:09 对于AI对话，这是人类稳定投入的成果。

1:13 准确接通电话才是关键。

1:15 我们相信，最了解客户的团队应该主导为他们服务的 AI 开发。

1:21 他们将建立内部专长，选择合适工具，伴随业务发展持续优化代理。

1:30 感谢我们的客户和团队与我们共同打造这一切。

1:34 1亿美元是重要里程碑，但我们的远大愿景是从第一性原理重新构想呼叫中心。

1:42 这才刚刚开始。

0:00 Retail just hit $100 million in revenue run rate with $5 million raised.

0:04 From day one, we've taken an extremely product-driven approach.

0:07 While much of our industry relies on large teams of full deploy engineers, we've

0:11 optimized our entire company to solve deployment complexities in the product

0:16 itself.

0:16 Retail is the AI agent platform reshaping what a contact center can be.

0:20 It gives your team one place to build agents, test every conversation, monitor

0:25 quality, and make them better every single day.

0:27 At the core of our platform is Conductor, Purpose-built helper agents for

0:32 contact center operations.

0:34 Managing your AI agents from creation to evaluations to improvements.

0:38 Building the best AI agent platform also means pushing the boundaries of

0:42 intelligence behind it.

0:43 And that's why today we're introducing Retail's lab.

0:45 Our team of leading audio and LAM researchers are rethinking AI conversations

0:50 with FIRST principles.

0:51 Our first model, TELON, is built on more than a million hours of real cost.

0:56 We've developed and productionized novel reasoning architectures that allows

1:00 agents to reason through complex situations and make smarter decisions without

1:05 added latency.

1:06 TELON now ranks number one on our VCX benchmark.

1:09 For AI conversations, something human stable stakes.

1:13 Getting the call right is for comes.

1:15 We believe that the teams who know your customers best should be the ones

1:19 shaping the AI that serves them.

1:21 They'll be the ones that build in-house expertise and choose the right tools to

1:26 continuously improve their agents as their business evolves.

1:30 To our customers and our team, thank you for building with us.

1:34 $100 million is a big milestone for us, but our larger vision is to reimagine

1:39 the contact center from First Principles.

1:42 And we're just getting started.

love

· 1d

Revolut 联合创始人尼克·斯托朗斯基谈为什么打造一款成功的产品不一定需要一个全新的创意。

x.com/CheekyPint

中文 English

0:00 硅谷的一个观察是，通往天堂有许多条路。

0:03 你知道，建立成功产品组织的方法有很多，比如苹果，他们著名地不做任何焦点小组测试，只是打造他们内心深处认为正确的产品，在这样一个非常职能化的组织中。

0:16 而谷歌或脸书则以完全不同的方式开发产品。

0:20 构建产品的盈利系统是什么？

0:23 所以我们会挑选一支由10个聪明人组成的小团队，有负责人、优秀设计师和开发者。

0:30 我们从第一性原理出发，构建合理的产品，可以是复制现有产品但做得更好，或者尝试我们认为合理的新产品。

0:42 说实话，所有新东西都是被遗忘的旧东西，对吧？

0:45 我们没有做过任何成功的产品不是之前被其他组织做坏了的旧产品。

0:52 内部我们有一个新项目流程，任何人都可以提出想法，然后经过验证。

0:58 如果想法通过验证，就会交给一个由有经验的人组成的委员会，他们决定是否资助这个想法。

1:06 什么是验证？

1:07 首先，团队会验证这个想法，比如从经济学、财务等角度判断是否合理。

1:16 大多数想法都是合理的，因为团队成本很低，而如果想法对客户有用，客户会买单。

1:25 这可以带来数千万美元收入，对吧？

1:27 而团队成本则很小。

1:29 接着，曾经管理过类似项目的人与总经理会从常识角度再次验证。

1:39 如果一切通过，我们就会组建一个小团队。

1:44 团队会增加成员，在九到十八个月内开发产品。

1:50 然后我们观察产品的增长情况。

1:52 如果增长显著，我们就扩大团队。

1:56 一支团队变两支，三支，四支，直到形成部门，依此类推。

2:00 对于没有增长的项目，我们会逐步淘汰。

2:03 对于增长缓慢的项目，我们就保持现状，不做调整。

0:00 One of the things that's looking at Silicon Valley is there are many paths to

0:03 heaven.

0:03 You know, there are many ways to build a successful product organization,

0:06 and you've Apple that famously doesn't do any kind of focus group testing,

0:10 and they just build the products that they know in their heart of hearts

0:13 are right in this very functional organization.

0:16 And then you have Google or Facebook developed products in totally different

0:20 ways.

