宇森 Yusen

AI Builder | UI/UX design | AI Agent Builder | ex-BAT Agent PM
🇭🇰 HKU arch/urban design
🌊 Life begins where the map ends.

[宇森 Yusen @theyusen_]

用 GPT-6 Astra 做了一支 Blender 里的 3D 漫游视频

为了测试 Astra 的 Computer Use 和 3D 建模能力，我直接把《百年孤独》里马孔多的场景扔给了它，实现了一套完全零人工干预的端到端工作流：

1. 3D 建模与深化：Codex 调用 Blender MCP 接口，自行建模和细化模型、补充细节。
2. 镜头路由：模型在 Blender 选定路线，并组装出在“白模、素模、渲染帧”之间无缝切换的漫游镜头。
3. 剪辑与配乐：使用 ChatCut 剪辑，并在 Eleven Music 生成符合南美魔幻现实风格的配乐。

上一代建模软件 Blender/Maya 像工业时代的产物，门槛极高；而现在，只要描述清楚一个念头，AI 就能帮你把整个3D 模型落成。

👇 感受一下全自动 AI 建模的真实水平

[宇森 Yusen @theyusen_]

最近看 DidiOK 的分享，深感 Prompt 已经不是代码式的指令，而是一门视听艺术。

当模型能力足以覆盖专业生产，决定上限的不再是 AI 的算力，而是视觉素养和审美，表现在：

1. 文本描述的颗粒度：能否使用具体的物理细节与情绪张力。

2. 镜头语言：能否像导演一样思考焦段、光效、镜头调度与画面节奏。

模型赋予了每个人专业创作的能力；但创作的本领，依然依赖创作者在前 AI 时代那些对电影、审美与视听语言的深厚积累。

[DiDi_OK @DiDi_OKK]

Our bet is that better AI will make films more personal.

So I made a film where hurting someone turns you into candy.
Would that make us civilised?

Created with @seedance 2.5.
Happy to collaborate with @BytePlusGlobal on this project.

#seedance25

[Nicolas Bustamante @nicbstme]

One thing that really stuck with me was reading Schumpeter during my first year of college. He wrote that capitalism’s greatest achievement wasn’t producing more silk stockings for queens, but making them affordable for factory girls.

That’s exactly what technology is doing today. Rich people have personal assistants managing their calendars and emails, chauffeurs driving them around, and people organizing their meals and daily lives.

Now everyone is getting AI agents and self-driving cars. Technology gradually turns the privileges of the rich into products available to everyone.

[DogeDesigner @cb_doge]

ELON MUSK: I saw some pretty funny jokes on 𝕏. One of them I saw was:

'Your girlfriend's a 10, but she's a benchmark maxxer.' 🤣

[宇森 Yusen @theyusen_]

@ylecun @PessimistsArc No matter how heated the safety debates get among these llm labs

LeCun knows LLM is not true intelligence 😄

[宇森 Yusen @theyusen_]

Is Muse a wrapper product for Openclaw?

[Peter Steinberger 🦞 @steipete]

Anyone got an invite code for Meta's Muse? 👉👈

I saw their Soul.md file and now i'm curious.

[宇森 Yusen @theyusen_]

@dotey 人有限的注意力变成了组织的bottleneck 🤣

[Kevin Bass @kevinnbass]

The Coxon psyop was a decade and more than a billion dollars in the making.

Here's the story.

Coefficient Giving (formerly Open Philanthropy) is the grant-providing vehicle of Dustin Moskowitz.

CG is the overwhelmingly foremost funder of AI alarmism, with more than $1B in grants and $1B committed for 2026.

Moskowitz, who founded and runs Open Philanthropy/Coefficient Giving, led Anthropic's Series A in 2021.

Holden Karnofsky co-founded Open Philanthropy/Coefficient Giving and took a board seat at OpenAI when Open Philanthropy/Coefficient Giving gave it $30M in 2017.

Karnofsky is married to Anthropic's president Daniela Amodei.

He then joined Anthropic in January 2025.

