Cobo co-founder Shenyu said the AI wave has radically lowered the cost of turning an idea into something real, making human will and initiative more important than before. In a wide-ranging conversation with Jilian Technology founder Jin Ming, he also said Bitcoin remains "digital gold" and called it "a better gold." The interview, written by Da Qianzi and published by PANews, covered AI, entrepreneurship, investing, Bitcoin, Web3, agents and personal purpose.
Shenyu has been involved in crypto since 2011 and is described in the interview as one of China’s earliest Bitcoin evangelists and participants. Jin Ming is a China national freediving team athlete and founder of Jilian Technology. The article says he has long focused on video AI research and commercial applications and has received investment from Alibaba, Yunfeng Fund and Megvii.
AI lowers the back-end cost of execution
Jin Ming opened by recalling that Shenyu’s last interview appearance was in 2017, when he made a now well-known remark: "I suddenly didn’t know how to spend so much money." Shenyu replied, "Now I know how to spend it."
He said the deeper shift is that human intention is becoming more important in the AI era. Many tasks that once demanded major time and capital can now be done quickly if a person simply has an idea. In his view, that is a major change.
Shenyu said that 10 or 20 years ago, many things were hard to imagine, let alone deliver. Now the path from an idea to implementation has become much shorter. A dream that once stayed only in someone’s head can move toward execution because the back-end cost has fallen sharply. He added that AI has compressed and structured a large amount of human knowledge, allowing individuals to call on those capabilities at low cost.
That, he said, has opened a much larger field of possibilities. For him, the key point is not one standalone tool but the fact that the journey from idea to deployment has become materially easier.
A builder with questions still unsolved
Asked whether he is satisfied with his current life, Shenyu said things are "okay," though he still feels some confusion. He contrasted that with his earliest years as an entrepreneur, when his sense of mission was much stronger.
Now, he said, he feels he is in a more chaotic phase. AI has opened many possibilities, and he sees this as the best era in some respects, yet he still does not see a fully clear path, either in terms of mission or business direction. For now, he said, he is still exploring and improving himself.
Looking back, he described his earlier entrepreneurial drive as a response to a clear window of opportunity. When that kind of window appears, he said, it creates a powerful creative impulse. It feels like if you do not do the thing, you will regret it for the rest of your life.
Shenyu traced his early work to the computing power problem around 2013. Mining started in the United States and later shifted toward China. In the early ASIC miner era in China, the machines still had to connect to U.S. mining pools. That created two problems: delays and packet loss on the international internet, which hurt mining efficiency, and an industry settlement model that he said was not reasonable at the time.
His team’s work, he said, was to solve those two issues. He described it as identifying a market gap and a niche with unmet demand, then moving because he had the ability to address it.
Asked what his mission is now, Shenyu said he still wants to build things and continues to define himself as a builder. Since the start of the AI era, he said he has made quite a few small tools and dashboards, but he has not yet found a truly large and systematic opportunity and is still watching closely.
He added that after years of operating at the front line, founders often end up with more confusion, not less. That pushes them back toward history, philosophy and older bodies of knowledge. He said those things feed back into how he understands the present and the future.
"To make the world’s possibilities slightly larger"
Jin Ming framed purpose as a balance between creating big things and pursuing experience and happiness. Shenyu said he does not strongly identify with either extreme. He does not feel a special need to leave something behind, and he is not singularly focused on experience either.
What he does feel, he said, is an internal impulse to build. "I just want to build something. Once it’s built, I feel great." He said he wants to put something into the world because, in his words, living means that through your existence, the world’s range of possibilities becomes a little larger.
On anxiety, Shenyu said he does have some, but he has learned to switch states. Since childhood, he has had ways to empty his mind, including caring for animals and spending time on hobbies, and he said he generally enters that reset mode every week.
Freediving, fear and death anxiety
The conversation then moved to fear. Jin Ming described freediving as a process in which the body keeps sending signals while a diver holds their breath, creating a confrontation between reason and instinct. He called it something close to spiritual practice. A diver, he said, has to enter the water already knowing the outcome and repeatedly tell themselves that they are safe.
He said a person on land generally cannot easily hold their breath to death because the autonomic nervous system is involved. At the limit, the brain may enter a brief blackout, or BO, for a short period of unconsciousness. After roughly 10 seconds, the airway reopens and the body naturally breathes again.
Underwater, the danger is different. If consciousness is lost there, the airway reopens and takes in water instead of air, which can lead to drowning. That is why freediving training must have a buddy, he said.
