Kevin Kelly was a guest at Chen Yuan 139 and held a dialogue with the emerging new forces of China's technology sector.
In 2026, AI is no longer unfamiliar to everyone. Large models have delivered sufficiently impressive results in the digital world, and technologies have further endowed intelligent bodies to step into the real physical environment. For the AI industry, this is an industrial offensive that breaks through human limitations. For a broader audience, this is a new industrial revolution that is enough to subvert all historical experience.
However, on the way forward for AI, many practical problems remain to be solved:
Is there a bubble in the AI industry?
Can embodied intelligence really understand the physical world?
Will robots completely replace humans? Can they coexist with humans, and where is the boundary?
Can Token cost determine the winner of AI?
What attitude should we take towards the associated risks of AI?
On September 21, Kevin Kelly, the founding editor of *Wired* magazine in the United States and a technology thinker, was a guest at Chen Yuan 139 in Beijing. He launched an in-depth dialogue of nearly 2 hours around AI with Wang He, Founder and CTO of Galaxy General Robotics, Zheng Qingsheng, Partner of HSG, Shi Yaqiong, Vice President of Jinqiu Fund, and Feng Dagang, CEO of 36Kr. Many practical issues including the technical inflection point of the world model, cost game, human-robot collaboration boundary, risk governance and technological optimism were discussed in this conversation, leaving six clear judgments.
Event Site
01
Judgment 1: Embodied Intelligence Has Broken Through the AlphaGo Moment
In August 2026, at the World Humanoid Robot Games, a humanoid robot developed by China's Galaxy General Robotics (Galbot) completed fully autonomous tennis match without remote control. The robot ran all over the court, received and served the ball, competed, stood up quickly on its own after falling, predicted the trajectory of the ball, and then hit the ball precisely.
After the video of the robot playing tennis was released on social media, Andrew Kang, co-founder of Mechanism Capital, commented: "AlphaGo for every sport is coming." Elon Musk also left a comment: "The AlphaGo moment for embodied intelligence."
When Wang He, founder and CTO of Galaxy General Robotics who developed this robot, met face to face with Kevin Kelly, the latter commented: "What is most talked about right now is the large language model that almost everyone is using, but the large language model itself knows very little about the real world. It is not trained in the real world, but trained on text and text data that describe the real world. Now, people are starting to give AI a body, and use robots to connect AI to the physical world. However, if you directly connect the large language model to the robot and let it act autonomously and cope with the environment in the real world, it will be very difficult. So our current goal is no longer just to build this kind of intelligence that 'knows all book knowledge', but to develop AI that understands the real world. Sometimes it is called a world model. Its training materials are no longer just texts from all over the world, but all physical laws, and massive amounts of data describing how the real world works. This is the next major breakthrough, and the whole industry is moving in this direction."
Event Site
How did AI complete its evolution from being able to perform one action to mastering a skill?
Wang He revealed the underlying logic of Galaxy General Robotics' R&D: "We found that the brain of a mouse is very small, and the number of neurons is relatively small compared with humans. But it has all the intelligence to find food and escape in the wild, and it does not need to have a brain as large as a human to survive. This kind of intelligence is a different form from the knowledge intelligence behind the text model. On this path, we also hope to build step by step a bionic brain system that connects the cerebrum, cerebellum and pons like a human, which can not only make high-level judgments, but also perform low-level balance execution, so that our humanoid robots can truly have wisdom."
From a global perspective, Google VLA and Open AI's world model represent two mainstream technical routes, but both have their own shortcomings. VLA training data must be bound to robot action labels, and massive public videos cannot be reused; the world model is good at environment prediction, but does not necessarily output action instructions that robots can directly execute.
To this end, Galaxy General Robotics proposed the World Action Model (WAM), trying to integrate the two paradigms. The model can not only predict the evolution of the environment, but also directly output robot action instructions. This technical route is called "the endgame technology for robots" by NVIDIA.
Wang He
Wang He said: "In the past, when robots danced, they needed to extract human trajectories in advance, put them on the robot, learn according to the trajectories, and deploy them after debugging. Now our general cerebellar model can make our robot dance the same dance immediately in front of a person after the person dances once. The intelligence behind the rapid imitation of hip-hop athletes' movements even surpasses human's ability to master limb movements. Two years ago, this path was not very realistic. But today, optimistic people in our industry, including Galaxy General Robotics, have built a pyramid of embodied intelligence data, from Internet data to human data, to synthetic simulation data, to data collected by remote-controlled robots, to data returned from autonomous robot operations. This five-layer pyramid is continuously being scaled up for implementation, promoting the robot's cerebrum and cerebellum to become more intelligent."
Facing these new breakthroughs, Kevin Kelly said: "I think the next big breakthrough will be about actually finally getting robots to work. What we have been working on will have a huge impact. So that's where some of the excitement goes."
02
Judgment 2: The Underlying Cognition of Robots Is Completely Different From That of Humans
When robots break through the AlphaGo moment, become smarter and more capable, will they compete with humans for ecological niches?
Kevin Kelly reminded: "When we give AI a body, that is, a robot, we need to understand that they are different from us, they are not human." He calls robots "heterogeneous intelligence" or "alien intelligence".
Wang He explained the "underlying" differences between robots and humans: "Robot intelligence is very different from human intelligence which has advanced thinking, judgment and reasoning logic. It is an intuitive intelligence. Most of human reasoning and thinking belong to the slow system in the brain, or System Two. And these human operations belong to the fast system in the brain, System One. The fast system requires a lot of practice to turn abilities into an innate, inadvertent subconscious reaction. It is a very different kind of intelligence from the large model that uses a very complex chain of thought."
In other words, the cognitive combination formed by AI is completely different from that of humans.
