Silicon Valley Study Tour Summary: Crypto, AI, and the Secrets of Silicon Valley's Innovation

From August 17th to 23rd, I participated in the 2025 Silicon Valley Crypto Industry Study Tour organized by Uweb. This was my first time participating in a full study tour of this kind, a first for me. I'd previously viewed study tours from afar, and distance created a bias and a certain skepticism. I was skeptical about what a group of strangers, each with their own unique goals, could learn from a brief, cursory visit. This was especially true when I read some of the eloquent, overly exaggerated study tour summaries others wrote, which felt like advertisements and were therefore suspicious.

This time I participated in the whole process myself, and to my surprise, I gained a lot. Silicon Valley is the global center of technological innovation, and I have many old friends in Silicon Valley, and I have been there many times. Especially in the past year, I have been there several times for work and personal reasons, so I didn’t have high expectations before this trip. It’s a place I often visit, and it’s also a popular attraction for Chinese people to make a technological pilgrimage to the United States. I have seen so much sincere sharing and boasting, what else can be new? However, this time, under the leadership of Principal Yu Jianing, more than 20 group members spent a week full of rich content and explosive information. To be honest, I feel very rewarded. Afterwards, I asked myself, why did I go there several times and meet so many people, but didn’t gain so much? The answer seems obvious. Going to Silicon Valley on my own only partially taps into my own network of resources, whereas a study tour allows me to aggregate and superimpose the resources of the organizer and all the group members, creating a concentrated energy field that connects with my own thoughts and sparks a lot of reflection. After returning to Australia and reflecting for a few days, I now feel the need to summarize some of the key takeaways from this study tour and share them with my readers. Because of this study tour, my previous series of summaries of my Hong Kong trip has been delayed. However, it was a blessing in disguise: my trip to Silicon Valley, with its comparative perspective, gave me a clearer understanding of the Hong Kong crypto landscape. For the time being, I'll focus on writing summaries of my trips to both Silicon Valley and Hong Kong. While it's still fresh in my mind, I wanted to jump in and post my Silicon Valley summary first. Some readers have been urging me to update my Hong Kong summary, so please bear with me. I won't delay this time. This study tour was incredibly rich, and it's impossible to fully document it. I've simply summarized the key insights that struck me. Each section focuses on a specific topic and can be considered a standalone article. These represent only my personal reflections and cannot fully and objectively reflect the entire study tour, nor do they constitute investment advice. A New Crypto Boom is Becoming a Consensus My main observation from this trip to Silicon Valley was that there's widespread consensus within the American tech community that a crypto boom is imminent. Everyone we met and interacted with, regardless of their level of crypto involvement or understanding of the industry, even including some who harbored biases, believed that crypto was poised for a boom. When reaching this conclusion, they didn't give much thought to the attitudes of other countries and regions, such as whether China and Europe would follow suit. Instead, they believed that even if the rest of the world didn't follow suit or respond, relying solely on US policies, funding, and technological capabilities would be enough to trigger such an industry boom. The root of this optimism, needless to say, was primarily the Trump administration's new crypto policy. In addition to actively promoting the GENIUS Act and the CLARITY Act, core members of the Trump administration publicly supported crypto on various occasions and actively engaged in related businesses. Therefore, as various conditions gradually converged, people in Silicon Valley generally believed that the crypto industry was about to usher in a long-term, large-scale boom, and that the United States would not only be the source of this crypto boom, but also its center. This judgment conveys a sense of urgency. On the last day of my trip, I visited the Jensen Huang Building at Stanford University's School of Engineering and saw the following exhibit: Google's first server, hand-built by the two founders in their Stanford doctoral student dormitory in 1996. The outer chassis was made of many Lego bricks. This is a famous exhibit at Stanford Engineering, and I'm sure many people have seen it. But for someone my age, seeing it still holds a special feeling. Someone familiar with the history of China's internet told me that around the time this server was built, China was fortunate to have chosen the right direction for internet management. At the time, there was a plan for the telecommunications department to manage the internet similarly to landline telephones. Had that path been taken, China's internet today would likely be lost in the crowd, and the story of China over the past two decades would undoubtedly have been completely different. Fortunately, China made the right choice, and the internet industry has achieved tremendous success over the past three decades. Times have changed. I wonder if people 30 years from now will still be able to talk about China's blockchain