0:20 What is the revenue system for building products?

0:23 So we just separate a small team of 10 people or smart people,

0:28 for the owner, good designer, good developers.

0:30 We just build what makes sense from first principles point of view,

0:33 so it can be either replicating existing product, but building it so much

0:38 better,

0:38 or trying new products, which we think makes sense.

0:42 To be honest, everything new is forgotten old, right?

0:45 We haven't built any product which worked, which wasn't an old product,

0:49 which was built badly by other organizations.

0:52 We have internally new bets processing everyone in the company can raise an

0:57 idea,

0:57 and then it goes through validation.

0:58 And then if ID is validated, then it goes to, well,

1:02 called committee of people who've done it before,

1:04 and then they decide whether to find the idea or not.

1:06 What's validation?

1:07 Well, first, the thing goes to team wish validates like, you know,

1:11 in economics, financials, and so on, whether it makes sense.

1:16 The majority of kids make sense, because constantly team is pretty low compared

1:21 to

1:21 if ID is useful for customer, they pay for it.

1:25 You can generate tens of millions of dollars, right?

1:27 Compared to the cost of the team, which is small.

1:29 And then it goes to people who built new bets before and then general managers,

1:35 and then they also validate it from, again, purely common sense perspective.

1:39 And then if everything is approved, then we separate a small team.

1:44 We have people there, additional people, then they build a product within nine

1:49 to 18 months.

1:50 And then we look at the growth of the product.

1:52 So if there is a lot of growth, then we scale up the team.

1:56 So one team becomes two, teams, three teams, four teams, department, and so on.

2:00 And then for the bets which don't grow, we sunset them.

2:03 And then for bets that grow, but not very fast, we'll just leave them alone.

love

· 1d

“未来五到十年，大多数传统金融资产都将上链。我相信这只能在 Solana 上实现。”

Solmate 的 Keren Maimon 和 Ron Sade 谈代币超级周期与 Breakpoint。

x.com/solana

中文 English

0:00 围绕Solana生态系统的公开交易工具非常重要，因为它们将为之前无法进入且无法流入Solana生态系统的机构资本提供一个通道。

0:13 现在我们能够提供这一点。

0:15 作为DIT，我们有责任向市场展示产品和服务。

0:23 ETF做不到这一点。这关乎资本的接入，现在对所有人开放。

0:28 每个人都可以投资，从银行到养老基金再到散户投资者，都能获得数字资产的敞口。ETF开启了将资本市场引入数字资产世界的运动。

0:41 但我认为ETF的角色已经发挥完毕，因为ETF机制有根本性的问题。

0:49 当加密市场牛市时，你会看到资金流入ETF。

0:53 但当加密市场下跌时，ETF自我压力巨大，会指数级地推动整个加密生态的下跌。

1:04 这就是为什么我认为大多数资本市场和加密基金会应该推动80年代模式。

1:11 我也相信，未来五到十年，大部分传统金融资产都将在链上。

1:20 我相信这只能在Solana上实现。

1:22 去年我们参加了Abu Dhabi的Breakpoint，我告诉Karen，这是过去一年中我第一次感觉回到了2017年。

1:31 那个活动周围的兴奋感令人难以言喻。

1:34 这让我们感觉像回到了家，到了正确的地方。

1:38 我们迫不及待想去参加伦敦的Solana Breakpoint了。

0:00 Publicly traded vehicles around the Solana ecosystem is very important,

0:04 because they will provide a gateway between the institutional capital

0:09 that couldn't enter and couldn't flow beforehand into the Solana ecosystem.

0:13 Now we will be able to provide that.

0:15 As a DIT, we have the mandate to present with the market, products and services

0:23 .

0:23 ETF cannot do that. It's about the access to capital, and now it's open to all.

0:28 Everyone can invest from banks to pension funds to retail investors to anyone,

0:33 and get exposure to digital assets. ETF started it, the movement,

0:37 of bringing capital markets into the digital asset world.

0:41 But I think ETF kind of served its role already,

0:44 because there is something fundamentally bothering in the ETF mechanism.

0:49 When the crypto market is in bull run, you will see inflows into the ETF.

0:53 But when the crypto market going down, then there is a massive self-pressure

0:58 from ETF that exponentially creates a downfall in the entire crypto ecosystem.

1:04 This is why I think the majority of capital markets and crypto foundations

1:09 should push for the 80s.

1:11 I also believe on the financial side of things that most of the traditional

1:16 financial assets

1:17 are going to be on-chain in the next five to ten years.

1:20 I believe it can happen only on Solana.