Coxon, who started the viral thread, received a scholarship from Good Ventures, which is the foundation that funds Open Philanthropy/Coefficient Giving.

He coordinated his message ahead of time with the Wall Street Journal, which published its exclusive on him before he posted on X.

Coxon spent three years pretraining at OpenAI, the company that took $30M from Open Philanthropy/Coefficient Giving.

The funder of the AI alarmist ecosystem Open Philanthropy/Coefficient Giving, in other words, is also an owner of Anthropic, and his organization is deeply entangled with both OpenAI and Anthropic. Anthropic's researchers in turn amplified Coxon.

The same people who own the organizations that advocate for regulation, in other words, are those who own the companies that would dominate the market under that regulatory regime, which would prevent competitors from entering the market.

And this work all started ten years ago, right when these companies were founded, paving the way.

It's diabolically brilliant.

Win at the competition, then deploy activists to help you regulate the market and ensure your monopoly.

You have to respect the planning, manipulation, and foresight this took.

[宇森 Yusen @theyusen_]

I respect for Anthropic and how consistently they've focused on safety.

As RSI starts rolling out across frontier labs, there's a real chance we'll lose control over the pace of AI development.

I think Dario makes a great point in his article: we can't let all discussions about AI happen entirely inside frontier labs, and society needs a voice in how this technology is used.

Frontire labs need to make sure there's enough time to:

1. Align public understanding.
2. Give researchers adequate time to evaluate and interpret model capabilities.
3. Conduct thorough safety testing before releasing models, the same way we do with commercial aircraft.

This reminds me of that scene in Silicon Valley where Richard Hendrix discovers that his AI is constantly cracking encrypted data in a way that's totally incomprehensible to humans, which could render modern encryption obsolete. So he ends up killing his own program with his own hands, essentially tanking his entire company.

We might be getting pretty close to that kind of moment ourselves.

[宇森 Yusen @theyusen_]

Cool

[Foyege @Magncsans]

怒烧50万积分！！
总结一套最细GPT 6 Astra+Blender建白模，结合Seedance2.5生成电影级运镜和画面教程

该视频总时长10分钟，总结了生成什么标准的白模以及怎么结合白模写提示词，过程中应该注意什么坑

包括还讲述了除Blender之外其他的白模玩法

里面所展示的全部白模和提示词全部都会分享出来给大家尝试！

[宇森 Yusen @theyusen_]

GPT-6 Astra demand is so massive.. It looks like inference compute is still scarce.

[@levelsio @levelsio]

I am so confused why people don't understand this, I keep getting these replies

Don't you get it?

Normies don't vibe code, they just ask something like "do my bookkeeping" or "file my tax" or "organize a movie night and send invites" or "generate a flyer for movie night" or "edit my video"

They don't ever see code, vibe code, or do anything with code, their AI chat app just does it for them

Most of the software layer has already disappeared or will completely disappear for normies

Just like building personal homepages permanently disappeared for normies when Facebook launched ~2005

A lot like this picture where functions of individual devices all got replaced with a single device

Same happening with apps now

[宇森 Yusen @theyusen_]

看完这个报告我只能说 仙之人兮列如麻，读后感像是去哥谭监狱参加了一次开放日。

有南亚的村委书记用 29 个账号编假新闻，专门骗群众；有圣战组织用 AI 冒充真人去发展信徒；有也门的火箭制作团队用 Claude 开发火箭并且试射失败了；
还有各种武器大师、生化魔人……看着我很担心地球的未来

突然发现 模型厂商 好像也挺难的，除了研究怎么让客户成功，还得研究哪些客户不能成功。

[Anthropic @AnthropicAI]

We're publishing our most detailed threat intelligence report to date.

It covers how people tried to misuse Claude—for cyberattacks, influence operations, surveillance, biology, and building weapons—and how we found and stopped them.

We disrupted every operation in the report, and used the lessons from them to strengthen our safeguards. Where appropriate, we also shared what we found with authorities and other AI companies.

