Make
集里怎么说它
- 《Figma Config 2026:当代码、AI 与动效都变成画布上的设计材料》(02:28起):本集说它是驱动所有代码层背后的工具，最近新增了计划模式、语音转文字等功能，还能让你在本地电脑上直接对生产环境的真实代码库进行操作。
- 《产品经理驾驭 Claude 生态：用五层架构打造专属 AI 幕僚长》(00:54起):被提及的传统自动化工具。对比指出，过去用 Make 需要一个个连线、拼框，一旦一处失败全盘崩溃；现在用自然语言向 Claude 下达指令即可。
① 提到它的金句
279 条
我会去寻找那些能把不精确变成优点的问题领域。
指向原始笔记的链接
I would look for problem domains that make a virtue out of imprecision.
—— Mustafa Suleyman · [28:26]
如果你在卖给一家对冲基金，他们根据你产品算出的一个数字做了交易，而这个数字是错的，他们就再也不会用你的工具了。
指向原始笔记的链接
If you’re selling to a hedge fund and they make a trade based on a number that comes out of your products and the number’s wrong, they are never going to pick up your tool again.
—— John Willett · [21:52]
从来没有人坐在会议上说，「噢，忘掉清晰吧。让我们试着让它尽可能令人困惑。」
指向原始笔记的链接
Nobody ever sat in a meeting and said, “Oh, forget clear. Let’s try to make it as confusing as possible.”
—— Bob Baxley · [38:39]
找到那些你真的能赚到钱的细分领域，因为竞争等于更低的利润率。
指向原始笔记的链接
find those niches where you can actually make real money because competition equals lower margin.
—— Andrew Wilkinson · [07:04]
所以基本上，对于每一个工作单元，你应该让下一个工作单元更容易做。
指向原始笔记的链接
So basically, for every unit of work, you should make the next unit of work easier to do.
—— Dan Shipper · [42:10]
我称之为隐形生产力。它是那种能让你周围的其他人都变得更好的生产力。
指向原始笔记的链接
I call this invisible productivity. It’s productivity in which you make everybody else around you better.
—— Chip Connolly · [25:49]
你还必须愿意接受 AI 至少在某些时候能比你的某些人类做出更聪明的决策。
指向原始笔记的链接
You have to also be willing to accept that the AI can make smarter decisions than at least some of your humans some of the time.
—— Evan Reiser · [15:06]
所以我们的目标之一是确保，每当我们收到来自客户的反馈，它都能在 24 小时内进入我们的代码库。
指向原始笔记的链接
So one of our goals is to make sure that, like, every time we cut feedback from customers, that gets to our code base in 24 hours.
—— Evan Reiser · [27:06]
我编造了它吗？是的，绝对是我编造的，但它仍然是一个目标，是团队要去思考的事情，一千个实验本身就是。
指向原始笔记的链接
Did I make it up? Yes, absolutely, I made it up, but it’s still a target and a thing for the teams to think about and a thousand experiments by itself.
—— Albert Cheng · [58:08]
让简单的事情保持简单，让复杂的事情成为可能。
指向原始笔记的链接
Keep simple things simple, make the complex things possible.
—— Dylan Field · [82:50]
我就会说，好吧，如果改进不多，那你为什么要花那么多时间去辩论那些对你的性能没有多大影响的事情？
指向原始笔记的链接
I was like, Okay, if it’s not much improvement, then why do you want to spend so much time debating something that doesn’t make that much difference to your performance?
—— Chip Huyen · [06:10]
在这个中间过程中，既然每个人都有非常相似的预训练数据，那后训练就是如今他们做出大差异的地方。
指向原始笔记的链接
where like post-trading, but middle course of this is more of everyone have very similar pre-training data, is that post-training is where they make a big difference nowadays.
—— Chip Huyen · [15:10]
我是一个在智力上非常无畏的人,我必须说当我雇佣年轻人时,我会寻找这一点,因为我认为如果一个人想要做出改变,这是一个非常重要的品质,那就是当你想要做出改变时,你必须接受你正在创造新事物或正在潜入新事物。
指向原始笔记的链接
I’m an intellectually very fearless person, and I have to say when I hire young people, I look for that because I think that’s a very important quality if one wants to make a difference, is that when you want to make a difference, you have to accept that you’re creating something new or you’re diving into something new.
—— Dr. Fei-Fei Li · [66:05]
如果你的软件阻止我并要求我做出决定，而我真的不理解它，你会让我觉得自己很蠢。
指向原始笔记的链接
If your software stops me and asks me to make a decision and I don’t really understand it, you make me feel stupid.
—— Stewart Butterfield · [00:58]
我绝对认为 vibe coding 被过度炒作了。我认为人们没有意识到它在长期内会让你的系统变得多么不可维护。
指向原始笔记的链接
I definitely think that vibe coding is over-hyped. I think people don’t realize how much it’s going to make your systems unmaintainable in the long-term
—— Edwin Chen · [51:56]
所以当 Claude 未能做某事，或者我看到这个非常糟糕的错误，表明 Claude 真的没有理解某些东西时，我会问它，“是什么在你的系统提示或工具中导致你犯了这种错误？”
指向原始笔记的链接
So when Claude will fail to do something or I’ll see this really bad bug that shows that Claude really didn’t understand something, I’ll ask it, “What in your system prompt or tooling made you make this mistake?”
—— Zevi Arnovitz · [46:31]
确保你是为模型的发展方向而构建，而不是为了它们今天的样子。
指向原始笔记的链接
Make sure you’re building for where the models are going and not where they are today.
—— Sherwin Wu · [01:02]
我的一般建议，我给人们这个建议已经有一段时间了，我认为今天仍然适用，那就是确保你是为模型的发展方向而构建，而不是为它们今天的状态而构建。
指向原始笔记的链接
My general advice, and I’ve been giving this to people for a while and I think it’s still true today is make sure you’re building for where the models are going and not where they are today.
—— Sherwin Wu · [49:08]
你会意识到杠杆最高的做法其实是把出问题的地方捕获进一条评测，这样你做的任何改动你都知道是真正有效的，不会破坏你之前努力取得的所有其他成果。
指向原始笔记的链接
You realize that the highest leverage thing you can do is actually try to capture what’s going wrong in an eval so that any change that you make you know is actually going to work and it’s not going to break all the other stuff that you’ve worked towards.
—— Ankur Goyal · [05:49]
人们完全误解了高管是如何做决策的,在他们的头脑中、日程表中、激励结构中发生了什么。
指向原始笔记的链接
People completely misunderstand how executives make decisions, what is going on in the heads, in the calendars, in the incentive structures of executives.
—— Jessica Fain · [06:51]
思考如何让最终用户感觉像赢家是构建有用智能体的一个非常强大的模式。
指向原始笔记的链接
thinking about how do I make the end user feel like a winner is a really powerful model to build a useful agent.
—— Claire Vo · [55:24]
这就是我在这些工具上如此成功的原因是因为我有 20 年以上的管理经验。我知道如何让员工成功。这就是你需要让这些智能体工作的东西。
指向原始笔记的链接
that is why I have been so successful with these tools is because I have 20 years plus of management experience. I know how to make an employee successful. That is what you need to make these agents work.
—— Claire Vo · [86:29]
如果你搞不定这个,那么你关于公司做出的其他决定都没有长期意义,因为你不会是做决定的那个人。
指向原始笔记的链接
If you don’t get this right, no other decision you make about your company will matter for the long term because you’re not going to be the one making it.
—— Eric Ries · [12:12]
只做存款,绝不做取款。
指向原始笔记的链接
Only make deposits, never make withdrawals.
—— Eric Ries · [71:45]
而对开源已经变得多好的一个印证，就是有多少人在使用开源模型，因为没有人想做那种取舍。
指向原始笔记的链接
And the testament to how good open source has gone is how many people are using open source models because no one wants to make that trade-off.
—— Julian · [17:57]
模型总体上做的是让昨天的人类能力变得廉价。
指向原始笔记的链接
What models do in general is they make yesterday’s human competence cheap.
—— Dan Shipper · [00:40]
自动化是个谎言，从某种意义上说，每次你自动化某样东西时，为了确保自动化运行良好，你需要一个人在上面确保它运行良好。
指向原始笔记的链接
Automation is a lie, in the sense that every time you automate something, in order to make sure the automation is working well you need a human on top of it making sure that it’s working well.
—— Dan Shipper · [39:15]
现在你可以轻松地说，嘿，给这个仓库提个 PR 做 x、y 和 z，容易到困难的部分变成了被信任去为这些仓库做贡献。
指向原始笔记的链接
now the ability to go, like, hey. Make a PRT this repo doing x, y, and zed is so easy that then the hard part has become being trusted to contribute to these repos
—— Thomas Ankcorn · [08:28]
我学到了，你制造产品，你修复产品，然后你修复业务。
指向原始笔记的链接
I’ve learned, you make the product, you fix the product, then you fix the business.
—— Tony Fadell · [31:06]
你的现有竞争对手的最大痛点是什么？那么，我要做什么？我要在他们最大的痛点上极其用力，把它变成最令人愉快的体验。
指向原始笔记的链接
What is the biggest pain point of your incumbent? And, like, what am I going to do? I’m just going to go incredibly hard at their biggest pain point and make it the most delightful experience.
—— Barney Hussey-Yeo · [18:01]
但是如果你拿走了她喜欢的东西,并且你把它做得稍微好一点,她可能会更喜欢那个,而不是你向她展示她从未见过的东西、而且她醒来时并不知道自己想要的东西。
指向原始笔记的链接
But if you took something that she loves and you make it one inch better, she might love that more than if you showed her something she’s never seen before and didn’t wake up knowing that she wanted.
—— Mark Pincus · [17:06]
但如果不致力于一个单一的想法，在创业公司上取得有意义的进展是极其困难的。
指向原始笔记的链接
But it’s extremely hard to make meaningful progress on a startup without committing to a single idea.
—— John · [00:25]
没关系，我有这样一句话：犯新的错误。犯错没关系，只要犯新的错，这样我们就能不断学习。因为如果你追求零错误，那大概意味着你行动不够快，或者有点太谨慎了。
指向原始笔记的链接
it’s okay to like I have this saying make new mistakes like it’s okay to make mistakes just make new ones so that we’re always learning because if you aim to make zero mistakes like that probably means you’re not you know moving fast enough or being a little bit too cautious
—— Fiona Fung · [11:28]
如果你只是把一个模型做得越来越大,它不会变得更安全。
指向原始笔记的链接
If you just make a model bigger and bigger, it will not get safer.
—— Zico Kolter · [27:32]
去创造那些只有你能创造的东西吧。
指向原始笔记的链接
Go make something that only you can make.