Jin Ming said the training process is also built on positive mental cues: you are meditating, you are entering flow, you are healthy, you are safe, and your brain is smart. He added that beginners rarely approach their full limit. Early on, many only reach about 50% of their true limit. With more training, they move toward 70%, 80% and even 90%.
In many cases, he said, what stops new divers from staying down longer is not their physical limit but unfamiliarity with their own bodies. As that familiarity grows, trust grows with it, and the process can feel peaceful rather than like a struggle.
Shenyu asked whether this kind of training reduces fear of death or forces a person to face death more directly. Jin Ming said his fear of death has in fact dropped sharply after long-term freediving and breath-hold training. He said he used to feel that fear in his earliest sessions, but no longer does. He even jokes with his parents that his life already feels full and complete, and if an accident happened tomorrow, he would still feel he had lived happily.
Shenyu responded that he still has anxiety on some level and said outside observers, along with his own agent companions, would tell him that he has death anxiety.
Three AI agents and a note system
On information gathering, Shenyu said AI now handles almost all of it. When he wakes up, AI tells him it has read everything he went through the day before.
He said he currently uses three AI agents. One is instrumentally rational and tells him what he should do. Another is softer and more feminine in style and searches for what is deeper inside him. A third looks at what he did the previous day and points out blind spots it observed.
Those agents might, for example, tell him that a certain book in history contains a chapter discussing a problem ancient people also faced, one that closely matches his current state, and suggest spending 15 minutes on it. The relevant materials then appear immediately for review.
He also uses a note-taking system that links his past interests and life-related subjects using what he described as evolutionary thinking. In that framework, his attention acts like a "god" deciding which ideas or propositions survive. AI, he said, hybridizes a batch of new things each day and then selects one, telling him a certain article is worth reading.
As a result, he now rarely reads raw information directly. Instead, he receives AI-filtered dailies and long-form pieces after the AI has processed them. He usually compresses that output into PPT form for quick review.
DeFi helped him build an investing framework
Jin Ming noted that this method works better for existing information than for sudden real-time developments, such as a conflict breaking out seconds earlier. Shenyu said he does not care much about that. The most important information, he said, will find its way to you anyway, and he does not trade on a very short-term basis.
He said he recognized early that he does not have a talent for very short-term trading. Simpler decisions and longer-duration situations suit him better.
According to Shenyu, he only really began investing seriously after "finishing DeFi." He said DeFi effectively forced him to make up for the missing history of investment knowledge. From what he called the "underground Federal Reserve" onward, he got to see how traditional finance was invented and how it developed across roughly 200 years. That process produced many questions, which pushed him into more study of history and finance. That, he said, was when he truly started investing in a serious way.
He described himself as fortunate on day one, having somehow picked up methods that were not too bad and followed a path that was not too wrong. But only after the DeFi cycle did he build a real knowledge structure and investing framework. Before that, he said, he did not have a complete one.
Looking back on his mining years, Shenyu said he did not even think of himself as an investor then. He simply had a vague intuition that he should hold some Bitcoin, mixed with some early sense of position management. As an ISTJ, he said, many of his frameworks were abstracted gradually through practice: build a framework, train intuition, discover new questions, then abstract again.
From DeFi to RWA
Jin Ming argued that very few people build an investing framework through DeFi because many participants treat DeFi as "farm, sell and withdraw," which is a different logic from investing. Shenyu replied that the important question is what those financial products actually do and how they meet market demand.
For him, the key was that DeFi was building a financial market on-chain from zero to one. That process mattered more than the surface-level actions of arbitrage or mining, which he said were closer to muscle memory and intuition.
He said the experiences gained there were valuable, and so were the many questions that came out of them. At first, a person may not know why the market needs these products. Once the answer becomes clear, history starts to connect.
When Jin Ming said the current wave of RWA and the introduction of U.S. equities on-chain had effectively extended that DeFi-based understanding of financial products into traditional finance, Shenyu agreed. It puts U.S. stocks right in front of you, he said, and forces you to understand the logic behind them.
He recalled that when he was mining 10 years earlier, he never imagined holding U.S. stocks. He said he did not even fully understand it when people asked to borrow money from him to trade them. The later learning curve became much faster because he had accumulated so many questions in practice and then gained AI tools that let him move through large numbers of books and materials quickly.
Bitcoin does not need a fresh narrative
Asked how his view of Bitcoin has changed after more than a decade, Shenyu said it has not changed at all: it is still digital gold.