industry with the confidence that "we didn't miss out." The "Stablecoin War" Didn't Arrive as Expected When the US Stablecoin Act passed in July, a wave of discussion about stablecoins swept the world. At the time, I predicted that once the bill passed, all institutions and companies eligible to issue dollar-denominated stablecoins would quickly do so, and a stablecoin war would ensue in a short period of time. One of the purposes of my trip to Silicon Valley this time was to find out if this was happening. Why is this important? Because it affects the speed of adoption of "stablecoin payments." Anyone familiar with this field knows that stablecoins offer significant efficiency improvements over traditional payments in cross-border payments. However, traditional payments are currently a highly competitive field, with strong players in each sector. While stablecoin payments are advanced, there are currently no robust solution providers, and creating a niche is challenging. What will truly drive stablecoins' rapid impact on traditional payments is for existing payment companies and banks to proactively issue stablecoins. A major stablecoin war could create a sense of urgency and greatly accelerate this process. Unfortunately, my trip to Silicon Valley forced me to admit that my original view was incorrect. We haven't witnessed a "stablecoin war." This doesn't mean that new stablecoins aren't being planned and designed. As I write this, Hyperliquid announced the issuance of its new US dollar stablecoin, USDH, proving that new players will continue to join the stablecoin fray. However, the anticipated "rush" into the market hasn't materialized. This is especially true for banks and major internet platforms, which stand to benefit the most from issuing stablecoins and are best positioned to drive stablecoin payments into the real economy. So far, they've remained quite restrained and calm. Why hasn't the stablecoin war unfolded as expected? I have three hypotheses. First, perhaps these banks and internet companies aren't ready yet. Trump's election caused a 180-degree U-turn in US stablecoin policy within six months, leaving these large institutions perhaps too little time to make a decision. The second reason is that the threshold set by the bill acts as a tripwire. During the study tour, the organizers invited a senior legal counsel from Coinbase to speak with us. This counsel mentioned that Coinbase, through highly effective political donations, has strongly influenced the strategy of the US crypto industry, including the passage of bills like GENIUS. This led to the inclusion of carefully designed, targeted blocking measures to precisely hinder potential competitors from entering the US dollar stablecoin market. This can be considered a self-protection strategy for "first-in-comers" in the crypto industry. While it may sound unethical, it is actually reasonable. The third reason, and perhaps the most profound, is the "innovator's dilemma" identified by Christensen. This refers to the fact that because new technologies conflict significantly with a company's current core business and existing interests, the innovation departments of market-leading companies are often suppressed by corporate politics, making them unable to promote disruptive innovation, especially when such innovation is controversial. This effect has been widely discussed, with recent examples like Kodak and Nokia serving as cautionary tales, but it is difficult to overcome. Stablecoin payments are a classic example of a controversial disruptive technology. If, like AI, there were unanimous agreement, the innovator's dilemma wouldn't arise. The concern is that technologies like blockchain, with their divergent opinions, are particularly prone to plunging large organizations into the innovator's dilemma. Even now, many people remain skeptical of the technical and economic value of blockchain payments, or even stubbornly deny it. They often assert that blockchain payments offer no advantages over their own carefully crafted proprietary technologies. Ordinary users fail to grasp the seemingly intangible yet powerful forces of open systems and network effects. Based solely on superficial user experiences, they are unable to discern what is right and wrong, nor can they convey clear signals to decision-makers. As a result, blockchain departments within large companies are severely disadvantaged, often losing ground in the competition for internal resources and struggling to convince CEOs of their potential success. I've encountered numerous blockchain departments within banks and payment companies, and they often struggle to establish a meaningful presence within the company, struggling to drive meaningful change. Even today, many outsiders still believe that the opportunity for stablecoin payments lies with large banks or online payment companies with established user bases. I, on the other hand, believe that exchanges and cross-border e-commerce companies have a much higher chance of success in the stablecoin payment market than banks and online payment companies. However, I must admit that the anticipated stablecoin war hasn't happened yet. I still believe such a situation will occur, but it's still brewing. The crypto industry has entered a RWA cycle, and a large-scale stock swap is imminent. Before traveling to the US, one of my biggest questions was, "Will there still be an altcoin season?" I hope to form a clear judgment through this exchange. The so-called "altcoin season" is the industry's abbreviation for the "altcoin bull market." The term "altcoins" originally