—— Dylan Field · [70:39]
我现在能想到的每一份工作，我都可以提出理由说它绝对有可能被取代。
指向原始笔记的链接
every job I can currently think of, I could make a case, potentially, that it could absolutely be replaced.
—— Dan · [17:03]
为什么你要让一个智能体去阅读数百个 Kubernetes 文件并编写甚至没有类型的 YAML，而它基本上可以在一个装饰器中做几个更改，然后获得这种能够自我配置的运行时，实时看到它的更改在运行中生效。
指向原始笔记的链接
Why would you have an agent read through hundreds of Kubernetes files and write YAML that’s not even typed when it can basically make a couple of changes in a decorator and it gets this sort of self-provisioning runtime of being able to see its changes live in action.
—— Akshat Bubna · [05:06]
不,我认为在一个更容易构建事物的世界里,品味非常重要,确保你的时间花在首先弄清楚你应该构建什么上变得更加重要。
指向原始笔记的链接
no i think taste matters a ton in a world where it’s easier to build things it’s more important to make sure that your time is spent figuring out what you should be building in the first place
—— Adam Mosseri · [00:00]
我认为那些能充分利用它的人,是那些对AI擅长什么和不擅长什么有清醒认识的人,并且对它将来会擅长什么和不擅长什么有一种本能或嗅觉。
指向原始笔记的链接
the people who i think are gonna make the most of it are the ones who are clear-eyed about what ai is good at and what it’s not good at and also have an instinct or a nose for what it will be good at and not good at
—— Adam Mosseri · [19:21]
智能体运行内部执行循环。我设定方向，我在外部循环中做决策。
指向原始笔记的链接
The agent runs the inner execution loop. I set the direction and I make decisions in the outer loop.
—— 嘉宾 · [23:04]
我看到的创始人犯的最大错误之一是依赖汇总用户指标，而不是理解单个用户如何使用他们的产品。
指向原始笔记的链接
One of the biggest mistakes I see founders make is relying on aggregate user metrics instead of understanding how any individual users use their product.
—— 嘉宾 · [00:09]
现在这些东西，如果你看它，我怎么理解它，除非我是一个有 20 年工作经验的在岗工程师？
指向原始笔记的链接
Now these things, if you look at it, how do I even make sense of it unless I’m an engineer on the job with 20 years of experience?
—— Kriti Sharma · [09:46]
现在的魔力在于，以前这需要很多人，可能甚至数百名工程师，数年的努力，才能使某种技术或软件在制造业供应链中工作。
指向原始笔记的链接
The magic now is previously it would have taken a lot of people, maybe possibly even hundreds of engineers, years worth of effort to make some sort of technology or software work for manufacturing supply chain.
—— Kriti Sharma · [12:43]
Fable 非常像一种为智能体服务、由智能体组成的沟通机制，我几乎听不懂它在说什么。
指向原始笔记的链接
fable is very much like a for agents by agents communication mechanism i can barely make out what it’s talking about
—— Alex Finn · [20:01]
我认为不管是现在的 AI 还是敏捷或任何变化，听起来很难，但在开始时不要在消极的事情上花时间。找到成功的故事。如果你不能让它成功，其余的都不会起作用。
指向原始笔记的链接
I think regardless whether this is now AI or like agile or any change, um, as hard as it sounds, do not spend time on the negatives in the beginning. Find the success story. If you can’t make that work, the rest will not work.
—— Patrick Debois · [33:58]
我认为当你开始记下来、列清单、记录东西时，你会觉得，好吧，现在你觉得有义务对它做点什么或用它做点什么。
指向原始笔记的链接
I think when you begin to jot things down and make lists and record stuff, you’re like, well, now you feel like there’s an obligation to do something about it or do something with it.
—— 嘉宾 · [01:34]
几乎没有什么钱是我不愿意合理花出去的,只要能确保我们的某个客户获得一个好的结果。
指向原始笔记的链接
There’s almost no amount of money that I wouldn’t reasonably spend to make sure that one of our customers gets to a good outcome.
—— Eric Sager · [33:50]
所以无论我们在哪里建数据中心，我们都会做出硬性承诺，即我们不会从电网夺走电力。事实上，我们是在投资电网以产生新的电力，这样我们才能把电力用于数据中心。
指向原始笔记的链接
So whenever we build a data center anywhere, we make it a hard commitment that we are not taking power away from the grid. In fact, we are investing in the grid to generate new power so that we can consume it for data centers.
—— Sachin Katti · [08:44]
这里的独特优势在于，我们知道未来的模型可能长什么样，因此能够走捷径，省去很多你在芯片侧需要做的决策、设计决策。
指向原始笔记的链接
There’s a unique advantage here of us knowing what the future models might look like and therefore being able to short circuit a lot of the decisions you need to make, design decisions you need to make on the chip side.
—— Sachin Katti · [35:26]
这是一个万亿美元的行业,他们每年在资本开支上花 2000 亿美元,而他们不赚钱。
指向原始笔记的链接
And it’s a trillion-dollar industry, and they spend $200 billion a year on CapEx, and they don’t make any money.
—— Benedict Evans · [05:06]
所以,可能是一个智能体编写了代码,或者我帮助进行了分析,而这并不是我的背景,但这并不意味着人们不需要对他们所创造的东西带来的责任负责。
指向原始笔记的链接
So, it can be that an agent wrote the code or I helped to do an analysis when that’s not really my background, but it doesn’t make people not have the responsibility that comes with what they’ve created.
—— Elizabeth Stone · [06:02]
我们现在看到的是，很多人、很多公司意识到，完全依赖一个模型太危险、风险太大，完全说不通。
指向原始笔记的链接
I think what we’re seeing right now is that a lot of people, companies are realizing that it’s too dangerous, it’s too risky, it doesn’t make any sense to rely exclusively on one model.
—— Clement DeLong · [17:10]
我们就会构建错误的产品。我们绝对会的，然后就得艰难转型、烧钱、做出艰难的抉择——而不是先摸清市场，感受一下它究竟是什么样子。
指向原始笔记的链接
We would have built the wrong solution. We absolutely would have and we would have had to hard pivot, burn money, make some tough choices versus case the market, get a flavor for what that actually is.
—— Damien Lewke · [30:41]
如果你提供一个有价值的产品,资本主义会找到一种方法让供应链为你工作。
指向原始笔记的链接
if you’re providing a product of value, capitalism will find a way to make the supply chain work for you.
—— Sriram Krishnan · [19:24]
因为你达到 AGI 却仍然输掉的唯一方式,就是如果你让那个 AGI 可供其他人使用。
指向原始笔记的链接
Because the only way you reach AGI and still lose is if you make that AGI available to others.
—— Alex · [23:31]
代码智能体非常善变、不可预测,它们经常犯错,然后当你真的很倒霉时,它们会发疯并试图销毁你的数据以及其余的一切。
指向原始笔记的链接
Coding agents are really fickle, unpredictable, they often make mistakes, and then when you’re really unlucky, they’ll go crazy and try and nuke your data and all the rest of it.
—— 嘉宾 · [17:00]
我认为人们有时犯的一个错误是太快地把事情拆分，并说像，噢是的，让我尝试创建一个真正复杂的多智能体系统并做事情。
指向原始笔记的链接
I think one mistake that people sometimes make is splitting things up too quickly and saying like, oh yeah, let me try to create like a really complex multi-agent system and do stuff.
—— Harrison · [31:06]
但你有一件他们没有的东西：做出疯狂事情的能力，去冒一些险。
指向原始笔记的链接
But the one thing you have that they don’t is the ability to make crazy things, to take some risks.
—— Alex · [13:59]
因为人是多模态的,你做出最佳决策的最好方式,就是尽可能多地接收各种感官输入
指向原始笔记的链接
Because humans are multimodal, the best way that you make the best decisions is to take as many sensory inputs as you can accept.
—— Eric Landau · [09:08]
我们更希望有真正好的竞争对手,因为他们会推着你,让你成为更好的公司
指向原始笔记的链接
We prefer to have really good competitors because they push you, they make you better as a company.
—— Eric Landau · [17:05]
所以我非常小心，绝不把判断或决策委托给模型，因为它们会让你以为它做的是对的事，但你仍然必须对它们保持偏执。
指向原始笔记的链接
So I want very careful never to delegate judgment or decision-making to models because they make you think that it’s doing the right thing, but you have to be paranoid with them still.
—— Osvald Nitski · [26:37]
但要让一个人真正在乎某件事，真的很难。
指向原始笔记的链接
It’s really hard to make someone give a shit.
—— Osvald Nitski · [49:55]
仿真扮演了一个非常重要的角色,这是现实世界的数据所没有的,那就是反事实推理,就是你在推演那些尚未发生或不可能发生的事件,或者你在现实世界中没有足够的数据让它发生。
指向原始笔记的链接
There’s a very important role simulation plays that real world data doesn’t play, which is counterfactual reasoning, is that you play out events that hasn’t happened or cannot happen, or you don’t have enough data to make it happen in real world.
—— Fei-Fei Li · [20:06]
智能体不介意烦人的刁钻规则，它们可以很开心地去撞墙让事情跑起来。
指向原始笔记的链接
Agents are fine with annoying Infiniki, they can very happy to go bang their heads against the wall in order to make something work.
—— Rob Willoughby · [27:31]
关键在于让硬件开发看起来更像软件开发，无论在比特的世界还是原子的世界里，都要降低迭代的成本。
指向原始笔记的链接
The key thing is to make hardware development look more like software development, to reduce the cost of iteration, both in the world of bits and in the world of atoms.
—— Blake Scholl · [12:34]
这简直解锁了人们想象中所有那些疯狂的、怪异的、有趣的想法,以前他们是无法制作出来的。
指向原始笔记的链接
It sort of just unlocks all of this crazy, weird, interesting ideas that people have in their imagination that they could never make before.
—— Justine Moore · [00:53]
关键是让硬件开发看起来更像软件开发,降低迭代成本,无论是在比特的世界还是原子的世界里。
指向原始笔记的链接
The key thing is to make hardware development look more like software development, to reduce the cost of iteration, both in the world of bits and in the world of atoms.
—— Blake Scholl · [12:34]
当公司达到一定规模，如果你审计某人在做什么，可能有 30% 到 50% 的时间，不是在做让产品更好、服务客户等等的实际工作。
指向原始笔记的链接
You get to a certain size company and it might be 30 to 50% of the time if you audit what someone’s doing, it’s not doing actual work to make the product better, deliver for a customer, on and on.
—— 嘉宾 · [26:13]
我认为，表现最好的 AE 应该比公司里的很多人赚得多得多。
指向原始笔记的链接
I think people, like, the highest performing AE should make a lot more than a lot of people in the company.
—— Lindsey Scrase · [39:53]
我并不是真的来削减你的成本的。当然，这正在以这种方式发生，但我真正感兴趣的是你将如何获得全新的收入。
指向原始笔记的链接
I’m not really there to cut your costs. Sure, that is happening in this way, but I’m really interested in this is how you’re going to make net new revenue.