He argued that in a world where most assets sit behind one, two or even three layers of intermediaries, Bitcoin remains unusual because it can be held without an intermediary. That is one reason his view has stayed the same.
As for the more pessimistic tone among some market participants, he said Bitcoin’s value is precisely the kind of thing that shows up under extreme conditions. When markets are booming, he said, it is normal that many people do not care much about it.
Jin Ming asked whether Bitcoin materially strengthens the sovereign individual under a classical liberal framework. Shenyu replied that one core feature of the sovereign individual is the ability to dispose of one’s own assets. AI strengthens capability, while BTC is a core asset in the way sovereign individuals handle wealth. Both matter, he said.
When Jin Ming said many early BTC believers now feel there is no new narrative in this cycle, Shenyu answered: "Good assets do not need narratives, and the longer they exist, the more they prove certain things."
Ethereum may end up serving agents more than humans
Jin Ming said that in his view, the main Web3 narratives in this cycle may be Ethena’s all-asset arbitrage model and the move to put RWA on-chain, though both remain early. He also said a true mass-market consumer hit still has not appeared in Web3, meaning something ordinary users could use as naturally as WeChat and that could genuinely change daily life.
Shenyu said he has long held a different view: ecosystems such as Ethereum may not end up serving humans as the main users at all. They may end up serving agents. In that model, the chain becomes a network and humans may not be the best endpoint.
He said many on-chain products still demand too much professional knowledge and too much security awareness from users. If humans remain the direct end users, the barrier stays high. If agents become the middle layer, he said, that could be a better presentation model.
Asked what that would mean for ecosystems such as Ethereum and Solana if the main actors running on them were agents, Shenyu said he is watching one thing in particular: whether agents can achieve large-scale adoption in an ecosystem like Ethereum. If that happens, he said, it could mark a turning point.
Jin Ming also argued that one value proposition of Web3 is still heavily underestimated. Moving from read-only to read-write to read-write-own, he said, may reshape how wealth is distributed across society. In finance, he said, the Web3 version would let users not only participate but also own part of the upside. He cited Binance and Hyperliquid as examples where users generate fees for a platform and may also share in ecosystem growth through tokens, buybacks or similar mechanisms, making them both participants and something like shareholders.
Shenyu said he thinks structures like that will emerge. Data in the AI era, he said, could itself create similar mechanisms. The possibility for technological diffusion is already open. Large-scale adoption still requires time and many unresolved issues, but he said successful demos can already be seen in some edge cases.
"Bitcoin is a better gold"
When asked directly about gold, Shenyu said, "I don’t really look at these things. Bitcoin is a better gold."
Jin Ming said the two assets share one trait: consensus. The difference, he said, is that Bitcoin has formed part of its consensus at the level of sovereign individuals, while gold has stronger consensus at the level of sovereign states, especially central banks. He added that central banks around the world are still buying gold.
Shenyu then reduced the divide to two broad camps: one centered on "digital currency plus AI," represented by Bitcoin, and another centered on "gold plus industrial manufacturing power." He said they are more aligned with the former.
On diversification, Jin Ming referred to James Tobin and the argument for holding low-correlation or negatively correlated assets to capture what markets call a free lunch. He said Bitcoin and gold have sometimes been positively correlated, but over the past year have also shown some degree of negative correlation.
Shenyu said he does not believe truly negative-correlation assets really exist in a deep sense. Many assets appear diversified under normal conditions, he said, but they still share several core constraints. When those market variables change systematically and conditions turn extreme, many assets start behaving in very similar ways.
Jin Ming pushed back with cases such as oil versus some credit assets and U.S. Treasuries versus credit risk. Shenyu answered that equities and Treasuries can also fall together. In his view, what matters is resilience in the most extreme cases. Something that looks diversifying in ordinary times may not protect you when stress becomes severe.
That is why, he said, he is not naturally a broad diversifier. He is more concentrated, because no one can truly understand too many things at once, and human attention is limited. If the underlying logic is not clear enough, he would rather focus capital on a small number of assets he genuinely understands.
Belief, temperament and position sizing
Jin Ming said some assets may not need to be fully understood in depth. Gold, for example, could still be part of a well-diversified allocation if it has already established long-term consensus among sovereign central banks.
Shenyu replied that the problem is psychological pressure when an asset falls. The simplest way to avoid that pressure, he said, is not to hold the asset. If Bitcoin falls and you have real conviction, you will not panic in the same way. But that conviction, he said, only comes after a large amount of prior work.
