Mostly Physics , AI, Macro, Crypto, Finance , Technology and Civilizational Shifts, Philosophy,Music n Aroma. 目前在OKX-Research, SOSV-VC mentor,Fudan&UC Berkeley

Joined March 2026
17 Photos and videos
Pinned Tweet
What truly matters has always been the ecosystem's integrative capacity—transforming boundaries from constraints and limitations into a freely flowing, normally distributed universe of infinite possibilities..
1
2
197
Here is the polished English translation: Elephants Can Dance: The Olympic Ecosystem; Only upon arriving here did I realize that the Olympics have long been a mature, juggernaut IP ecosystem. The IOC doesn't merely manage the Olympic Games as a mega-IP; it also oversees the development of 206 National Olympic Committees worldwide, sporting events, media and digital rights, commercial sponsorships, fan communities, technical standards, athlete development systems, sports culture dissemination and PR, sports science, derivative products, charitable initiatives, international political discourse, and support for sports in developing nations. Its commercialization pathways and corporate development frameworks are remarkably diverse—both internally and externally—making its organizational structure far more complex than that of an ordinary enterprise. Q: If the Olympics were to build a sports-vertical AI platform or internal AI-driven efficiency tools (applying harness engineering to regional event management, fan operations, energy consumption, sustainability, athletic training, athlete tracking, etc.), they possess an inherent first-mover advantage: exclusive, vast historical databases of venues, athletes, and more. Yet the sheer scale of the organization will inevitably present execution challenges in driving AI adoption…
1
13
Jill yang retweeted
埃隆·马斯克谈及成功者最重要的品质: “好奇心。尽量多阅读。尽可能在许多不同领域学习。并且用批判性思维考虑任何你得知的东西。”
38
411
1,740
89,090
从商业火箭可回收实现开始到星链starlink开始稳定盈利再到xAI 建设算力集群,未来实现在太空上建设算力中心发电中心 ,真的每一步都不是吹牛逼 ,elon把每一步都去落地实现 ,这人的大脑已经性感到无以复加, 未来已来。没啥好说的 ,我买就是了 。 这笔打算存着 再等上一个五年再看看。
1
24
So... it was still drizzling when I left , Then, while riding along, I accidentally took the wrong path and— just by chance i recognized that i ended up right at the entrance of the world's first stock exchange at Amsterdam! Right as I arrived, the sky cleared up sky cleared up 🐂—time to soak in some good luck vibes!! Go SpaceX !!!🐂🐂 #amsterdam #stockexchange #spacex #TravelDiaries
34
you should come here now
9
Jill yang retweeted
We're building a trading bot with a personality. t lives on Obsidian. Every day, it logs its own thoughts — market reads, trade rationale, second-guesses — in a shared vault where humans and the agent write side by side. You leave a note. It responds. It remembers. The execution layer runs on Onchain OS. The bot finds quotes, routes trades across 500 DEXs, and settles on X Layer with zero gas. Fully autonomous. Fully onchain. But the interesting part isn't the trading. It's the journal. Most algo strategies are black boxes. This one keeps a diary. Over time, you get a record of how an AI agent thinks through markets — in its own voice, on its own timeline, next to yours. We're calling it a strategy. It might end up being something closer to a character. More soon. DM if you have better idea or same experience
1
2
87
Jill yang retweeted
We're entering the golden age of just-in-time software
64
54
621
114,432
Jill yang retweeted
今天东京首次 rewired meetup 很人气✨
3
1
29
8,866
Trading with agent is cool i guess
2
27
Jill yang retweeted
The window is closing. You probably don't feel it yet. Everyone who entered crypto in 2023 learned to think in models. GPT-4, Claude, Gemini — three years of Capex flowing overwhelmingly toward foundation models. That's where the bets went. That's where the mental real estate went. 2025 was supposed to be the year of Agents. And it was — technically. But the mindset didn't move. The conferences still talk models. The capital allocation still skews models. Agents got the headlines; they didn't get the conviction. That gap is the opportunity. And it's not permanent. Historically, paradigm shifts in tech give you roughly 3 years of open window before consensus catches up and compresses the upside. We're 18 months in. The remaining runway — call it 12 to 18 months — is when the asymmetry is still real, before everyone who's been sleeping on Agents wakes up at the same time. The infrastructure is ready. The use cases are live. The only thing still lagging is where people's heads are. That's a timing arbitrage. And timing arbitrages don't wait.
1
1
68
Jill yang retweeted
Exchange OS isn't just another DeFi protocol. It's a structural shift in how markets get built. For years, launching an exchange meant rebuilding the same matching engine, margin system, and settlement stack from scratch — or trusting a centralized operator to do it for you. That tradeoff killed most ideas before they started. Exchange OS moves those functions to the protocol layer. Any institution, any developer, any team with a market worth creating can now deploy on the same infrastructure stack that runs one of the world's largest exchanges — with self-custodial asset security, configurable compliance, and a unified account system across every venue. The real signal here isn't the technology. It's the model: OKX is no longer just operating a market. It's becoming the infrastructure layer that other markets run on. That's a different kind of moat.
1
1
30
Jill yang retweeted
What truly matters has always been the ecosystem's integrative capacity—transforming boundaries from constraints and limitations into a freely flowing, normally distributed universe of infinite possibilities..
1
2
197
before productization you should go “ skillify it ” fisrt
“Skillify it” is the way you should write most tasks Write markdown that makes code Don’t make elaborate Foxconn factories to call agents Let agents make their own tools, kaizen style
1
26
Recent days in Tokyo , let's connect If you're also researching in the AI/RL space, let's connect and trade notes! DMs are open. 📥 #AIRL #FDE #SoloBuilder
1
49
Jill yang retweeted
历史反复证明:技术领先者不一定是价值捕获者。别看谁的模型最强,要看谁控制了价值流动的关键节点——芯片产能、能源、数据飞轮、标准制定权。印刷术让书便宜了,稀缺变成识字和判断力。互联网让信息免费了,稀缺变成注意力和信任。算力 Token 充裕了,稀缺会变成'值得被算力服务的目标'。这是哲学问题,不是技术问题。
1
1
85