referred to all digital currencies other than Bitcoin. However, as digital currencies like Ethereum have stood the test of time and established themselves, the term now refers to an asset class alongside "mainstream currencies," broadly referring to digital assets with small market capitalizations, low liquidity, and low rankings. During the past two major bull markets, altcoins experienced collective surges of tens or even hundreds of times, often referred to as "alt seasons." The typical pattern of a crypto bull market is that Bitcoin recovers first, then breaks through its previous high, experiencing a significant surge. Then, Ethereum awakens, growing at a rate and magnitude exceeding Bitcoin's, subsequently triggering the arrival of an altcoin season. This altcoin season typically marks the climax of a bull market, fostering new assets within the industry and incubating a new generation of mainstream projects, while also laying the groundwork for a market crash. Since the beginning of this bull market, Bitcoin, Ethereum, and other mainstream digital currencies have achieved breakthroughs. The aforementioned pattern is only halfway through. The key question now becomes: will an alt season arrive as expected? During this trip to Silicon Valley, we met with numerous crypto institutions and experts. Through our conversations with them, I have reached a clear conclusion: this bull market will not see an alt season. Alternatively, while there will be an alt season, the stakes will differ from previous ones. Instead of a dazzling array of exotic crypto altcoins, the focus will be on RWA concept assets. Therefore, it can no longer be called an alt season. I have three reasons for this conclusion. First, the dominant players have changed. The previous bull market took place amidst an unprecedented global monetary easing campaign, with many governments providing direct cash to households. Retail investors became unprecedentedly powerful and played a major role in the bull market. They not only fueled the cryptocurrency market's surge, but also enabled them to compete with mainstream institutions on Wall Street. However, starting in 2022, a monetary tightening cycle and a market crash wiped out most of the wealth held by retail investors, leaving only institutional investors with funds. Therefore, starting in 2023, it became clear that Wall Street and institutional investors had become the dominant players in the market. Following the shift in US crypto policy, a large number of professional institutions entered the market, further consolidating their dominance. The entry of these professional institutions will lead to fundamental changes in the market's liquidity preferences and compliance awareness. I doubt whether the altcoin projects of the past, boasting a few demos and boasting big stories, can gain recognition from leading institutions. Second, the market mindset has changed. With the entry of mainstream institutions and mainstream capital, the mindsets of entrepreneurs and investors have also shifted. Some Silicon Valley crypto VCs I used to be familiar with have adjusted their perspectives, focusing more on projects related to stablecoins and RWAs, and placing greater emphasis on equity investments. The reckless issuance of altcoins is now seen as a negative indicator. Third, the industry theme has shifted. This bull market is undoubtedly centered around RWAs. Please note that in the Chinese community, when RWAs are mentioned, many people quickly think of very "real" and "tangible" assets like real estate, land, minerals, jade, and antiques. However, in reality, real-world "virtual assets" like bonds, equity, copyrights, and securities are the more scalable and easier to manage RWAs. The logic behind RWA implementation is to first put these already virtualized, homogenized, and securitized RWAs on the blockchain, and only then will we gradually move on to more tangible and "real" assets. Currently, in the US, the current focus of RWAs is particularly specific and focused on US stocks. If we expand this focus a bit, high-quality private company equity is also a concern. One thing we realized during our time in Silicon Valley is that if all types of corporate equity could issue tokens, then the competitiveness of pure-play altcoins would be very weak. Most of these altcoins are solely based on the trading and speculation of digital assets, with little or no real-world connection, and their teams lack real-world resources and experience. In contrast, in traditional markets, a vast number of high-quality company stocks, equity, and other interests have yet to be tokenized. If high-quality projects in AI, biomedicine, new energy, and smart hardware issue tokens based on their equity and enter the crypto market, can those altcoins, often hunkered down in their own little corners, truly compete? Of course, I also know that within the crypto community, there are some experienced "market manipulators" who are well versed in human nature and adept at manipulating the market during bull markets, profiting from them. They reap rewards in every bull market. Overall, however, I believe that with the influx of high-quality RWA assets, the market will shift towards RWAs, as liquidity is limited. Currently, the on-chain implementation of RWAs is still in the works, as the CLARITY Act has not yet been passed. However, institutions interested in crypto are already adjusting their preferences, and I believe a large-scale market rotation will soon occur. Copycat projects must somehow seize the opportunity presented by RWAs, or the future will