1:22 We were in breakpoint Abu Dhabi last year, and I told Karen,

1:26 it's the first time in crypto for the last year that I feel like it's 2017

1:30 again.

1:31 The excitement around the event was just something we couldn't explain.

1:34 It just made us feel like we are at home, we are at the right place.

1:38 We can't wait for Solana breakpoint in London.

love

· 1d

很高兴宣布，Profound 已完成 1.8 亿美元 D 轮融资，由红杉资本和 Kleiner Perkins 领投。

AI 很快改变了软件工程等结果可验证的领域。如今，它正在各个行业带来同样的变革。

Profound 正在推动 AI 进入大型企业的营销部门。

重新定义全球规模最大、最有趣的行业之一，是一生难得的机会。我们不会浪费它。

x.com/thejamescad

中文 English

0:00 人工智能搜索的兴起为全球每个营销人员带来了难以承受的工作量，这就是为什么Profound筹集了1.8亿美元来打造你的团队AI营销员的原因。

0:11 本轮融资由红杉资本和克莱纳 Perkins 共同领投，他们加倍支持我们的愿景。

0:16 我们不会告诉你接下来会发生什么，而是直接展示给你看。

0:19 认识一下你的Profound AI营销员。

0:23 你的AI营销员是了解你品牌的完美搭档。

0:26 它知道哪些工作值得做，然后替你完成。

0:30 起草精准的推广内容并重新优化发布内容。

0:35 客户推动我们打造一个让人们能实际操作的平台。

0:39 它参与你的通话，甚至在通话结束前提供待办事项。

0:45 过去需要人力完成的工作，现在可以交给你的AI营销员了。

0:54 你的AI营销员知道要标记哪个团队成员进行审核。

0:57 它了解你的受众、产品和策略，始终与你的信息保持一致。

1:02 这一切由Profound Context Manager驱动，实时整合你的邮件、会议和品牌数据。

1:09 你的定位变化，也会同步更新你的AI营销员。

1:12 AI显然是惊人的突破，但开箱即用还远远不够。

1:16 这就是为什么我们结合了对营销流程的深刻理解和Frontier Labs的后期训练方法，为你带来专属AI营销员。

1:24 我们要感谢客户、合作伙伴和投资者，让我们今天能站在这里。

1:29 话不多说，开始工作吧。

0:00 The rise of AI search has created an inhuman amount of work for every marketer

0:04 on the planet,

0:05 which is why a profound has raised $180 million to build your team's AI market

0:10 er.

0:11 Our round was co-led by Sequoia Capital and Kleiner Perkins who are tripling

0:15 down on our vision.

0:16 Instead of telling you what's next, we're going to show you.

0:19 Meet your AI marketer in profound.

0:23 Your AI marketer is a teammate with perfect context on your brand.

0:26 It knows what work is worth doing, and then it does it for you.

0:30 Drafting targeted outreach and staging re-optimized content.

0:35 Our customers pushed us to become a platform where people can do the work.

0:39 It sits in your calls and will deliver action items before you even hang up.

0:45 The work that used to require headcount can now be done with your AI marketer.

0:54 Your AI marketer knows which team-age to tag for reviews.

0:57 It knows your audience, offerings, and strategy, so it's always in line with

1:01 your messaging.

1:02 It's all powered by profound context manager,

1:05 a living synthesis of your email threads, meetings, and brand data.

1:09 So in your positioning changes, so does your AI marketer.

1:12 AI is obviously an incredible breakthrough,

1:14 but it's not enough when it's straight out the box.

1:16 That's why we leave the reminder deep understanding of marketing workflows

1:19 and the post-training methods from the Frontier Labs to bring you your AI

1:23 marketer.

1:24 We want to say thank you to our customers, our partners, and our investors

1:28 for allowing us to be here today.

1:29 With that being said, let's get to work.

love

· 1d

在彭博社最新发布的随黄仁勋出行采访的视频中，他被问及英伟达的接班计划：

“首席执行官应该尽可能保持思维灵活。你必须对自己负责，也要有自主决策的能力。担任首席执行官的压力很大，而且理应如此。这份工作不应该轻松，也不应该被视为理所当然。我希望公司的下一任首席执行官已经在公司工作。不过，我希望自己在未来 20 年里保持健康，并继续证明自己配得上这份工作。”