Event Site
Wang He believes that robots will become as intelligent as humans, but "robots are essentially AI, not humans". AI will not pose a fundamental threat to human ecological niches: "Does it have the desire to reproduce? What can robots get from reproduction? It does not have our genes, and there is no mechanism to screen and retain its own genes. Personally, I think many things we imagine today treat robots too much like humans, thinking that they will have the desire to replace humans. But I am quite optimistic about the future of robots. Under constraints, they will become human partners."
Kevin Kelly added that the future will also be diversified within the world of robots. AI products will differentiate into rich and diverse categories, and different products will form differentiation in pricing, cost, response delay and output capabilities. Eventually, the entire industry will evolve into a highly complex ecosystem. "In my opinion, these intelligent carriers with physical bodies are low-cost and valuable resources. We will build a society where we work hand in hand with these 'aliens'. " AI alone cannot solve all problems, but there is huge room for imagination for humans + AI.
03
Judgment 3: Token Cost Is Reshaping the Survival Logic of the AI Industry
For robots, optimistic people foresee a huge space for imagination brought by technological progress. However, there are still two thresholds between reality and optimistic imagination for the future: First, from the cognitive generalization ability to the real completion of physical interaction in complex environments, robots still need to continue to prove that they know "how to do it" and "can do it"; second, at present, the rapid development of AI is supported by huge data and computing power investment, and it still needs to be cheaper.
Kevin Kelly observed a clear change in Silicon Valley: In the past, when engineers called models, they rarely calculated Token costs. It was not until competitors could achieve similar effects at one percent of the cost that the entire industry suddenly woke up. Some complex reasoning tasks consume millions of Tokens at a time, and computing power overhead will directly rewrite the product business model.
"We are entering an era of huge gaps, and cost will become increasingly critical." Kevin Kelly believes that in many scenarios, price is crucial.
Cost pressure has been transmitted to the venture capital market. Shi Yaqiong, Vice President of Jinqiu Fund, provided a set of industrial observations in the roundtable: Large models have greatly lowered the threshold for AI product prototype development, but the user acquisition threshold has risen sharply. From 2022 to 2026, AI applications have grown by about 17 times, and the customer acquisition cost of AI products has generally increased by more than 50%.
Who can take the lead in the competition? Technical capabilities, as well as first-mover advantages laid by cost and scale, are all important. In the past, the industry spread the hypothesis of the Jeans Paradox: when the cost is reduced by ten times, the usage scale increases by ten times. But Shi Yaqiong believes that reality is far more radical than theory, and the expansion of usage brought by AI is not a hundred times. In the code generation scenario, the per capita code output has increased by 10,000 times. "This has brought many changes, allowing us to reassemble business logic in a coding way."
Shi Yaqiong
In the embodied robot track, the cost logic will change further. Wang He said that text large models can tolerate a certain reasoning delay, but humanoid robots and autonomous driving are strongly real-time edge systems. It is extremely risky to completely rely on the cloud to call Tokens. Exceeding the network and reasoning delay standards may cause robots to fall and cause operation accidents. Cloud-edge-end hybrid deployment will become an inevitable choice for the industry.
04
Judgment 4: AI Creates a World Without Upper Limit
Investors who have experienced the mobile Internet cycle can easily apply the Internet experience directly to the AI track. Zheng Qingsheng, Partner of HSG, warned at the roundtable: This reference framework has failed.
In the mobile Internet era, there are hard constraints on industry growth, and the ceiling comes from the total human attention. The physical world changes slowly, and cities and production modes will not undergo subversive reconstruction for decades. No matter how iteratively the Internet develops, it is essentially distributing in the fixed human time pool, and there is a clear boundary for the total user duration.
But AI is breaking this growth logic, and it is creating a world without upper limit. Zheng Qingsheng compares AI to basic energy sources such as steam and electricity, but it is essentially different from all energy revolutions in history.
Event Site
Traditional energy sources are constrained by mineral reserves, installed capacity and transmission networks, and there is a clear physical upper limit for supply. The core raw materials of AI are computing power and data, which can theoretically continue to expand. We can schedule thousands of AI Agents to perform tasks in parallel at the same time.
"Oil and electricity do have an upper limit, but I don't think AI does." Zheng Qingsheng said that AI can not only act as an energy source itself, but also create itself, with the ability of self-proliferation, which has never appeared before.
05
Judgment 5: Technological Risks Need to Be Resolved Jointly by New Technologies and Systems
As the shipment of humanoid robot hardware increases, the division of safety and responsibility has changed from theoretical discussion to an urgent practical issue in the industry.
Wang He took the deployed pharmacy robot as an example: In deterministic closed scenarios such as factories and pharmacies, the rights and responsibilities are relatively clear. Galaxy General Robotics has deployed nearly 100 robot pharmacy points in China, which are responsible for taking medicine at night. To avoid the risk of wrong delivery of medicines, the team did not fully rely on the general reasoning of the large model, but added an additional deterministic verification program. The medicines must complete three verifications of text, image and barcode at the same time before they are allowed to be packed and delivered from the warehouse. In the past year and a half, nearly a million boxes of medicines have been processed, with zero wrong delivery. In such scenarios, if the robot malfunction causes losses, the responsible subject points to the manufacturer.
The family is a completely open and complex scenario, full of unpredictable interference factors, and the risk complexity increases exponentially. The robot may knock over the water cup and cause a short circuit and fire, and the accident may even occur when no one is at home. The robot is still unable to independently deal with the chain consequences caused by the accident, which is the core reason why the industry will not sell unattended humanoid robots to ordinary families at this stage.
Kevin Kelly believes that in the short term, AI enterprises must assume corresponding responsibilities, and can also support supporting insurance mechanisms and so on. Conversely, the AI enterprises that can survive must be those that can handle the responsibility problem well.