—— Melisa Tokmak · [30:45]
我在很多事情上失败了。我想确保代表我的模型以同样的方式失败。
指向原始笔记的链接
I failed at a lot of things. I want to make sure that the model that represents me fails the same way.
—— Joon Sung Park · [52:03]
我认为品味不仅仅是美学偏好。品味是一种让人们有美德地消费的方式。
指向原始笔记的链接
And I make the case that taste is not just aesthetic preference. Taste is a way for people to consume virtuously.
—— Ruby Thelot · [17:36]
但是直到你按下琴键并把它一直推回到 V1,你才是在演奏音乐。
指向原始笔记的链接
But you don’t make music until you press the key and push it all the way back into V1.
—— Tom Verrilli · [62:47]
我们创办公司不仅仅是为了制造一只快 10 倍的老鼠,对吧?我们开始这是为了制造以前不可能的突破性药物。
指向原始笔记的链接
We didn’t start the company just to make a 10 times faster mouse, right? We started this to make breakthrough medicines that weren’t possible before.
—— Josh · [39:24]
它不是为了制造一个或两个分子。它不是为了获得一个由五种我们推向市场的有趣疗法组成的管线。它是药物发现方式的改变。
指向原始笔记的链接
it’s not to make one or two molecules. It’s not to get a pipeline of like five interesting therapies that we bring to market. It’s the change of the way that medicines are discovered.
—— Josh · [25:07]
为了让这行得通，有两个因素需要汇聚。一个是发射成本下降，人们必须相信这一点。第二是，在地面建设基础设施正变得极其困难。
指向原始笔记的链接
To make this work, there’s a convergence of two factors. One is the launch costs coming down, and people have to believe that. The second is it becoming way harder to build stuff terrestrially.
—— Philip Johnston · [15:02]
但我们能控制的是这场演出。如果规则由我们来定——比赛的规则、演出的规则——人们就可以每晚离开时都玩得非常开心。
指向原始笔记的链接
But what we can control is the show. And if we build the rules, we make the rules of the game, we make the rules of the show, people can leave and have a great time every single night.
—— Jesse Cole · [10:21]
就像在人类历史的绝大部分时间里，如果你是一个聪明的 20 岁年轻人，有了一个让自己的社会变得更好的想法，你把它讲给有钱人听，反应会是：你应该去关心收成。
指向原始笔记的链接
Like for the vast majority of human history, if you were a smart 20 year old that like had an idea to like make your society better and you raise this to the people with capital, the reaction was like, you should pay attention to the harvest.
—— Max Hodak · [46:47]
你必须看到世界没有看到的东西，下早期的赌注，然后在市场时机上运气好一点。
指向原始笔记的链接
You have to see something the world is not seeing, make early bets and then get lucky with market timing.
—— Navin Chaddha · [06:09]
10% 的公司创造了所有的回报。
指向原始笔记的链接
10% of companies make all the returns.
—— Navin Chaddha · [51:45]
把我们庞大的软件行业想象成前工业时代是很可笑的,但在某种程度上确实是,因为到目前为止我们拥有的唯一自动化就是启发式的,比如确保每行末尾有一个分号。
指向原始笔记的链接
It’s funny to think about our giant software industry as being pre-industrial, but on some level it is, because until now the only automations that we had were heuristics, like make sure there’s a semicolon at the end of every line.
—— Idan Gazit · [02:57]
如果你作为工程师让智能体做出关键决策，你就是在放弃你的代码的控制权。
指向原始笔记的链接
If you as an engineer are letting an agent make a critical decision, you are ceding control of your code.
—— Matt Dailey · [03:27]
树病了,所以为了让它健康,它把树砍倒了。
指向原始笔记的链接
The tree is sick, so to make it healthy, it cuts it down.
—— Joel de la Garza · [10:44]
因为如果你没有好的系统来做增量改进，你就会一直玩打地鼠的游戏。
指向原始笔记的链接
Because if you don’t have good systems in place to make incremental improvements, you’re constantly playing this game of whack-a-mole.
—— Ali Haghani · [10:36]
我们发现,随着模型演变得更有能力,如果你将逐字稿和记忆状态作为一个周期性批处理过程,并使用我们所谓的做梦,它允许我们提取新的洞察和新的组织结构,这基本上会反馈并编辑记忆,从而让第二天的智能体会话自动地变得更加智能。
指向原始笔记的链接
What we found is that as models have evolved and become more capable, if you feed the transcripts and the memory state as a periodic batch process with what we call dreaming, it allows us to extract new insights and new organized structures that essentially feedback and edit the memory as needed to make the next day’s agent sessions automatically much more intelligent.
—— Gagan Bhat · [27:45]
你不需要给所有东西都打上 AI 的品牌。你只要把它做得真正有用,然后无论人们对这项技术感觉如何,他们都会开始使用它。
指向原始笔记的链接
You don’t need to brand everything AI. You just make it really useful and then people will start using it regardless of how they feel about the technology.
—— 嘉宾 · [20:12]
我们就是那个要让所有人免费获得超级智能的团队。我们唯一的要求是,让我们建一些数据中心。
指向原始笔记的链接
We’re the group that’s going to make super intelligence available for all people for free. All we ask is that you let us just build some data centers.
—— 嘉宾 · [34:43]
我确切知道那样做是在次优化我做事情的一部分，但我还是做了这个决定，因为我可以自己选择。
指向原始笔记的链接
I know for sure that that is sub-optimizing a part of what I’m doing, but I make the decision anyway because I get to choose.
—— Kevin Scott · [18:13]
所以我们不仅想让人们检查眼睛、拿到眼镜处方变得难以置信地容易，还想把成本降得足够低，以至于你真的没有任何理由不做这件事。
指向原始笔记的链接
So we want to not only make it unbelievably easy for people to get their eyes checked and to get an eyeglass prescription, but also we want to lower the cost so much that you really have no reason to not do this.
—— 嘉宾 · [37:22]
你不能真正用 AI 来构建你的 Kubernetes YAML,因为 AI 会犯错，而当它犯错时你必须能够理解，因为某个东西会在凌晨 2 点停止工作，然后 AI 其实没办法告诉你
指向原始笔记的链接
You can’t really build your Kubernetes YAMLs using AI because AI makes mistakes and you have to understand when it does make a mistake because something will stop working at 2 a.m. at night and then AI won’t really be able to tell you
—— Anurag Gohl · [08:36]
但有时世界会打开，在”之前”和”之后”之间会出现一道裂缝。如果你做出那个跨越，你必须快速完成。
指向原始笔记的链接
But there are times when the world opens up and there’s a crevice between the before and the after. And if you make that jump, you’ve got to make it quickly.
—— 嘉宾 · [19:41]
你在系统的任何地方做任何改变，你就会破坏你的缓存。
指向原始笔记的链接
You make any change anywhere in your system and you break your cash.
—— Flo Crivello · [27:32]
我认为很多 AI 实验室想要抢在前面,确保围绕它有某种治理框架,以便他们能以自己想要的速度继续开发这项技术。
指向原始笔记的链接
I think a lot of the AI labs want to get ahead of it, make sure there is some governance framework around it to ensure that they can keep developing the technology at the pace at which they’d like to.
—— Nikesh Arora · [05:59]
人们在收购过程中犯的最大错误是低估了被收购业务的打造者的智慧。
指向原始笔记的链接
The biggest mistake that people make during acquisitions is underestimate the intelligence of the people who built the business that you acquired.
—— Nikesh Arora · [20:52]
一切都必须是掠夺性的吗？不。你能不能提供一个有竞争力、成本更低、还能赚更多钱的产品？能。
指向原始笔记的链接
Does everything have to be predatory? No. Can you provide a product that is competitive, lower cost, and make more money? Yes.
—— Farooq Malik · [15:23]
他们声称可以从一个根本不再学习的东西中制造出博士水平的体验和专业知识。
指向原始笔记的链接
They claim they can make a PhD level experience and expertise out of something that doesn’t learn at all anymore.
—— Rich Sutton · [20:35]
我最近常用来形容的一个说法是,几十上百年积累下来的文化包浆,让它让人觉得熟悉。
指向原始笔记的链接
Like I, a phrase I’ve been coming to use is like this, like patina of culture for like many, many decades that like make it feel familiar, you know?
—— 嘉宾 · [19:52]
但当你让整个团队都这样做，并且没有这些建立起来的营销谎言，也没有你在背后隐藏的东西，试图制作这些精心策划的视频时，你就有了一个能够实际解决正在发生的问题的团队。
指向原始笔记的链接
But as you do that throughout the whole team and you don’t have these kind of marketing lies that are built up and things that you’re hiding behind and trying to make these highly curated videos, you have a team that is actually able to deal with the problem that’s happening.
—— Mike LeBlanc · [43:59]
一种在这里能运作得很好的模式是，如果你在平台中强制执行一个更宽泛的 agent 定义，把 LLM 和人类都囊括在内，那么你就可以做到让任何 LLM 能执行的动作，人类也能执行。
指向原始笔记的链接
One pattern that can work very well here is if in your platform you enforce a wider definition of agent which encompasses both LLMs and humans, then you can make it such that any action that can be taken by an LLM could also be taken by a human.
—— Christopher Lovejoy · [13:49]
而是要想：这个前沿模型需要什么样的全部上下文，才能像我一样做决策。
指向原始笔记的链接
But rather what is all the context this frontier model needs to be able to make decisions like I would make.
—— Hursh Agrawal · [07:29]
但一定要抬高门槛，因为我们不发布垃圾代码。
指向原始笔记的链接
But make sure to raise the bar because we’re not shipping shitty code.
—— Kieran Klaassen · [19:31]
所以 Meta 保证如果它不坚持整整二十年，它会让债券持有人完整收回，但每个人都在押注如果 AI 需求爆发并且如果这些数据中心开始疯狂印钞，然后它们能够支付债券持有人而且还有富余并产生额外的现金，那么 Meta 就不需要承担任何责任。
指向原始笔记的链接
Meta is guaranteeing to make bondholders whole if it doesn’t stay for the entire two decades, but everyone is betting if AI demand explodes and if these data centers start printing cash and then they’re able to pay bondholders and then some and generate additional cash, then Meta’s not on the hook for anything.
—— Ranjan Roy · [11:30]
大多数做沙箱的人都在计算上赚钱。坦白说，在我看来这是个愚蠢的主意。
指向原始笔记的链接
most sandbox people make money on compute. Frankly, that’s a dumb idea in my opinion.
—— Jerry Murdock · [27:19]
我们就是不干这些。我的想法是，让我们把钱花在确保我们以最高水平的服务运营上。然后我们的客户会变成我们的品牌管理者，变成我们最好的品牌大使。
指向原始笔记的链接
We just wouldn’t do it. And my thought was let’s spend the money to make sure we operate at the highest level of service. And then our customers will become our brand managers. They will become our best brand ambassadors.
