早晨的思考
今天早上我写下了这段文字，虽然一开始不确定是否会发布，但最终决定分享出来。
I wrote this early this morning and I wasn't sure if I would actually publish it, but here it is:
https://t.co/7Dw9UFpeep
AI 领袖动态
2026-04-11 · 31 条
...
https://t.co/UWxeZ1n4hr
https://t.co/UWxeZ1n4hr
最后的主题演讲
我为今天的最后一场主题演讲做了准备，欢迎收看。
gave myself the last keynote of the day.
tune in https://t.co/OljQpdhQHh
对中国AI工作文化的观察
1. 很多人11点到达工作，晚上11点才下班。2. 由于这种作息，很多员工都很年轻，父母难以维持同样的作息。3. 这些公司的每个人都在使用最好的美国AI。
Observations about Chinese AI work culture:
1. Many arrive at work late (11 am) and work until late at night (11 pm).
2. Due to the schedule above, many employees are young. Hard for parents to sustain the same schedule.
3. Everyone at these companies is using the best US AI
为孩子们培养AGI-proof职业
我正在为我的孩子们培养一种能够抵御AGI影响的职业技能。
Training my kids for an AGI proof career https://t.co/eQ6Wesw2w0
Codex团队的机会
Codex团队和@OpenAI现在有一个巨大的机会：1. 告诉@openclaw用户如何切换到GPT订阅（我认为只需告诉机器人切换模型即可）。2. 修复GPT的个性（也许分享一个提示能在短期内有所帮助？）。
Codex team and @OpenAI have a huge opportunity right now to:
1. Tell @openclaw users how to switch to gpt subscription (I think it’s just telling the bot to switch the model?)
2. Fix GPT’s personality (maybe even sharing a prompt will help in the short term?). This is the main
Replit与Databricks的整合
Replit现在可以直接部署到Databricks。你的应用程序在Databricks环境中运行，同时继承其安全性、治理和数据访问。测试版已上线，企业已经开始使用，并在商业智能和内部工具方面取得了巨大的加速。
Replit now deploys directly to Databricks.
Your apps run inside your Databricks environment while inheriting its security, governance, and data access.
Beta is live. Enterprises are already building with it and seeing massive acceleration in BI and internal tools. https://t.co/O33uJHohgo
企业CIO与AI领导者的见解
我越是接触非科技领域的企业CIO和AI领导者，就越明显，如果你开发的软件没有良好的无头模式，未来几年将面临风险。
The more I meet enterprise CIOs and AI leaders outside of tech, the more it’s obvious that if you’re building software that doesn’t have a great headless mode, you’re going to be at risk in the coming years.
Asked a group of 20 IT leaders across banking, media, finance, and
软件和自动化的需求被低估
我认为大家严重低估了在那些看起来不像“软件”的领域对软件和自动化的总需求。我不是在谈论手机上的另一个应用，而是那些全天候为公司自动化事务的软件。
I think everyone is substantially underestimating the total demand for software and automation in areas that don’t feel like “software”. Not talking about software that’s another app on your phone.
Software that just automates things for companies all day long. Most companies
提示的重要性
认为提示毫无用处就像认为给刚加入团队的新同事清晰的指示是无用的。“提示”应该涵盖给予代理完成任务所需的所有信息。这是高杠杆的做法。
The idea that prompting would be useless is like if giving clear instructions to a brand new colleague who just joined your team is useless. “Prompting” should just encapsulate the entirely of giving the agent everything it needs to perform the task. This is high leverage.
AI 采用的两种模式
AI 的应用呈现出两种截然不同的模式。一方面，大多数用户通过聊天工具与 AI 互动；另一方面，一些人则在部署代理来执行长期任务，创造实际工作成果或自动化工作流程。前者非常有用，但...
AI adoption is a tale of two cities. On one end (most) users right now are interacting with AI via chat tools and on the other end people are deploying agents to do long running tasks that create and produce real work output or automate workflows.
The former is super useful but
我一直在告诉你们
Look I've been telling you too
Look I've been telling you too
与 Kulveer 在 YC 合作的自豪感
我非常自豪能与 Kulveer 一起在 YC 工作。
Super proud to be working with Kulveer at YC
即将集成 GBrain
我们将在 3... 2... 1... 的倒计时中集成 GBrain。
Integrating into GBrain in 3... 2... 1...
对未来的警惕
我们将努力反对这种结果，但在进入新的不确定时代时，这个想法令人信服且令人恐惧。
We're going to fight against this outcome, but it's a compelling and scary idea to be hyper-aware of as we go into new and uncertain times
必读书籍
这将是一本必读的书！
This is going to be a must-read!