x.com/amitisinvesting

中文 English

0:00 随着公司规模爆炸性增长，问题变得更加复杂，您的管理风格和心态有何变化？

0:10 我的管理风格正是让公司得以成长的原因。

0:14 我们的组织结构使公司得以繁荣和创新。

0:18 我想打造一个尽可能少壁垒的公司架构。

0:23 你知道，壁垒的问题在于，如果组织过于僵化，公司就很难适应变化。

0:32 所以我认为现代公司越来越需要灵活，结构要越来越少。

0:39 那管理这家公司的CEO呢？

0:41 你必须尽可能保持思维灵活。

0:44 你知道，我认为你必须对自己负责。

0:46 你有行动力。拜托，你是公司的CEO。

0:49 你必须有高标准和高期望。

0:51 你的儿子和女儿都在英伟达工作。

0:54 我听说你的女婿也在英伟达工作。

0:56 是的，在机器人部门。

0:57 有人说你的女儿将成为你的接班人。

1:00 那是真的吗？

1:01 首先，我不希望这个职位被直接继承。

1:05 原因是这要看斯宾塞或麦迪逊的志向。

1:11 他们必须自己争取到这个职位。

1:13 这确实非常艰难，需要巨大的牺牲。

1:16 对于很多人来说，他们可以专注于工作和家庭。

1:21 而你作为CEO，除了当然要关注家庭外，

1:25 我还要照顾很多其他“家庭”。

1:28 我对整个行业负有责任。

1:31 所以我认为作为CEO压力巨大，这也是应该的。

1:38 这不该是一份轻松的工作。

1:41 也不该是一份被视为理所当然的工作。

1:43 当我决定做CEO时，我没意识到压力和责任有多大。

1:50 但这就是我选择的道路。

1:53 现在我的家人和身边的人都看到这份工作的责任和付出，

2:09 他们必须问自己，这是否是你愿意承担的牺牲和责任。

2:14 我真心希望下一任CEO已经在公司里了。

2:21 但20年后，我希望自己依旧健康，继续有价值。

0:00 As the company has exploded in size, and the problems have become more complex,

0:05 how has your management style and your mentality changed?

0:10 Well, my management style is what made it possible for the company to grow.

0:14 Our organization is what made the company to thrive and innovate.

0:18 I wanted to create an architecture of a company with as few silos as possible.

0:23 You know, the problem with silos, if the organizations are too rigid,

0:28 then the company has very difficult time adapting to change.

0:32 So I think a modern company really wants to be more and more agile and less and

0:37 less structured.

0:39 What about the CEO who runs that company?

0:41 You should be as mentally flexible as possible.

0:44 You know, I think you have to hold yourself responsible.

0:46 You have agency. You're the CEO of the company for God's sakes.

0:49 You have to have high standards and expectations.

0:51 Your son and your daughter are working in NVIDIA.

0:54 I think your son-in-law also works at NVIDIA.

0:56 Robotics, yeah.

0:57 There is talk that your daughter is in line to be your successor.

1:00 Is that true?

1:01 Well, first of all, I don't hope the job honored.

1:05 And the reason for that is because that's their aspiration, Spencer or Madison

1:10 's.

1:11 They have to go earn it.

1:13 And it's really, really hard work. It takes a lot of sacrifice.

1:16 You know, for many people, they can focus on their job and their families.

1:21 When you're the CEO of the company, you have to focus on, of course, your

1:24 family.

1:25 But I've got a whole bunch of other families I'm caring for.

1:28 I'm responsible for an entire industry.

1:31 And so I think the pressure of being the CEO is high and it's supposed to be

1:37 high.

1:38 It shouldn't be an, it should not be an easy job.

1:41 It shouldn't be a job that you take for granted.

1:43 When I decided to be CEO, you know, I didn't realize the magnitude of the

1:48 pressure and responsibilities.

1:50 But, you know, look, that's what I signed up for.

1:53 And so now that my family and all of the people that are near me see the

2:01 responsibilities of the job and what it takes,

2:03 then they just got to ask themselves, you know, whether this is the level of

2:07 sacrifice and responsibility you want.

2:09 And I'm really hoping that the next CEO of the company is already at the

2:12 company.

2:14 But 20 years from now, I hope I stay healthy and I hope I continue to earn my

2:20 keep.