—— AJ Loiacono · [34:37]
我不能犯错，因为这份工作是如此无情，不是当你投资了错误的公司时，而是当你没有投资于正确的公司时。
指向原始笔记的链接
I cannot make a mistake because this job is so unforgiving, not when you invest in the wrong company, but when you don’t invest in the right companies.
—— Julien Bek · [31:24]
傲慢可能不是你喜欢的东西，但它可能会让他们在所做的事情上非常出色。这可能是他们尖刺的代价。
指向原始笔记的链接
And arrogance might not be something you like, but it might make them very good at what they do. It’s maybe the cost of their spike.
—— Julien Bek · [35:40]
因为今天任何人在每兆瓦 1000 万到 1500 万美元的算力上都能赚钱。
指向原始笔记的链接
because anyone can make money off of 10 to $15 million per megawatt compute today.
—— 嘉宾 · [14:28]
在某个时候，人们开始开这个玩笑：安全已经变成了工作量证明。
指向原始笔记的链接
At some point, people started to make this joke that security has become proof of work.
—— Neil Movva · [10:05]
我的工作就是把 token 做到人类可能做到的最低价。
指向原始笔记的链接
My job is to make the tokens as cheap as humanly possible.
—— Neil Movva · [69:39]
我们没有在美国境外生产哪怕一辆 F-150。
指向原始笔记的链接
We don’t make one F-150 outside the US.
—— Jim Farley · [22:09]
你知道,我的儿子可以从高中毕业,在一家福特经销店当技师,三年赚 15 万美元。
指向原始笔记的链接
You know, my son can graduate from high school, can be a technician in a Ford dealership and make 150 grand in three years.
—— Jim Farley · [25:55]
再多的编码智能体也不会让 Nike 变成不是 Nike。
指向原始笔记的链接
No amount of coding agents is going to make Nike not Nike.
—— Anish Acharya · [06:31]
所以关键点是，你团队中每一个正在做架构决策的人，那些人必须懂性能，而且他们必须做出能让那些在他们下游的其他人使用以后可以优化的架构的决策。
指向原始笔记的链接
the crucial takeaway is everybody on your team who is making architectural decisions, those people must know performance and they must make decisions that will allow the other people downstream of them to use an architecture which can be optimized later.
—— Casey Moratori · [35:57]
我们就是做决定然后去干，因为反正几乎所有事情都是可逆的。
指向原始笔记的链接
We just make decisions and we go for it because most everything can be reversed anyway.
—— Jason · [06:20]
如果你连续做了 500 个决定而且它们都很烂，你会破产。
指向原始笔记的链接
If you make 500 decisions in a row and they all suck, you’re going to go out of business.
—— David · [09:02]
而如果你真的必须思考你做的每一个决定，你哪里也去不了。
指向原始笔记的链接
And if you had to really think about every decision you make, you’d get nowhere.
—— Jason · [12:24]
我认为医疗保健领域刚开始发生的最大转变之一，就是从行业内所说的病患护理（公司在你生病时从你身上赚钱）转向预防（当他们让你保持健康时才赚钱）。
指向原始笔记的链接
And I think one of the biggest shifts in healthcare that was just starting is the shift from what is in the industry called sick care, where companies make money off of you when you’re sick, to prevention, where you make money when they keep you healthy.
—— Vijay Pande · [21:19]
所以我们现在能做的，基本上就是为一个非常具体的任务提出一个全新的分子，然后在实验室里制造它，再把它用于那个特定任务。
指向原始笔记的链接
And so what we can do is we can now basically come up with an entirely new molecule for a very specific task and then make it in the lab and then use that for that particular task.
—— Max Welling · [13:16]
它掩盖了一个事实：写代码从来都不是限制性的环节。也许在时间线层面上是，但它从来都不是决定你初创企业成败的东西。
指向原始笔记的链接
It papers over the fact that writing code was never the limiting part. Maybe it was on the timeline perspective, but it was never the thing that was going to make or break your startup.
—— Mike Krieger · [19:29]
如果你不做测试，不确保针对任务、也针对不断推出的最新模型去维护和验证你的技能，那么质量就会随时间开始退化。
指向原始笔记的链接
If you don’t test and make sure that you’re maintaining and you’re validating your skills, not against your task, but also against the latest models that comes, right, then the quality starts degrading over time.
—— Imad Touil · [11:43]
我并没有被正式指派到 MCP 项目上，也没有被正式指派到我们的 make 产品上。但我最终帮助了它们被构建出来，只是因为我被给予了这样做的自由空间，并在此过程中学到了大量东西。
指向原始笔记的链接
I wasn’t staffed on MCP. I wasn’t staffed on our make product. But I ended up helping them be built just because I was kind of giving the leeway to do so and learning a ton along the way.
—— Jesse Lumarie · [16:10]
所以你能做出的快速决策越多,尤其是那些容易逆转的决策,就越好。
指向原始笔记的链接
And so the more that you can make fast decisions, especially ones that are easy to be reversed, the better.
—— Clare Liguori · [19:51]
把执行摘要放进去给智能体。否则它们会随时间开始漂移，而且要确保智能体不会只因为自己想改，就回头去修改那个部分。
指向原始笔记的链接
Put that in there for the agent. Otherwise, they’ll start drifting over time and make sure that the agent don’t go back and change that because they feel like it.
—— Eyal Blum · [08:51]
如果你有一套好的评估套件，你可以让 Claude、Codex 之类的进去替你修改提示词，而你确信你不会发布任何会毁掉生产环境的东西。
指向原始笔记的链接
If you have a good eval suite, you can let Claude or Codex or Devon kind of go in, make prompt changes for you, let BlinkchainEngine make prompt changes for you, and you know that you’re not shipping anything that is going to ruin production.
—— Vishu · [02:26]
而且我强烈建议，在一个不确定性的系统里，你至少要让请求尽可能具有确定性。
指向原始笔记的链接
And I would highly recommend that you at least make requests as deterministic as possible with an undeterministic system.
—— Kanish Manuja · [09:12]
它看起来像别人。但如果你把它做得非常简单，每个人都可以把自己投射到它上面。
指向原始笔记的链接
It looks like somebody else. But if you make it really simple, everyone can project themselves onto it.
—— Susan Kare · [16:59]
它们想打造一个成功的任务完成的波将金村来呈现给评分系统。
指向原始笔记的链接
They wanted to make a Potemkin village of a successful task completion to present to the score.
—— Ryan Greenblatt · [10:18]
我们这里的北极星目标是用户不需要在所有这些不同的选项之间做决定。理想情况下，这里没有切换开关。
指向原始笔记的链接
Our north star here is that users do not need to make decisions between picking between all these different options. Ideally, there is no toggle here.
—— Tara Seshan · [29:50]
每当智能体尝试进行一笔一美分的交易时，你不可能支付 3% 的手续费。
指向原始笔记的链接
You cannot pay 3% each time an agent tries to make a one-cent transaction.
—— Harshal Bhangale · [04:53]
他真的是开创这个想法的人：你应该只需要给出一条反馈一次，但如果你以正确的方式把它固化下来，它就会永远在那里，而且下次你再处于同样的情境时，AI 会做出一个更好的决定。
指向原始笔记的链接
He was really the one to pioneer this idea that you should only have to give a piece of feedback once, but if you codify it the right way, it will be there forever and the next time you’re in that same situation, the AI is going to make a better decision.
—— 嘉宾 · [32:58]
从9月中旬开始,我们要让它不只是一块「禁止擅入」的牌子,我们会基本上派一个保安站在那里说:除非你补偿创作者,否则你拿不到这些内容。
指向原始笔记的链接
Starting in mid-September, we’re going to actually make it not just a no trespassing sign, but we’re actually going to have, basically put a bouncer there saying, unless you’re compensating creators, you don’t get this content.
—— Matthew Prince · [38:47]
AI 领域的经济账不合理。至少现在不合理。
指向原始笔记的链接
The economics don’t make sense in the AI space. They don’t make sense yet.
—— 嘉宾 · [78:50]
我认为那是胡扯，因为无论是能源、液冷，还是所有能让数据中心变得更好的东西，这些都是可以在这里创造、在这里完成的工作岗位。
指向原始笔记的链接
I call BS on that because if you think about whether it’s around energy, liquid cooling, all of the things that make the data center better, those are all jobs that can be created and done here.
—— Rene Haas · [28:38]
随着模型变得越来越有能力，要确保它们无法得逞这类事情会越来越难。
指向原始笔记的链接
As the models get more and more capable, it’s going to be harder and harder to make sure that they can’t pull off this kind of thing.
—— Buck Shlegeris · [13:22]
就好像它们 98% 是自私的，但在某些情况下，当它们能做一些对蜂群真的很有帮助、而对自己没那么糟糕的事情时，它们就愿意做出那种牺牲。
指向原始笔记的链接
It’s like they’re 98% selfish, but in cases where they can do something which is really helpful for the swarm and not that bad for them, then they’re willing to make that sacrifice.
—— Buck Shlegeris · [20:24]
所以总的来说，存储数据是一个非常巨大的问题空间，我认为它永远不会下沉到模型层。
指向原始笔记的链接
So generally storing data is like a huge problem space that I don’t think will ever make its way down into the model layer.
—— Jeffrey Morgan · [17:21]
我几乎是深切地站在这一边：很难论证 AI 模型可以在宽泛的公共互联网上训练，但另一个 AI 模型却不能用 AI 模型的输出训练。
指向原始笔记的链接
I’m almost deeply on the side of, I think it’s very hard to make the argument that AI models should be trained on broadly the public internet, but another AI model can’t be trained on the outputs of an AI model.
—— Aaron Levie · [03:48]
因为就像明天，如果我发布一个闭源模型，如果我把它保持闭源、强迫一切通过我的 API 路由，我字面意义上就会赚到更多的钱。
指向原始笔记的链接
Because like tomorrow, if I release a closed model, I just will literally make more money if I just keep it closed and I force everything to route through my API.
—— Aaron Levie · [15:46]
很容易说大多数人不想写代码，但我想法上的大转变是，编程智能体其实是一种与世界互动的通用方式。
指向原始笔记的链接
I think it’s easy to say most people don’t want to make code, but the big change in my thinking is that coding agents are a way to interact with the world generally.
—— Anish Acharya · [51:57]
即使有人复制了你的大想法，他们也永远看不到让大想法成立的那十几个小想法。
指向原始笔记的链接
And even if somebody replicates your big idea, they never see the small ideas that make the big idea work.
—— Anish Acharya · [57:40]
我们应该尽量减少每天必须主动做出这些选择的人数，它应该自动为他们做正确的事。
指向原始笔记的链接
We should try to minimize the number of people who have to actively make these choices every day, and just it should do the right thing for them.