大模型的好消息
大型模型的良好表现意味着小型模型的表现也会更好。这使得智能对每个人都变得可及，这是我希望继续发展的一个超级趋势。
The good news about bigger models that work well is that it also means smaller models will work even better too
And that is how intelligence becomes accessible to everyone and this is one super trend I hope continues
奥克兰的枪支暴力问题
奥克兰不到 0.5% 的人口造成了大部分枪支暴力。奥克兰的 Ceasefire 项目在针对正确人群时将枪支犯罪减少了一半。这不是枪支问题，而是政治意愿的问题。
Less than 0.5% of Oakland's population drives most of its gun violence. Oakland Ceasefire cut gun crimes in half when the city targeted the right people. This isn't a gun problem. It's a political will problem.
https://t.co/Gx60J6BiPI
自身经历
我自己也遇到过这个问题。下一阶段的代理将专注于简易入门和开箱即用的安全性。
I’ve definitely hit this myself
The next phase for this level of agents will be about easy onboarding and security out of the box
AI 需要开放市场
AI 需要开放市场和开放访问，而不是目前发生的事情。
AI needs open markets and open access not whatever is happening here
GBrain助力完美记忆
如果你希望你的OpenClaw或Hermes Agent能够完美记忆所有10,000多个Markdown文件，GBrain可以提供帮助。这是我OpenClaw/Hermes Agent的设置，采用MIT许可证的开源软件。希望能帮助你构建迷你AGI。
If you want your OpenClaw or Hermes Agent to be able to have perfect total recall of all 10,000+ markdown files, GBrain is here to help.
It's exactly my OpenClaw/Hermes Agent setup. MIT-licensed open source. Hope it helps you build your mini-AGI.
https://t.co/yFpFU4pn5b
GBrain与Hermes Agent兼容
GBrain可以在Hermes Agent上运行，只需粘贴相同的安装脚本即可开始使用。
GBrain works on Hermes Agent, paste the same install script and you're good to go.
https://t.co/yFpFU4pn5b
从VC转型的11条经验
我曾写过一篇关于我两年前转型为风险投资人的长文，但决定将其缩减为11个要点（虽然有些是陈词滥调，但确实如此）：1. 上飞机——不要质疑。2. 交易往往在...
I was writing a (relatively) long essay on all my learnings of transitioning to a VC ~two years ago..
But, decided to whittle it down to a tight 11 bullet point list (some are cliches, but oh so true):
1. Get on the flight -don’t question it.
2. Deals often die on the
VC提问的目的
有时风险投资人会提问只是为了测试你的深度（市场、技术、你的执着）。在我看来，最优秀的创始人能够以极其清晰的方式论证论点的两面。风险投资人提出的问题很少会让你真正思考你的业务。
Sometimes VCs will ask questions just to test your depth (market, technical, your obsession).
The best founders (imo) are able to argue both sides of an argument with extreme clarity.
There’s rarely something VCs ask that should make you *really* think about your business.
寻求WhatsApp CLI的帮助
有人想帮助WhatsApp CLI吗？它需要关注，而我现在无法专注于此。
Anyone here who wants to help with WhatsApp CLI? It needs love, and I can't focus on it right now. https://t.co/J28AfwqmeZ
首个使用Claude管理代理的应用
首个使用Claude管理代理的应用程序来了！
first @every app using Claude Managed agents!!
@TrySpiral https://t.co/LpLNPQkyev
我笑死了
我笑死了，哈哈。
im dead lol
AI的下一个大趋势
AI的下一个重大领域无疑是生物学和医学。如果你想知道原因，这里有我三年前为@every写的一篇文章。
the next big thing in AI is undoubtedly biology and medicine.
if you want to know why, here’s a piece I wrote for @every 3 years ago: https://t.co/kHiw3WpOMd
Shopify实现代理原生化
Shopify实现了代理原生化！
Shopify goes agent native!
OpenAI的澄清
我刚与OpenAI交谈，这实际上是错误的。OpenAI正在与一个受信任的测试小组合作开发一个网络产品，但这与他们最新的模型Spud无关。不幸的是，Axios的报道混淆了这两者，现在已经更新。
I just spoke to OpenAI, and this is actually false.
OpenAI is working on a cyber product with a trusted tester group.
But this is NOT related to Spud, their newest model.
Unfortunately seems like the Axios story conflated the two, and has now been updated.
自由软件的影响令人困惑
自由软件的影响确实让人难以理解。如果你不喜欢某个界面，可以立即设计并推出一个新的。如果性能较慢，只需重新构建数据层并自动优化性能。这真是疯狂。
The implications of "free" software are truly hard to wrap my head around.
You don't like an interface? Just design and push a new one immediately.
Performance is slow? Just re-factor the data layer with auto-performance optimization.
Insane.