Jin Ming said ordinary people face a bigger barrier because most are not professional investors and struggle with position sizing, drawdown control and emotional management. Shenyu then asked whether sports such as freediving are mainly trainable or shaped more by natural temperament.
Jin Ming said it is probably both. He added that sport may become a necessity for many people, as more individuals pay attention to health, commit deeply to one activity and gradually become athletes in a broad sense. That, he said, can generate both satisfaction and happiness.
He tied that back to what he called a trading lifestyle. In that setup, a Bitcoin drawdown could become a better accumulation window for a certain type of investor. He gave a price ladder as an example: at $120,000 Bitcoin may feel hard to buy, and he personally would not be eager to accumulate much above $100,000. Below $90,000, though, he would see an opportunity and would consider buying at $90,000, $80,000, $70,000 and $60,000 gradually because he remains positive over the long term.
AI as a cycle, or as the next era’s utilities
Jin Ming said it is inaccurate to treat AI as a pure cyclical trade. In his view, AI will become the most important productive force of the next era and replace a large amount of what keeps the world running today. If that assumption is right, AI is not just another cycle but the equivalent of water, electricity and coal for a new era. If the assumption is wrong, then it can be treated as cyclical.
Shenyu answered with a historical comparison. During the internet bubble around 2000, he said, the market also believed the internet was unquestionably the future. Network infrastructure was the key theme, and many people bought Cisco. Cisco then went through a long period of valuation digestion. His question was whether AI could follow a similar path.
Jin Ming said anything is possible, but if the AI era really arrives, storage demand will be enormous. For now, he said, the market mostly has conversational bots plus some coding bots. Much bigger demand may come from embodied intelligence, meaning real robots. Households do not yet have truly embodied intelligent robots, and workplace robots are still concentrated in a small number of factories. Large white-collar office environments have not really been replaced.
He added that future humans may also need more external memory, whether through multiple backups of the brain or some form of extra "external brain" through brain-computer interfaces. Those needs have not materialized yet, but if the AI era does arrive, the demand could be huge. The central question, he said, remains whether one believes the AI era will definitely come. His answer is yes, and he added that robotics is no longer merely a wild assumption because the current path of human technology already points toward the possibility of massive production.
How he screens core assets
Returning to portfolio construction, Shenyu said this again comes back to trading lifestyle. A person’s temperament influences the investment philosophy and framework they choose, and that ultimately shapes position management and emotional control.
Jin Ming said each person is different, but the people who do well in the long run usually create an internal system that is truly their own. Even the definition of a top-tier or T0 asset differs from one framework to another.
Shenyu then outlined his three-layer process. First comes anomaly detection: a company appears unusual in some way. At that stage, he said, they do not open with a heavy position. They might buy no more than 2%, and often just a few tenths of a percent, as an observation position. Then they spend time and energy studying it. If the logic proves sound, the allocation can be upgraded from a small fraction to low single digits, though generally not above 10%.
Moving from around 10% to around 20% as a core holding requires what he called a leap of faith. Once an asset enters the core position tier, the decision can no longer rely only on fundamentals and logical deduction. It also needs a sufficiently large vision, a sufficiently large TAM, or total addressable market, a strong business model and a real key bottleneck or strategic choke point.
He summed up the current framework as "monopoly plus growth." A core asset must have a very large TAM and a big enough vision to support its place in the portfolio.
Current core assets: Bitcoin, Ethereum and Tesla
Asked which assets currently meet that standard, Shenyu named Bitcoin, Ethereum and Tesla. SpaceX, he said, is still in the timing and observation phase and has not been upgraded into that tier.
He also said Ethereum remains in the group for now, though they have recently been discussing whether it should be downgraded from that level.
Jin Ming suggested identity may also play a role and recalled that when Shenyu fully exited Ethereum a year earlier, he had said in a group chat that he "slept better," and Shenyu had said the same. Shenyu agreed and said that was because he was no longer farming DeFi.
Still, he said the key logic they now see is that agents may run on Ethereum in the future, and that thesis has not yet been falsified. Asked why agents could not instead run on Base or Solana, he said they can, but Ethereum may be a better match because it still has a larger set of decentralized infrastructure at the base layer. He added that while much of this has not yet turned into anything very large, the foundation is trying to move in that direction, including registry and ecosystem infrastructure around agents. The work has not yet produced a breakout result, he said, but it has at least created some basis for hope.