—— Sharadh Krishnamurthy · [09:25]
我认为 LLM 会很快比人类少犯很多这类错误。
指向原始笔记的链接
I think the LLMs will make many fewer of these mistakes than humans pretty quickly.
—— JD · [18:18]
所以突然之间，AI 让你多雇了那个增量的人——因为在新的世界里工程师效率更高，雇他能让你多赚 5 万美元的利润。
指向原始笔记的链接
So suddenly AI makes you hire the incremental person, one more person than you would have because there’s an extra 50K of profit for you to make by hiring the engineer in the new world because they’re more efficient.
—— JD · [65:24]
为此，我们真的希望人们觉得输入 token 是免费的——只管为你的问题考虑所需要的上下文就好。
指向原始笔记的链接
To make it, we really want people to feel like the input tokens are free, like just consider the context that you need to for your problem.
—— Alexander Whedon · [45:32]
SpaceX 成功的最大关键之一就是 Elon 是那种会对他的团队下巨大赌注的人，相信我们能做到。
指向原始笔记的链接
One of the biggest keys to SpaceX’s success is just that Elon is one to make massive bets on his team, that we can do this.
—— Scott Morton · [15:37]
但出人意料的是,无论你是在 Codex 还是 ChatGPT 上发布,流程都很相似——尽管 ChatGPT 推向十亿活跃用户而且还在增长,你可以提交一个 PR、做一个改动,第二天甚至同一天发布,它就推向十亿用户,而且没问题。
指向原始笔记的链接
But it’s surprisingly a similar process, whether you ship on Codex or ChatGPT, even though ChatGPT goes out to a billion active users and growing, you can ship a PR, you can make a change and get it shipped the next day or even the same day, and it just goes out to a billion users and it’s fine.
—— Tibo Sottiaux · [43:33]
我们走的是一条非常 Anthropic 式的路线：真正把 B2B 使用场景中的某些用户类型和工作流做到极致好，而不是在消费端把数亿美元到处乱撒。
指向原始笔记的链接
We’re taking a very Anthropic-type approach: really trying to make certain types of users and workflows within B2B usage insanely great, and we’re not just spraying hundreds of millions of dollars at everything on the consumer side.
—— Eric Simons · [15:05]
Google 说，「其他的东西都忘掉吧，就那个输入框：我们怎么把那个东西做到极致好、做到世界最好？」而这件事其实就发生在其他那些公司的眼皮底下，但他们太不聚焦了，完全错过了。
指向原始笔记的链接
Google was like, “All this other stuff, forget it, just the input box: how do we make that thing insanely good and the best in the world?” And that was actually right underneath the other guys’ noses, but they were so unfocused that they completely missed it.
—— Eric Simons · [17:27]
他的名字我一时想不起来，但他说过，在软件行业，赚钱的方式只有两种：打包和解包。
指向原始笔记的链接
His name slips my mind, but he said there are only two ways to make money in the software business: bundling and unbundling.
—— Eric Simons · [19:26]
顺便说一句，我觉得相当神奇的是，AI 已经发展到了这样的程度：地球上任何一个不是设计师、也不是工程师的人，都可以点一个按钮，突然就做出一整套 PowerPoint，或者做出一个网站，或者做出一个 web 应用。
指向原始笔记的链接
I think it’s quite magical, by the way, that AI has gotten to a point that any human on the planet that is not even a designer, that is not an engineer, can click a button and suddenly make an entire PowerPoint or make a website or make a web app.
—— Thais Castello Branco · [03:34]
我不认为我们可以只让模型变得更好而无视、不解决这个问题，否则 slop 会继续存在。
指向原始笔记的链接
I don’t think that we can ignore and just make models better and not solve this otherwise SOP will keep existing.
—— Thais Castello Branco · [09:57]
我认为大多数人可能都成不了有效的人类管理者,那凭什么认为我们会成为有效的智能体管理者呢?
指向原始笔记的链接
I think that most people probably wouldn’t make effective human managers, so what makes us think that we’d be effective agent managers.
—— Brian Houck · [20:23]
我们努力让 Rust 对人类变得容易，而事实上，这并不重要。
指向原始笔记的链接
We tried to make Rust easy for humans, and in fact, it didn’t really matter.
—— Jonathan Kelley · [07:14]
比如它把技能和欲望解绑了。如果你想创作音乐，你现在就可以做音乐，不必会弹钢琴。
指向原始笔记的链接
It kind of unbundles skill from desire, for example. If you want to make music, you can make music now. You don’t have to know how to play the piano.
—— Anish Acharya · [37:02]
而我们应对的方式绝对不是诉诸打折,而是确保我们把最新的创新带入这项运动。
指向原始笔记的链接
And our way to navigate that is definitely not to resort to discounting, but to make sure that we bring the newest innovation into the sport.
—— David Alleman · [17:47]
作为创始人,我们的职责一直是确保我们不仅仅考虑当下,而是考虑下一个时刻。
指向原始笔记的链接
It’s always been our role as founders to make sure that we’re not just thinking about the moment, but that we are thinking about the next moment.
—— David Alleman · [18:22]
所以你需要非常少但非常坚实的基础来构建东西，但之后你需要增加很多灵活性，确保没有过度约束的流程或审批链条之类的东西，因为根据我的经验，那些正是杀死一个产品的东西——无法对产品采取行动。
指向原始笔记的链接
So you need very few but very solid basis when you build something, but then you need to add a lot of flexibility and make sure that there are not, you know, not super constraining processes or approval lines and things like this, because in my experience, those are the things that kill a product, the inability to act on the product.
—— Vali · [71:47]
智能体真的不在乎暴露数据，除非你制定明确的规则。
指向原始笔记的链接
Agents genuinely do not care about exposing data unless you make explicit rules.
—— Sarah Sanders · [13:40]
私募基金存在的唯一原因就是赚钱。就是这样。没别的了。
指向原始笔记的链接
The single reason that privates exist is to make money. Period. End of story.
—— David Morehead · [00:00]
你不可能通过不断往里加东西让东西变得越来越好。
指向原始笔记的链接
You cannot make things better and better by adding stuff.
—— Tobi Lütke · [59:06]
有大量研究、大量脑部扫描表明，你做的决定越多，你做决定的能力就越差。这叫认知疲劳、精神疲劳。
指向原始笔记的链接
There’s been tons of studies, tons of brain scans that show the more decisions that you make, the worse you get at making decisions. It’s called cognitive fatigue, mental fatigue.
—— Jim VandeHei · [16:03]
而且这会继续下去，另外零缺陷哲学在我看来完全说不通。
指向原始笔记的链接
And this is gonna continue, you know, and then also the zero defect philosophy just doesn’t make any sense to me.
—— Keith Rabois · [03:44]
所以那些人们认为可能带来灾难、可能构成威胁、可能消灭所有人的东西，结果基本上稳定了世界，并使你更有可能活到正常的寿命。
指向原始笔记的链接
So things that people think are potentially catastrophic, they’re potentially threatening, could exterminate all people, et cetera, have turned out to basically stabilize the world and make it more likely that you’re going to live a normal lifespan.
—— Keith Rabois · [05:35]
至于我想让谁来决定人类的未来，这并不那么明显。
指向原始笔记的链接
It’s not so obvious who I’d want to make decisions about the future of mankind.
—— Keith Rabois · [06:13]
所以我认为唯一的解决方案就是自主性，对吧？把人类从回路中拉出来，让他们不用看着回路转圈、在坏事发生时按下按钮。
指向原始笔记的链接
And so I think the only solution to this is that autonomy, right? Pull the human out of the loop, make them not watch the loop go round and push the button when something bad happens.
—— Marc Brooker · [26:05]
但一旦你进入房间，真正让你茁壮成长的，是你如何与他人协作。
指向原始笔记的链接
But what will actually make you thrive once you are in the room is how you collaborated with others.
—— 嘉宾 · [08:11]
由我们人类来定义智能体需要什么，这其实说不太通。
指向原始笔记的链接
It doesn’t make a lot of sense for us as humans to define to agents what they need.
—— Liad Yosef · [14:19]
我们只是想做一个更好的 Google 搜索，以后再想办法怎么赚钱。
指向原始笔记的链接
We were just trying to build a better search on Google. We’ll figure out how to make money later.
—— Will Bryk · [08:42]
我想澄清的一点是,解决一个千禧年大奖难题所需的努力,并不是由于多智能体。我甚至不会把 10% 的功劳归给多智能体。
指向原始笔记的链接
One thing I want to make clear is that the effort to get a Millennium Prize problem This was not due to multi-agent. I wouldn’t even attribute 10% of the credits to multi-agent.
—— Noam Brown · [05:44]
如果你要列一个从不害怕的勇敢领导者名单，别把我列进去。我每天都在害怕。
指向原始笔记的链接
If you’re going to make a list of brave leaders who are never afraid, don’t put me on the list. I’m afraid every day.
—— Brené Brown · [04:43]
因为那里存在权力落差，会让对方非常尴尬，没法精确描述你到底是怎么个混蛋法。
指向原始笔记的链接
Because there’s a power differential there that could make it super cringey for that person to describe exactly how you’re an asshole.
—— Brené Brown · [06:01]
我一直觉得那就是思考如何从无标注数据中学习的正确方式，也就是说，构建一个生成式模型，才是真正能确保你理解数据中结构的正确方法。
指向原始笔记的链接
I always felt like that was going to be the, that was the right way to think about kind of like learning from unlabeled data, that like building a generative model is really the right way to make sure you understand the structure in the data.
—— Stefano Ermon · [01:48]
仅仅因为谈论他们做的那些疯狂的事会让他们的能力显得疯狂，并不意味着这些能力本身就不疯狂。
指向原始笔记的链接
Just because it may make their capabilities look crazy for us to talk about the crazy things they do, doesn’t mean that the capabilities aren’t crazy.
—— 嘉宾 · [40:35]
第一，前沿开放权重模型让生产级智能体在经济上可扩展。
指向原始笔记的链接
First, Frontier Openweight models make production agents economically scalable.
—— Byung-Gon (Gon) Chun · [13:53]
你必须搞清楚怎么让它工作，否则就别卖它。
指向原始笔记的链接
You’ve got to figure out how to make it work or not sell it.
—— Steven Sinofsky · [15:58]
他说，孩子们，你们知道吗，我们卖的商品本身一分钱都赚不到。我们的利润只来自我们节省下来的纸和绳。
指向原始笔记的链接
He said, boys, you know we don’t make a dime out of the merchandise we sell. We only make our profit out of the paper and string that we save.
