GeekPark 极客公园

GeekPark here. For 16 years, we’ve been there — early. Close to the moments before they became obvious. At GeekPark: - Lei Jun explained why to build Xiaomi — with zero users - Zhang Yiming sat on the floor at a basement meetup, sharing early ideas on machine learning - Elon Musk presented his EV vision — his first public appearance in China We’ve also been early believers: - Angel investor in Unitree Robotics - Angel round investor in XPeng Today, this has grown into: - A community of 40,000+ AI builders - Our 5th fund in progress - 10+ editions of the GeekPark Innovation Festival Before the next inflection point — let’s connect.

Perhaps the most misunderstood thing about the AI era is this: people assume the scarce resource will remain generation. But generation is quickly becoming ambient. Words arrive instantly now. Images emerge from a sentence. Music, code, films, voices, simulations — all increasingly obtainable at the speed of intention. We are entering a world where creation itself begins to feel less like craftsmanship, and more like invocation. The old friction between imagination and execution is dissolving. But something else becomes visible once that friction disappears. The real mystery was never how content was made. It was why humans felt compelled to make anything at all. Why someone stays awake trying to describe a feeling precisely. Why a founder spends years pursuing an idea that statistically should not work. Why certain melodies make people cry. Why some people need to leave traces of themselves in language, companies, paintings, architectures, children, civilizations. Models can produce infinite variations of form. But they do not ache toward anything. They do not long. They do not grieve. They do not feel humiliation, envy, tenderness, transcendence, loneliness, religious awe, or the strange human instinct to search for meaning even inside suffering. All artifacts still originate somewhere prior to generation: inside desire. And perhaps that is what becomes newly valuable in the AI era. Not the ability to produce - But the ability to genuinely want.

@UnitreeRobotics x.com/geekparkhq/sta…

Unitree Unveils: GD01, A Manned Transformable Mecha, from $650,000 👏 The world's first production-ready manned mecha. It can transform. It's a civilian vehicle. It weighs ~500kg with you inside. Please everyone be sure to use the robot in a Friendly and Safe manner.

I have a feeling some of you might prefer this heavier version of the @UnitreeRobotics GD01 😏

@KaichaoYou @vllm_project

Open source brings us together🔥 All 5 model drops Matt mentions — Kimi K2.6, Qwen 3.6, MiMo V2.5, Ling-2.6, DeepSeek V4 — shipped with day-0 vLLM support @vllm_project . Not a coincidence. That's open-source model vendors and global maintainers building in the open, together. 🌏

@goodhunt Welcome to China!

Join me & the whale bros across China!

The most interesting part about the rumored @Kling_ai spinoff is the valuation math behind it. When a fast-growing AI business sits inside a slower-growing parent company, the market usually prices the AI business using the parent company’s overall multiple. That means the AI upside gets diluted by the broader company’s slower growth and mature business expectations. Kling is a good example. Reports suggest Kling generated around $75M revenue in Q1 2026, with most revenue already coming from overseas markets like North America. The valuation is around ~$20B. But Kuaishou itself trades closer to ~1.5x PS. So even if Kling reaches a $1.3B annualized revenue run rate next year, using Kuaishou’s current multiple would still imply a valuation much closer to ~$2B than $20B. Feels like this is becoming a common pattern now. Google separated Waymo. Baidu spun out Kunlunxin. Now Kuaishou may do something similar with Kling. The faster the new business grows, the harder it becomes to fit inside the parent company’s existing valuation framework. That said, looking at past capital market cases, Kuaishou probably won’t do a fully detached spinoff. There have been enough controversial examples where companies stripped out their highest-growth assets during periods of weak parent-company valuations, leaving public shareholders with the slower-growing core business. A more likely structure is probably independent financing while Kuaishou retains controlling ownership and continues consolidating Kling financially. That gives Kling more flexibility around fundraising, incentives, and AI talent retention, while still allowing Kuaishou shareholders to participate in the upside. In practice, this may increasingly become the template: traditional internet businesses provide stable cash flow, while AI units operate with more independent capital structures and growth expectations.

@WillowVoiceAI Willow

Introducing Willow. The world's most accurate voice dictation tool. Willow learns from every correction you make. Automatically. Auto-learning plus context-awareness unlocks 99% word accuracy. No other dictation app comes close. RT + comment "Willow" to get a month free.

Wispr AI is reportedly in funding talks at a $2B valuation. Genuinely curious what the long-term capital stack for products like this looks like? bloomberg.com/news/articles/…

@elonmusk Welcome to Beijing again! Not sure if you still remember this but you visited the GeekPark office back in May 2014 haha Good memory for us! x.com/GeekParkHQ/sta…

On my way to Beijing in Air Force One

@luoyonghao 测试成功✌️

测试一下。

@lovart_ai Didn’t mean to accidentally promote you guys 🤣🙏

In case anyone wonders how Chinese tech media team decompress at work. - Location: GeekPark office in Beijing. - Today’s theme: practicing traditional Chinese archery (中国弓箭🏹)

Hosted alongside friends from @rtedevcommunity, Google Cloud × Webeye, and Parthenon Law.

Coming to @Google I/O and looking for good side events? Join us.

@yacineMTB True - or perhaps EQ more important than IQ in the future, do you agree?

You have to have a very high IQ to use AI to it's maximum potential

I’d be a DAU for this feature.🤣🤣🤣

@IterIntellectus the point is: if AI and robots are truly better than humans at every job we have today

i don't get people who say "there will be new jobs with AGI" like how? if AI and robots are truly better than humans at every job we have today, how is it possible for humans to still be competitive? "but every time new tech arrives, new jobs pop up" sure, for the AIs maybe, you don't see horses being hired for transport anymore. if any job were to emerge in the post-AGI era, definitionally AGI would be able to do it better. any company that could be founded would be founded by the AGI before you got there. if it needs dexterity, a humanoid robot already has it. i do not understand how people building cars can tell you with a straight face that there will still be an economy for carriage riders.

@thinkymachines World models and interaction models seem easy to confuse without clear definitions. - what’s the biggest difference

People talk, listen, watch, think, and collaborate at the same time, in real time. We've designed an AI that works with people the same way. We share our approach, early results, and a quick look at our model in action. thinkingmachines.ai/blog/interacti…

@Hesamation x.com/geekparkhq/sta…

AI research stands on the shoulders of: Chinese researchers in the US 🇺🇸 Chinese researchers in China 🇨🇳