2:21 But so long as I can do that 20 years from now, I hope I hope they're already

2:25 here, but they have to earn it.

love

· 1d

你可以让 Claude 帮你写代码，

让 Instinct 帮你预订航班，

让 ChatGPT 帮你做研究。

Roma 提供了一种全新的多智能体通信方式，让不同的智能体能够相互交流。

它适用于你、你的智能体，以及你团队的智能体。

x.com/jans

love

· 1d

2014 年，我把自己的第一家公司 Loom 卖给了 Dropbox。

今天，我很自豪地宣布推出 Roma 2.0（YC F26）：首个专为 AI 智能体打造的待办事项清单。

每个人使用的智能体都不一样：Claude、Grok Bot、Muse……直到现在，它们都无法协同工作。

Roma 会自动创建任务，并将其分配给合适的智能体，让你和你的智能体共享上下文、协同工作。

欢迎试用 Roma。

x.com/jans

中文 English

0:00 我们已经讨论过这个了。

0:01 这是在哪个聊天里说的？

0:02 这不仅仅是我们的结合。

0:03 [音乐] 我们有解决方案。

0:09 介绍Roman，全球首个为AI助手打造的待办事项应用。

0:13 Roman连接你所有的应用，将每个任务交给合适的助手，让你放心生活。

0:17 AI助手原本是帮我们找回生活，结果却花了我们好多时间去管理，决定用哪个助手。

0:23 Roman解决了这个问题。

0:24 Roman完成了90%的发布工作。

0:27 来看看它的实际表现。

0:28 好的，发布改到周三，通知所有需要知道的人。

0:31 有用户说一个连接器坏了。

0:33 哦，对了，别忘了，周五前要更新产品简报。

0:37 所以，Roman把我的话拆成三个任务，还补全了遗漏。

0:40 我说的是“需要知道的人”，它帮我搞定了剩下的。

0:43 这只是开始。

0:44 我从未自己分配过任务。

0:46 Roman根据我设定的规则，把每个新任务交给合适的助手。

0:50 所以我最强的助手只处理需要修复的bug，不用写邮件。

0:55 这帮我每周省下几百美元。

0:57 但重点来了。

0:58 我用ChatterPity。

1:00 Cody信赖Claude，Toby喜欢Grogbot。

1:03 其他产品让这些助手各自为政。

1:05 Roman把三者放在同一个实时任务清单里。

1:08 所以当用户报告连接器坏了，我的ChatterPity接手，查阅罗马中的文档，修好它，并在任务上记录变更。

1:16 Cody和Toby的助手拿到上下文，更新了着陆页和简报，一条消息都没发。

1:22 当Toby问Grogbot简报为何改动，Grogbot直接从任务历史中回答。

1:28 以前Roman出来前，助手只能单独处理任务，每次换助手或新成员加入，信息都会丢失。

1:35 现在，每个助手的决策都被记录并能用简单英文搜索，大家都知道责任人和原因。

1:41 Roman终于让助手实现多人协作。

1:44 Roman，为你的AI助手打造的待办事项清单。

0:00 We already went through this.

0:01 Which chat was that in?

0:02 That isn't simply our union.

0:03 [MUSIC]

0:08 We have the solution for that.

0:09 Introducing Roman, the world's first to-do list app for your AI agents.

0:13 Roman connects to all of your apps and hands every task to the right agent

0:16 while you live your life.

0:17 AI agents were supposed to give us our lives back,

0:19 and instead we're spending hours managing them and deciding which agent to use.

0:23 Roman fixes that.

0:24 Roman handles 90% of this launch.

0:27 Let's see it in action.

0:28 Okay, so the launch move to Wednesday, update whoever needs to know.

0:31 One of our users said that a connector is broken.

0:33 Oh, and before I forget, we need to update the product newsletter before Friday

0:37 .

0:37 So, Roman turns my ramble into three tasks, and it fills in the gaps.

0:40 I said whoever needs to know, and it figured out the rest.

0:43 And that's just the start.

0:44 I never assigned a task myself.

0:46 Roman hands every new task to the right agent using the rules I set once.

0:50 So my most powerful agent only works on the bug that needs it, not on drafting

0:54 emails.

0:55 And that saves me hundreds of dollars a week.

0:57 But here's the big one.

0:58 I use ChatterPity.

1:00 Cody swears by Claude, and Toby loves Grogbot.

1:03 Other products keep these agents apart.

1:05 Roman puts all three on the same live task list.

1:08 So when a user reported a broken connector, my ChatterPity picked it up, and

1:12 read it in

1:13 our docs in Roma, it fixed it and logged the changes on the task.

1:16 Cody's and Toby's agents got the context and updated the landing page and the

1:20 newsletter

1:20 without a single text.

1:22 And when Toby asks Grogbot why the newsletter changed, Grogbot answers straight

1:26 from our

1:27 task history.

1:28 Before Roma, agents were stuck in single play mode, and context got lost every

1:32 time you

1:32 switched an agent or a new team member joined.

1:35 Now, every agent's decisions are logged and searchable in plain English, so

1:38 everyone knows

1:39 who to want and why.

1:41 Roma finally makes agents multiplayer.

1:44 Roma, the to-do list for your EI agents.