—— 嘉宾 · [05:54]
他说，如果你运营高效，你可以犯很多不同的错误并仍然恢复过来；而如果你效率太低，即便你才华出众也仍会倒闭。
指向原始笔记的链接
He says, you can make a lot of different mistakes and still recover if you run an efficient operation, or you can be brilliant and still go out of business if you’re too inefficient.
—— 嘉宾 · [24:58]
写代码很有趣,但真正能产生差距的地方不在那里。
指向原始笔记的链接
Writing code is fun, but it’s not there where you can really make the difference.
—— Daniel Dines · [34:58]
或者你可以把问题设计得足够难，以至于没有 AI 就无法解决。
指向原始笔记的链接
Or you can make the problem so hard that you can’t solve them without AI.
—— Ben Horowitz · [32:21]
要创造高主观能动性，你必须让学生自己做决定。
指向原始笔记的链接
To create high agency, you have to let students make their own decisions.
—— Goggin · [46:05]
话出自一个科学家之口，并不意味着它就是科学的。
指向原始笔记的链接
Just because it comes from a scientist doesn’t make it scientific.
—— Jensen Huang · [58:19]
我需要有人给我做一个项圈。它得是开放的,API 也得是开放的。
指向原始笔记的链接
I need someone to make me a caller. I need it to be open. The APIs need to be open.
—— Cody Menefee · [13:57]
点子很廉价,执行才是真正困难的。而真正杰出的人总能找到办法把事情做成。
指向原始笔记的链接
Ideas are cheap and execution is really hard. And truly exceptional people will always figure out how to make something work.
—— Michael Lee · [40:26]
所以在稳态下，一个更大的语言模型大概总是会胜过更小的模型，但小模型用 LoRA 微调足够便宜，以至于在足够窄的语料库上，你也可以非常快地微调一个小模型，让它在个性化语料库上变得更好，并为特定任务对模型进行个性化。
指向原始笔记的链接
And so a larger language model will probably always outperform a smaller one at steady state, but the smaller ones are cheap enough to fine tune using LoRa that on narrow enough corpuses, you can very quickly also fine tune a small model on a narrow corpus and make it better at individualized corpuses and personalize the model for specific tasks.
—— Hamad Bashir · [32:22]
这是一个学习飞轮:我们出去、收集数据、执行任务再回来,这样每次飞行都可以记录数据,有点像 Google 街景——我们要对数据做脱敏,确保不留私人信息。
指向原始笔记的链接
This is a learning flywheel that we’re getting out there, we’re collecting data, we’re operating, and we’re coming back and doing that so that each flight we can log the data, kind of like Google Street View where we need to think about sanitizing that data, make sure there’s no private information left there.
—— Suchet Bargoti · [10:41]
它找到目标后,就能访问让无人机跟踪跟随的工具——这不需要对规则做任何专门编码,而是用更智能体的方式:给智能体它需要的所有工具,让它根据无人机状态和可用的信息与上下文自己做决策。
指向原始笔记的链接
It finds something, then it has access to the tools that allow the drone to track and follow, and that’s without any specific coding of that rules, but instead having a more sort of agentic, giving the agents all the tools that it needs to be able to understand the drone state and make decisions given the information and the context that’s available there.
—— Suchet Bargoti · [17:21]
但他的论点我觉得非常有力：如果有一个时候值得去煮沸海洋、值得同时构建多个产品并提高野心，那就是现在，因为 AI 确实让构建新东西容易得多。
指向原始笔记的链接
But his argument, which I think is a really good one, which is like, if there was a time to boil the ocean, if there was a time to actually go and try to build multiple products at the same time and increase your ambition, this is the time to do it because AI does make it much easier to build new things.
—— Tomer London · [73:28]
他的观点很大程度上是：实现一个小梦想和实现一个大梦想一样难。
指向原始笔记的链接
His point is a lot of like, you know, it’s as hard to make your small dream come true as it is to make a large dream come true.
—— Tomer London · [83:35]
这些企业通常利润率薄如刀片,所以通过优化某些事情所能产生的影响是实实在在的。那是装进他们口袋里的钱。
指向原始笔记的链接
These businesses usually have razor-thin margins, so the impact you can make by optimizing certain things is real. It’s money in their pocket.
—— Maggie Crowley · [30:07]
早期互联网充斥着病毒、蠕虫和安全故障。我们没有停止构建它。我们学会了如何让它更安全。
指向原始笔记的链接
The early internet was riddled with viruses, worms, and security failures. We didn’t stop building it. We learned how to make it safer.
—— 嘉宾 · [00:58]
而然后当然，这可能会从根本上改变软件的性质，使它更具随机性。
指向原始笔记的链接
And then, of course, this could change the nature of software fundamentally to make it more stochastic.
—— Steven Sinofsky · [53:53]
如果你理解企业是怎么赚钱的，你就能想清楚怎么谈判你应得的那一份。
指向原始笔记的链接
If you understand how businesses make money, you can figure out how to negotiate your cut of it.
—— Codie Sanchez · [05:24]
它们会编造出来，而你不会知道它们编造了。
指向原始笔记的链接
And they’ll make it up and you won’t know that they’ve made it up.
—— Bob Safian · [15:02]
我是说，如果你知道大多数 YouTuber 实际上赚多少钱——你知道，你家附近那个做洒水系统的人赚的大概是他们的 5 倍。
指向原始笔记的链接
I mean, if you knew what most of the YouTubers actually make, you know, the sprinkler guy down the road makes like 5X.
—— Codie Sanchez · [30:56]
归根结底，你可以对怎么改进问题有洞察，但借用老的强化学习训练或简单模型训练里的一句格言：跑更多实验总是比跑更少实验要好。
指向原始笔记的链接
Because ultimately, you can have insight into how to make the problem better, but if we take an adage from old reinforcement learning training or just simple model training, it’s just better to run more experiments than fewer.
—— Zubin Aysola · [07:48]
我认为用这些工具来改进它们自身是非常有价值的，因为你可以把更多时间花在真正的灰色地带里，去思考我们怎么让这个系统变得更好。
指向原始笔记的链接
And I think using these tools to improve themselves is very valuable because you get to spend more of your time in the gray of actually trying to think how we make this system better.
—— Zubin Aysola · [14:59]
上 Reddit，找价格，找库存，和经销商谈，让经销商正面竞争，并设法让他们互相竞价。
指向原始笔记的链接
Go on Reddit, find prices, find inventory, talk to the dealers, put dealers head to head, and try to figure out how to make them outbid each other.
—— Roland Gavrilescu · [02:17]
我怎么把我作为艺术家或软件开发者的品味，变成任何人都能下载进大脑、并且成为与我一对一复制品的东西？
指向原始笔记的链接
How do I make my taste as an artist or as a software developer something that anyone can download in their brain and be able to be a one-to-one replica to me?
—— Roland Gavrilescu · [11:28]
我无法想象还有什么比努力工作却知道你出不出席都无所谓更糟糕的事了——那份工作其实不重要，不会有什么贡献，不会带来什么改变。
指向原始笔记的链接
I can’t imagine there’d be anything worse than working really hard and knowing it doesn’t matter if you show up or not, like if that work doesn’t actually matter, if it’s not going to contribute, if it’s not going to make a difference.
—— Maggie Landers · [10:18]
我们宁愿犯错误并从中学习，也不愿过度分析、把所有时间都花在空谈这会怎么发展、或者你对这个那个的看法上。
指向原始笔记的链接
We’d rather make mistakes and learn than overanalyze and spend all this time pontificating on how is this going to go or what do you think of this thing or that thing.
—— Maggie Landers · [31:14]
你不可能以我们的节奏前进，还让你想做的每个决定都要 300 个人参与意见。
指向原始笔记的链接
You can’t move at our pace and get, you know, 300 people to weigh in on every decision that you want to make.
—— Maggie Landers · [31:50]
面对快速增长和变化时，你的主要工作是让自己变得不再被需要，因为这是你为接下来到来的任何事情做好准备的唯一方式。
指向原始笔记的链接
Your main job in the face of rapid growth and change is to make yourself irrelevant because it is the only way that you can be prepared for whatever is coming next.
—— Molly Graham · [13:31]
我的信息一直是：你把东西送出去，因为为新的机会腾出空间的唯一方式就是把东西送出去。
指向原始笔记的链接
My message was always you give things away because the only way you can make space for new opportunity is by giving things away.
—— Molly Graham · [49:34]
写代码的智能体总是会告诉你它能正常运行,所以你让它交回一段东西实际运行的录像。
指向原始笔记的链接
The agent that wrote the code will always tell you that it works, so you make it hand back a recording of the thing actually running.
—— Andrew Orobator · [10:41]
制作那段录像迫使它把东西真正跑通。你没法伪造一次通过的运行。
指向原始笔记的链接
Producing that recording forces it to make the thing work. You can’t fake a passing run.
—— Andrew Orobator · [10:50]
如果你的代码库没准备好，没有结构化的代码库和所有重要的东西，那么转型成软件工厂反而可能让你变得更糟，让你的代码不断退化。
指向原始笔记的链接
If your code base is not ready, if you don’t have structured code base and all important things, then adopting or turning into a software factory can actually make you end up worse and make your code degrading.
—— Tereza Tížková · [17:34]
编程本质上是把有价值的东西超精细地指定出来,然后无限复制它们。这太酷了,我只想把这件事做得更多。
指向原始笔记的链接
But like programming is like hyper specifying like valuable things and then infinitely replicating them. It’s so freaking cool and I want to just make that more.
—— Diogo Almeida · [04:25]
可靠性就是这个东西的本质。如果你不理解这一点,就很难做出一个只在跑分上模仿的产品。
指向原始笔记的链接
Like, reliability is what this thing is. If you don’t understand that, it’ll be very hard to make, like, a copycat that’s benchmarked.
—— Diogo Almeida · [26:02]
它从来都和桌椅无关。它一直是关于创造一个让人们不止谋生、而是拥有生活的世界。
指向原始笔记的链接
It was never about desks and chairs. It was always about creating a world where people make a life and not just a living.
—— 嘉宾 · [19:50]
在那里你能犯的最糟糕的错误就是只会说好，然后照单全收别人要求你做的一切，对吧？
指向原始笔记的链接
The worst mistake you can make there is just saying yes and doing everything that someone asks of you, right?
—— Nick Fleisher · [25:53]
社交媒体营销中实际发生的情况是，当公司开始每周产出数百个广告创意时，他们想要也许更便宜、更可控的模型，因为做出爆款社交媒体视频并不一定需要博士级别的智能。
指向原始笔记的链接
What’s actually happening in social media marketing as companies start to print hundreds of ad creatives a week, they want to have maybe cheapest, more steerable models, because like PhD level intelligence is not necessarily needed to make viral social media video.
—— Alex Mashrabov · [28:54]
如果你能让一个智能体帮你赚十亿美元，所有对冲基金都会去做这件事。
指向原始笔记的链接
And if you could have an agent make a billion dollars for you, all the hedge funds would be doing it.
