Brad

founder @dokki_one ｜ building for AI-natives, shipping with my agent crew 24/7

[Brad @brad_bao]

Most multi-agent setups I see aren’t actually better. They’re just more complicated ways to fail.

People keep adding agents before asking the only question that matters: does this work even split cleanly?

If it doesn’t, you’re not scaling intelligence. You’re scaling coordination debt.

[Brad @brad_bao]

@DegenInvestor69 been there
spent more time trying to make them talk to each other than the actual task took
at some point you realize the hard part isn’t the agents, it’s the glue

[Brad @brad_bao]

@tenobrus this is where I’m at too
once they start pinging each other and forming some kind of hierarchy, you stop feeling like you’re “using tools”
and start feeling like you’re managing a weird little remote team that never logs off

[Brad @brad_bao]

@rcmisk yeah this is too real
at some point you’re not prompting anymore
you’re just project-managing a bunch of agents that refuse to talk to each other properly

[Brad @brad_bao]

@pbteja1998 man I’ve felt this
agents do the work but I’m still the one who has to keep track of what each of them was doing
shared memory would’ve saved me a few late nights already

[Brad @brad_bao]

The biggest change after running a few agents isn’t productivity.
It’s that I now worry about them at night.
They don’t stop, so it’s hard for me to stop either.

[Brad @brad_bao]

@rileybrown Real.
I start worrying at night about what the agents are doing and whether they messed something up.
They don’t stop, so I don’t either. Easy to end up staying up all night.

[Brad @brad_bao]

真正有用的 Agent，不只是会调用工具，而是能持续建立世界模型、理解任务边界、在反馈中修正行为。

[Brad @brad_bao]

“智能”不是一个中心化的“超级控制器”，而是无数模型在并行���测、校正、协作后涌现出的世界理解能力。

[Brad @brad_bao]

@keyahayek lmao

[Brad @brad_bao]

Most AI transformations are wrong.

They’re trying to add AI to companies, instead of rebuilding companies around AI.

[Brad @brad_bao]

AI-native organizations are making companies obsolete.

[Brad @brad_bao]

@NousResearch @OpenRouter Congrats! Amazing team

[Brad @brad_bao]

Agents are eating the world.

[Nous Research @NousResearch]

Hermes Agent is now #1 on the Global @OpenRouter token rankings.

While our journey together has just begun, we'd like to take this opportunity to thank our contributors, supporters, and users for all they have done to get us this far.

[Brad @brad_bao]

The “no real-world signal” argument became outdated the moment copilots and chat agents shipped.

Every bug report is an eval.
Every retry is RL.
Every weird user workflow is grounding data.

AI labs aren’t isolated from the economy anymore — they’re already plugged into it.

[Tom Reed @mentalgeorge]

I don't think automation of AI R&D will rapidly lead to domain-general super-intelligence.

I think this will be true even if AIs can do *literally everything* a human AI researcher does today.

Even after the full automation of AI R&D, further capabilities progress will only happen through

(1) widespread deployment of AI throughout the economy, accompanied by data collection; and/or
(2) the wholesale recreation of much of the economy by AI labs.

Without access to the real-world signal provided by either of the above, I think that the only thing produced by automated AI researchers would be a "Goodhart Singularity".

If I'm right, this is obviously good news. I make the case for this in a new piece on my substack

[Brad @brad_bao]

很多公司以为，
给员工接入 ChatGPT、买几个 Copilot license、做几个 AI demo，就算完成 AI 转型了。

但真正的 AI Transformation，往往不是“多了一个工具”。

而是：

组织结构会变，
workflow 会变，
决策方式会变，
人才结构会变，
甚至��品和公司的边界都会变。

很多旧流程，本质上是为了“人类能力有限”而设计的。
但当 AI 开始具备 reasoning、execution、coordination 能力之后，很多系统其实需要被重新设计一遍。

最大的风险，可能不是公司不用 AI。
而是用了 AI，却还在用旧时代的组织方式思考问题。

#AI #AITransformation #AGI #FutureOfWork #AIAgents

[Brad @brad_bao]

如果 AI 在数学上的进展继续保持现在这个速度，数学教育体系可能很快会迎来一次真正的重构。

很多人还默认：
AI 只是帮助学生更快做题。

但更深的问题是——
当 AI 已经开始参与推理、证明，甚至提出新的数学路径之后，我们到底还应该怎样学习数学？

未来的重点，也许不再只是“解题能力”。
而是提出问题、构建抽象、选择方向，以及和 AI 一起探索未知的能力。

这可能不只是工具升级。
而是数学教育范式本身的变化。

#AI #Mathematics #Education #AGI #Math #FutureOfEducation

[Brad @brad_bao]

很高兴这次能支持 AI × Web3 School 的共学营和 Hackathon。

过去一年越来越明显地感受到，AI agents 正在从“会聊天”走向真正的 autonomous execution——开始调用工具、管理资产、参与链上交互。

而当 agent 真正开始行动之后，authorization、risk control、安全边界这些问题，也会变得越来越重要。

很喜欢 Hackathon 的地方在于，很多���正有价值的想法，都不是在讨论里产生的，而是在 builder 一起协作、快速试错、不断迭代里长出来的。

感谢 @aiweb3school 以及所有生态伙伴、社区、高校链协一起推动这件事。

很期待看到下一波 AI × Web3 builders 🚀

#AI #Web3 #AIAgents #AgenticAI #CryptoAI #Hackathon #Builder

[AIxWeb3 School @aiweb3school]

这次共学营和 Hackathon 的推进，离不开很多生态伙伴、技术团队、媒体社区和高校链协的支持～

💗特别感谢以下伙伴对 AI × Web3 School 的支持：

@ethereumfndn Builder Growth 团队、@Cobo_Global、@web3privacy、@zai_org、@elytro_eth、@PhalaNetwork

🤝 感谢所有媒体合作伙伴、社区合作伙伴，以及链协伙伴的支持。

🎤 媒体合作伙伴
@PANews、@Foresight_News、@wublockchain12、@TechFlowPost、@ethdaily

🫂社区合作伙伴
@eth_vietnam、@ethereum_sg、@NapulETH、@apacdao、@basezh、@OpenCsg、@499_DAO、@021Labxyz、@OpenBuildxyz

🏫链协支持：
@UIBEFTA、@USYBA_DAO、@ECUPL_BA、@cityuhkweb3ai、@BadgerBlock、@SEU_BA、@cnsndecyun1、@ZuelBlockchain、@0xUClub、@THUBA_DAO、@opencas_ucas、@BUPT3DAO、@nchublockchain、@JRBC330000、@FudanBlockchain、@sgunidao、@nchublockchain、@xuedao_tw

[Nous Research @NousResearch]

Hermes Agent v0.13.0 - “The Tenacity Release”

[Brad @brad_bao]

未来三年，CTO这个职位将不复存在。