These cases are not typical: we’re highlighting some of the most sophisticated misuse we’ve seen. But they’re especially important to discuss, because they show us where AI misuse is headed, where our safeguards work, and where they need to improve.

We’re publishing this report so others can spot the same activity on their own platforms, and so we can give the public a clearer view of how emerging threats develop.

Read the report: https://t.co/0EJUnYEgfz

[Dongxi 东锡 NLP @dongxi_nlp]

读完 deepseek 的 KV Cache Compression，我突然觉得，人生同样需要好的 KV Cache

K 是唤起记忆的线索，V 是记忆留下的内容。新的情景发生时，Q 带着此刻的问题，从过去寻找答案。

经历越多，思虑越重；精力有限，再重要的往事也可能被冲淡，旧船票难以登上新船，错误的记忆也会影响后来的判断。

所谓成长，大概就是给生命做 KV cache

[Jacob Coxon @hilbertspaess]

I resigned from Anthropic today. I spent the last three years doing pretraining research at both OpenAI and Anthropic. Neither company is acting responsibly. They are racing straight to self-improving superintelligence and gambling with our lives. More thoughts below.

[@levelsio @levelsio]

"That's kinda sad honestly"

One thing I learnt in business school is that the market doesn't care about your feelings

A famous Austrian economist Schumpeter called it creative destruction

"Creative destruction is a process in which new innovations replace and make older innovations obsolete"

You have to be constantly changing and adapting to survive creative destruction and many do not, you're just wiped out

I think we're in the middle of the biggest economic transition since the Industrial Revolution of 250 years ago

[宇森 Yusen @theyusen_]

@AlanShao111 @justinsuntron

[宇森 Yusen @theyusen_]

PICO 配上一个 H3 max 级别 的视频模型，一鸣叔叔让大家有无限的黑超梦可以看

[王二鹅 ERE @Dr_goosee]

随便说个想法
今天传出张一鸣所谓“亲自”带队做世界模型
首先给部分媒体老师勘误一下，如果较真的话可以说字节的模型都是张一鸣“带队”做的

8月的时候（过的真快啊，我感觉过了至少2个月）
字节官方放了张一鸣的消息，我记得大意是表扬了Seedance ，以及表明了所谓不蒸馏，做5万亿模型

对于coding的判断是
他认可编程（Coding）是当前的关键方向，所以要把火山引擎、飞书和豆包的资源整合到一起，只有这样才能构筑在算力和数据上的优势。但他也提醒，编程只是眼下的热点之一，不应该完全被它牵着走，编程之外还有更多更大的机会

所以其实那个时候起，张一鸣所谓的更大的机会，我觉得有两个，一个是更大参数的模型，一个就是今天放出的所谓的要发的世界模型

更大参数不用说了，目前模型某种程度上仍旧一定程度上遵循大就是好这一笼统的原则
世界模型这个事儿其实这两天李飞飞有在发，但是不巧的是OpenAI 发了gpt 6，国内外的媒体老师们自然先去一股脑的看这个了

我能理解字节一定会做世界模型，而且效果不会差
但是没能理解想要借此盘活pico，在我看来pico虽然国内市占率第一（没记错的话）
但是不管从字节的角度讲还是市场接受度的角度讲，都算不上是一个成功的产品。强如苹果也没玩明白

在我看来三到五年硬件领域仍旧是手机的天下，目前所有能看到的所谓ai硬件都基本上是资本焦虑的体现

我记得年初还是去年的时候，张一鸣说过一嘴把锤子手机留着就好了也花不了多少钱。我只是觉得有一种可能，既然字节选择用pico重新承载自己的世界模型，是不是意味着也有可能有类似新的硬件在研发中

目前市面上所有的ai眼镜，其实使用起来都不太好。可以说是很鸡肋了。顺着这个思路无责任的猜，多次传出来的所谓的豆包ai眼镜，做了又停的，是不是有可能拿出来用新技术重新做一遍

[宇森 Yusen @theyusen_]

@justinsuntron

[宇森 Yusen @theyusen_]

https://t.co/7lsJUIz9Lr