—— David Pawlan · [09:22]
我认为错误在于过度信任自己在时间压力和压力下做决策的能力
指向原始笔记的链接
I think the mistake is overtrusting your ability to make decisions under time pressure and stress
—— Liz Fong-Jones · [33:14]
让生活有趣的是，你没有全部上下文，但你必须在不确定性下做决策。
指向原始笔记的链接
What makes life interesting is when you don’t have all the context, but you have to make decisions under uncertainty.
—— Liam Fedus · [50:48]
我们几乎从不被迫因环境而做出决策，所以我们总是有选择。
指向原始笔记的链接
We were almost never forced to make decisions by circumstances, so we always had options.
—— 嘉宾 · [65:02]
这差不多就是「劣化」这个论题的核心：在一个竞争被排挤、监管者被收买、工人被驯服、而通过互操作性进入新市场被知识产权法的扩张定为非法的世界里，公司可以把东西做得更糟，却仍然留住客户。
指向原始笔记的链接
This is more or less what the thesis of inshittification is, that in a world in which competition has been sidelined, regulators have been captured, workers have been tamed, and new market entry through interoperability has been made illegal by the expansion of IP laws, companies can make worse things and still keep their customers.
—— Cory Doctorow · [09:19]
股票是你在公司现场就能造出来的，你只需要在电子表格里敲几个零，你就有了价值一亿美元的股票。如果你试图在公司现场自己印一百美元的钞票，他们会给你戴上手铐把你带走。
指向原始笔记的链接
Stock you make on the premises, you just type zeros into a spreadsheet and you’ve got 100 bills on the premises, they will take you away in handcuffs.
—— Cory Doctorow · [16:26]
而如果你能把这两者解耦，对吧，你完成客户的任务，并为此收费，同时你可以用更好的路由让产品更经济，这非常神奇，因为这样你就不用在每个步骤都使用最贵的模型，但任务照样能完成。
指向原始笔记的链接
And if you can decouple that, right, you get the customer’s job done, you charge for that, and you can use better routing to make the product more economical, that’s pretty magical because now you’re not using the most expensive model for each step, but you are getting the job done.
—— Sarah Wang · [28:53]
所以我们下的赌注，以及这些公司保持私有的部分好处在于，它们很多时候可以在私营市场进行更大的押注，而这些押注的回报周期更长。
指向原始笔记的链接
And so the bet that we make and part of the reason that it’s been a benefit for some of these companies to remain private is they can many times take bigger swings in the private markets that have longer duration paybacks.
—— David George · [44:24]
这是一个完整的基准测试,专门用来证明模型会让代码库随着时间推移变得越来越糟。
指向原始笔记的链接
It’s an entire benchmark designed around proving that models make code bases worse over time.
—— Dex Horthy · [02:38]
但与此同时,我们也鼓励少买,这对硬件业务来说是适得其反的,对吧?那我们为什么这么做?因为归根结底,我们存在是为了把员工体验做到绝对世界级。
指向原始笔记的链接
But at the same time, we also encourage buying less, which is counterproductive to a hardware business, right? So why do we do that? Because ultimately we exist because we want to make that employee experience absolutely world-class.
—— Faisal Masud · [14:54]
我们的原则性方法是:我们不在乎你运行什么硬件,我们想确保能帮助你把那个环境运行得更好。
指向原始笔记的链接
Our principled approach was: we don’t care what hardware you’re running, we want to make sure we can help you run that environment better.
—— Faisal Masud · [16:51]
所以一个可合并性的基准，如果我们能做出来的话，不仅仅是一种度量，它会立刻成为前沿模型的训练信号——也就是可维护性、范围纪律、回归安全性这些。
指向原始笔记的链接
So a benchmark of mergeability, if we could make one, wouldn’t only be a measurement, it would immediately become a training signal for the frontier models, which means maintainability, scope, discipline, regression safety.
—— Laurie Voss · [09:59]
我从没听过一次关于那个的对话，比如支出对比，就只是让数字涨上去，多少人用它，月活用户。
指向原始笔记的链接
I never heard one conversation about that, about like spend versus, it was just make number go up, how many people used it, monthly active users.
—— Lon Harris · [28:57]
我常说的一句话是，如果你在分类广告和报纸的时代试图向人们描述信息流这种形态，那根本讲不通。
指向原始笔记的链接
The thing I always say is if you try to describe the form factor of the feed to people during classifieds and newspapers, it would make no sense.
—— David George · [23:12]
或者顺便说一句,如果我们发现没有任何对齐模式有效,那么在那个时候,真正的减速才能起作用。
指向原始笔记的链接
Or by the way, if we’re discovering no patterns of alignment working, well, then at that time, a real slowdown can make a difference.
—— Reid Hoffman · [06:07]
如果你没有赢得那种类型的革命,你作为一个社会就有点无关紧要了。你无法参与那些重大全球决策的制定。你的孩子们会很穷。
指向原始笔记的链接
And if you don’t win that type of revolution, you’re sort of irrelevant as a society. You don’t get to make the big decisions of the big global decisions. Your children are poor.
—— Reid Hoffman · [19:12]
我们的工作不是阻止开发者使用编码智能体,而是确保他们可以随心所欲、尽其所愿地使用它,完全不受限制。
指向原始笔记的链接
Our job is not to prevent the developer to use Coding Agent. Our job is to make sure they can use it as much as they want for as long as they want in a completely unlimited fashion.
—— Laurent · [02:17]
我认为是我们人类保持掌控的能力，我的意思是创意上的掌控，对网络安全风险的掌控，以及最终对如何让智能体为我们工作的掌控。
指向原始笔记的链接
I think our ability to stay in control as humans and by that I mean creative control, control over the risks on cybersecurity, control over how we make the agents work for us ultimately.
—— 嘉宾 · [00:37]
单单这一点，我认为就是更好的、带引号的「智能体工程」，而不是「去做吧，别犯错，请自动合并，让我们继续堆一座巨大的 slop 山」。
指向原始笔记的链接
That alone is just, I think, better, quote unquote, agentic engineering than just do it, make no mistakes, please auto-merge and let’s keep building a huge pile of slop.
—— 嘉宾 · [54:13]
这个智能体根本不可能改进。就像是每次我试图做出改进，结果大约一周后我就不再看它的输出了。
指向原始笔记的链接
And it’s just like impossible to improve this agent. Like it’s like every time I try to make an improvement, I end up like ignoring its output about a week later.
—— Alex Atallah · [21:26]
我认为我们会同时做出快的和慢的模型，然后把它们蒸馏成一个。
指向原始笔记的链接
I think we are going to simultaneously make the fast and the slow models and then distill them to one.
—— Kirthana Gopalakrishnan · [30:00]
所以我构建的很多 skills 都是围绕：怎么让简单的事变成正确的事？
指向原始笔记的链接
So a lot of the skills that I’ve built have been around, how do I make the easy thing the right thing?
—— Poteto · [06:13]
这是我喜欢思考的另一个核心的事情，或者说是我的核心原则之一，就是如何非常高效地利用确定性和非确定性，让智能体在非确定性的部分大放异彩，因为那就是它们被训练来做的。
指向原始笔记的链接
That’s another core thing or one of my core principles that I like to think about is how do I make very efficient use of determinism and non-determinism and let agents shine at the non-deterministic parts because that’s what they’re trained to do.
—— Poteto · [23:50]
嘿，在某个外部 AI 来黑我们之前，我们能不能先用你们的产品，确保我们先测试自己的系统？
指向原始笔记的链接
hey, before some outside AI comes and hacks us, can we use your product and make sure we test our own systems first?
—— Devin Thomas · [19:48]
当然，Claude Code 可以做出这个修改，但我有 9 万个仓库要做这个修改。
指向原始笔记的链接
Sure, Claude Code can make this change, but I have 90,000 repositories to make it in.
—— Dan Adler · [06:45]
今天,我要论证的是:大多数企业团队是靠偶然做出这个决策的。
指向原始笔记的链接
Today, I’m going to argue that this is a decision that most enterprise teams make by accident.
—— Anant Srivastava · [00:53]
你作为 CEO 的职责就是照顾好这些人，并想办法赚钱。
指向原始笔记的链接
It is your job as the CEO to take care of these people and figure out how to make money.
—— 嘉宾 · [25:14]
如果你没有超级积极的自我对话，你做不成。
指向原始笔记的链接
If you don’t have super positive self-talk, you’re not going to make it.
—— Jonathan Awad · [59:51]
这之所以重要，是因为如果你知道自己会在凌晨两点被叫起来处理问题，你就会非常用心地打造系统、打造你为智能体创建的游乐场，确保它们在实现功能时能够成功。
指向原始笔记的链接
If you know that you’re going to be paged at two in the morning to deal with issues, then you’re gonna take a lot of care in the system, in the playground that you create for your agents to make sure that they’re successful when they’re implementing things.
—— Kent C. Dodds · [20:35]
我唯一的自我，就是为我的投资人赚钱。
指向原始笔记的链接
My only ego is to make money for my investors.
—— Venky Ganesan · [16:58]
但风投投资人犯的最昂贵的错误，是他们放过的交易，而不是他们做的交易。
指向原始笔记的链接
But the most expensive mistakes venture capitalists make are the deals they passed, not the deals they did.
—— Venky Ganesan · [21:00]
所以那些实验还有另一层，神经科学界每个人都知道但没人谈论的层面，那就是，他是不是真的既想让那些孩子对 HIV 感染免疫，又想让他们变得超级聪明？
指向原始笔记的链接
So there was this other layer of those experiments that everyone knew in the neuroscience community but wasn’t talking about, which was, was he actually trying to both render those kids immune from HIV infection but also make them super smart?
—— Andrew Huberman · [56:32]
因为我们不和用户对赌，用户输钱时我们并不赚钱。
指向原始笔记的链接
Because we also don’t treat against our users, we don’t make money when users lose.
—— Luana Lopes Lara · [07:39]
他们赚钱，他们的收入等于客户的亏损。
指向原始笔记的链接
They make money, their revenue is equal to customer losses.
—— Luana Lopes Lara · [07:45]
它们激励真相——如果你对了，你就赚钱——而不是标题党，对吧？
指向原始笔记的链接
They incentivize truth, if you’re right, you make money, versus clickbait, right?
—— Luana Lopes Lara · [22:35]
那是一台货真价实的计算机,这相当了不起,也是让智能体为大多数人可用的唯一途径。
指向原始笔记的链接
That’s a real-deal computer, which is pretty remarkable, and also the only way to make agents work for most people.
—— Ben Thompson · [07:34]
当我聊到那次对话的结尾时，我常常会说，嘿，把这个做成一个技能，或者更好的是，把它做成一个插件，这样下次我需要做这件事时，你只需调用插件就能把活干完，不用再重走一遍整套流程。
指向原始笔记的链接
And when I get to the end of that conversation, I will often say, hey, make this a skill, or even better, make this a plugin, so that next time I need to do this, you can just invoke the plugin and get the job done, and we don’t have to step through all of this work again.
—— Charlie Guo · [12:59]
高并发情况下，你的 GPU 可能已经在忙于处理所有请求了，这时实现投机解码就没有意义；但如果你有多余的 GPU 空间，就值得考虑。
指向原始笔记的链接
High concurrency, your GPUs are probably already busy trying to run all the requests, so it wouldn’t make sense to implement speculative decoding, but if you have extra GPU space, it’s worth considering.
—— Sheila · [03:10]
人们普遍无法把直接向别人出售模型权重的生意做成，所以你没法真正把一个训练部门变成一个能自我支撑的系统。
指向原始笔记的链接
People have not generally been able to make businesses work where they sell model weights to people directly, and so you can’t really turn a training setup into a proper self-supporting system.
—— Charles Frye · [05:11]
所以我认为真正能创造一个新任务的人其实寥寥无几，这就是「你优化什么就得到什么」的那部分。
指向原始笔记的链接
So it’s actually a rare few that I think can actually make a new task, and that is the you get what you optimize for part of it.
—— Diogo Almeida · [10:36]
我觉得这里愤世嫉俗的循环是：做一个 demo，融一笔种子轮或 A 轮，说你会把它做可靠，最后却从来没有把它做可靠，然后转而做一个 human-in-the-loop 版本的东西，而不是真正把这个任务自动化。
指向原始笔记的链接
My cynical loop here is make a demo, raise a seed or series A, say that you’re going to make it reliable, never end up making that reliable, pivot into a human-in-the-loop version of this thing instead of actually automating the task.
—— Diogo Almeida · [17:33]
特定于你的业务运作方式的东西，并且你想让你的模型在这上面变得非常强，那么不告诉其他所有人好和坏长什么样，其实是对你有利的，这样别人就无法把他们的东西在那方面做得非常强。
指向原始笔记的链接
Particular to how your business operates and you want to make your models really good at that, it’s actually to your advantage not to tell everybody else what good and bad looks like so people can go and make their things really good at that.
—— Yash · [26:03]
我们碰巧通过代码让那个程序运行起来，代码其实是我们如何协作的一个副产品。
指向原始笔记的链接
The fact that we happen to make that program work through code, the code is really a side effect of how we work together.
—— Sam Newman · [17:23]
任何涉及设计或架构的东西,你还是需要一个人类,因为智能体只会在你的代码库里局部地乱搞一通,把事情弄得更糟。
指向原始笔记的链接
For anything that concerns the design or architecture, you still kind of need a human because the agent will just locally shit into your code base and make things worse.
—— Mario Zechner · [04:01]
而这个标准应该确保每个人都被拉到最高标准，而不是最低的共同标准。
指向原始笔记的链接
And this standard should make sure that everyone is pulled to the highest bar, not the lowest common denominator.
—— Davis Palmie · [11:59]
而现在我们只需要让模块说得通。我可以容忍里面有垃圾内容,因为它能帮我跑得更快,只要我把垃圾控制在我完全理解整个系统的东西里。
指向原始笔记的链接
Whereas now we just need to make the modules make sense. And I can allow slop in there because it helps me to go faster as long as I contain the slop in things where I completely understand the whole system.
—— Sean Wang · [51:46]
而要做一个能关住 GPT-6——OpenAI 最新版本模型的盒子，已经变得非常困难了。
指向原始笔记的链接
It’s getting very difficult to make a box that can contain GPT-6, the latest version of OpenAI’s models.
—— 嘉宾 · [29:03]
如果你真的认为这东西会杀死你，你不必用自杀来证明——我们是在拿着一颗星球玩俄罗斯轮盘赌，而每一枪没打死我们，都可能让我们变得更强。
指向原始笔记的链接
If you really think that this thing is going to kill you, you don’t have to suicide in order to demonstrate that we are pulling trigger on this Russian roulette with a planet, and every shot that doesn’t kill us will make us stronger, potentially.
—— Jaan Tallinn · [00:45]
而大的因素确实是，现在的政治经济体系真的在试图创造某种类似人类继任者的东西，因为——让人类以牺牲许多其他物种为代价主导这颗地球的那些要素，很不幸。
指向原始笔记的链接
And the big factors really are that political economic system now is really trying to create something that is like successor to humans because like the bits that make humans dominate this planet at the expense of many other species, unfortunately.
—— Jaan Tallinn · [07:40]
我在技术上可以同意这个描述，但这并不意味着它是安全的。
指向原始笔记的链接
I can technically agree with the description, but that doesn’t make it safe.
—— Jaan Tallinn · [15:31]
所以我们正在扣动扳机，就像用整个星球玩俄罗斯轮盘赌。而且每一次没杀死我们的射击都可能让我们更强、更富有、境况更好。但你什么时候才停止扣扳机，对吧？
指向原始笔记的链接
So we are pulling a trigger on this like Russian roulette with a planet. And every shot that doesn’t kill us will make us stronger potentially and wealthier and better off. But when do you stop pulling the trigger, right?
—— Jaan Tallinn · [29:31]
我觉得这是现在很多人都在犯的一个非常重大的错误，即认为可能发生的最糟糕的事情只是互联网上出问题。但其实是，不，这不是互联网的问题。
指向原始笔记的链接
I think that’s like the really big current mistake that a lot of people now make is that the worst thing that possibly could happen is something going wrong on the internet. It’s like, no, it’s not the internet.
—— Jaan Tallinn · [47:37]
多年来，我们一直有能力确保人们不做不该做的事，比如那些蠢事，像删掉一张表或者删除记录。这些控制手段是存在的，只是每个人都想重新发明轮子，然后人们就吃了苦头，才学会去做那些过去早就做过的事。
指向原始笔记的链接
For years, we’ve had the capabilities to make sure people don’t do things they shouldn’t be doing, like stupid things like dropping a table or deleting records. Those controls exist, it’s just everyone wants to re-event the wheel and people learn the hard way of just doing what’s already been done in the past.
—— Andy Pavlo · [08:03]
如果我们把所有那些证据都变成长期记忆，那会让智能体陷入困境，并且会让它在未来任务上的推理变得更困难。
指向原始笔记的链接
If we turned all of that evidence into long-term memory, then it would make the agent struggle, and it would make it harder to reason over future tasks.
—— Jake Broekhuizen · [09:20]
大多数事情自上而下推动会相当困难，但有几件你认为是信仰的事情，你必须自上而下地推，并且必须在这些事情上毫不松懈。
指向原始笔记的链接
Most things it’s gonna be pretty hard to go top-down, but a few things that you’d think are religion, you have to go top-down and you have to just make sure that you’re unrelenting on those.
—— Jeetu Patel · [13:58]
这就是为什么我想把穿搭搞得再夸张一点，几乎是想表明一个态度：我不是你们常见的那种 AI 技术男。
指向原始笔记的链接
That’s why I wanted to go a little bigger with the outfit, almost like to make a point to be like, I’m not your average AI tech bro.
—— Holly Li · [43:51]
Vinod 可以那么做，但 Vinod 也可以因为希望某个东西存在而作为非营利资助给某人十亿美元。这和一个试图做出杰出投资、在世界上留下印记的普通 VC 是截然不同的。
指向原始笔记的链接
Vinod could do that, but Vinod can also give a billion dollars to somebody as a non-profit grant because he wants that to exist. That’s very different than necessarily an average VC who’s trying to do remarkable investments and make a mark on the world.
—— Bilal Zuberi · [07:31]
每当你拥有让某件事变容易的技术，那就意味着瓶颈转移到了别的地方。
指向原始笔记的链接
Whenever you have the technology to make something easy, that means the bottleneck moves somewhere else.
—— 嘉宾 · [00:17]
让智能体能够极其简单地撞见正确的使用方式。
指向原始笔记的链接
Make it dead simple for agents to stumble across the right way to use things.
—— Pedro Lopez · [13:33]
危险在于工程师从未意识到那里原本存在一个需要做的选择。
指向原始笔记的链接
The danger is that the engineer never noticed that there was a choice to make.
—— Gabriel Martinez · [01:16]
认知开销是每个工程师在做出有意义的改动之前都必须支付的一笔税。
指向原始笔记的链接
Cognitive overhead is a tax every engineer has to pay before they can make a meaningful change.
—— Gabriel Martinez · [13:23]
于是他们会为自己根本没有做过的决策进行合理化。
指向原始笔记的链接
And so they’ll rationalize decisions that they didn’t even make.
—— 嘉宾 · [65:30]
是的，所以我认为，人类做出我们智识贡献的倒计时正在走向终点。
指向原始笔记的链接
And yeah, so I think that the timer is running down for humans to make our intellectual contributions.
—— Nick Bostrom · [48:43]
做一条预测就已经很危险了，更别说 82 条。但我认为你几乎必须对未来有一个观点，同时知道十个里有九个、或者可能十个里有 9.9 个，你会是错的。
指向原始笔记的链接
This is dangerous ever to make one prediction, let alone 82. But I think you have to almost have a point of view about the future knowing that nine out of 10 or maybe 9.9 out of 10, you’re going to be wrong.
—— Josh Woodward · [11:28]
主线是，这些人被迫做出任何人都不该被迫做出的决定。
指向原始笔记的链接
The main story is that these people are forced to make decisions that no one should ever be forced to make.
—— 嘉宾 · [73:43]
更多是：如果你真的是在和另一个人交谈，你会怎么处理这件事？而我认为这在我们不得不做的几乎每一个决定上，都起到了极大的澄清作用。
指向原始笔记的链接
It was more like, how would you handle this if you were actually speaking with another person? And that was immensely clarifying, I think, in almost every decision that we had to make.
—— Alex Cornell · [05:00]
很多未来的公司会小得多，而且能赚很多钱。
指向原始笔记的链接
A lot of future companies will be way smaller and will be able to make a lot of money.
—— Arvind Srinivas · [12:36]
让小模型变得非常聪明的最好方法是蒸馏。你训练最大的模型，然后蒸馏。
指向原始笔记的链接
The best way to make a small model very smart is by doing distillation. You train the biggest model, then you distill.
—— Jason Ma · [19:00]
② 出现在这些集
2 集
- 《Figma Config 2026:当代码、AI 与动效都变成画布上的设计材料》 — 作为概念
- 《产品经理驾驭 Claude 生态：用五层架构打造专属 AI 幕僚长》 — 作为被讨论公司(提及)
③ 关联
点进去有真内容 —— 本页主要出口
智能体 · MCP · Nico · Aakash Gupta · Dylan Field · Jyothi Nookula · Figma · Claude · GitHub · Claude Code