be even more bleak. From the perspective of the crypto industry's development, the move toward RWAs is highly beneficial, as it signals a return to an open system. Looking back at the crypto industry's history, from 2009 to 2017, although the crypto industry was in its infancy, blockchain infrastructure was very primitive, and many scammers emerged, crypto was an open system, with people eager to transform the world with new technologies. However, after 2018, as major countries adopted negative policies toward crypto, crypto gradually lost its connection to the real world and became a closed system. Most crypto projects revolved around the "demand" of speculation and gambling, leading to an increasingly illusory industry. This is the inevitable fate of closed systems: after ceasing energy exchange with the outside world, entropy gradually increases, eventually leading to a "heat death" characterized by the absence of any meaningful physical laws. By the end of 2024 and early 2025, when the meme craze was surging, the entire crypto market had lost any apparent pattern. It was completely mired in a bot-driven gambling game between market makers and retail investors. This was the classic dead end of a closed system. Fortunately, the crypto industry didn't linger on this path for too long. With the launch of stablecoins and RWAs, the entire system reopened and began to exchange energy with the real world. Many people may not realize that this was actually the path to crypto's rebirth. Three Hot Topics in US Crypto: Summarizing our discussions with the crypto industry in Silicon Valley, we can see three hot topics currently in the US crypto industry: coin-stock linkage, US stock blockchains, and the ubiquitous exchange of everything. The so-called "crypto-stock linkage" involves creating a resonant relationship between the stock market and the cryptocurrency market. Currently, this linkage has a substantial effect on both the US and Hong Kong stock markets, though the two markets are at different stages of development. In Hong Kong, this linkage is still primarily driven by hype surrounding blockchain and digital assets, while in the US, it primarily takes the form of Digital Asset Treasury (DAT) companies. I met with several professional institutions in the US, and they are generally actively exploring and operating DAT listings. This model, thanks to the successful implementation of MicroStrategy, has become the most mature and effective path in the market. However, from the perspective of cryptocurrency-stock linkage, its development is relatively shallow, lacking deep business insights and the potential of the token economy, making it merely an "entry-level" model. The recent rapid market correction of DATs has raised concerns about the sustainability of this model. Some Silicon Valley institutions with experience in the cryptocurrency industry have begun planning DAT 2.0 models, but there are varying interpretations of what this 2.0 model actually is. Which model will prevail remains to be determined in the market. I firmly believe that the continued development of the DAT model will inevitably lead to a deep integration of the token economy and business. However, even a simple DAT still requires many operational details. While in Silicon Valley, we consulted with some experienced institutions about DAT implementations, such as SPAC mergers and acquisitions and RTOs (reverse-market offerings), and found them to be quite challenging and costly. I believe there should be more ideas for cryptocurrency-stock integration, and it won't be limited to DATs. The listing of US stocks on blockchains is a clear hot topic currently in development. Coinbase, Robinhood, and Kraken have all announced clear plans. Robinhood was quick to act, launching an on-chain US stock token based on Arbitrum in Europe, bringing over 200 US stocks and ETFs to market. Kraken, on the other hand, launched xStocks for non-US customers, enabling seamless trading on the Solana mainchain and currently supporting over 50 US stocks and ETFs. Coinbase, deeply engaged in the US market, views its stock token as part of its broader "Everything Exchange" strategy and is actively pursuing compliance pathways with the SEC, including seeking a no-action letter or enforcement waiver, to enable legal tokenized stock trading. I originally thought that after the passage of the stablecoin bill, the industry would take some time to digest the benefits of stablecoins and fully promote their implementation. However, it now appears that my assessment was wrong. The industry, far from clinging to stablecoins, has instead moved directly into US stocks. Promoting the implementation of stablecoin payments will impact the interests of banks and traditional payment companies, affecting many aspects, so this can only be done gradually. The industry's leading edge will undoubtedly advance along the path of least resistance: rapidly bringing more high-quality assets onto the blockchain and creating trading pairs with stablecoins. US stocks are undoubtedly the current hot topic. I believe that in a few months, the blockchainization of stocks will become the industry's hottest topic. The third hot topic is the so-called "Everything Exchange." This time in Silicon Valley, I carefully explored the areas of focus for venture capital firms and found that exchanges are a prominent focus. This is understandable. Exchanges are the leaders of the entire crypto ecosystem and the apex of the food chain, but competition in this sector is fierce, making it difficult for latecomers to break through. However, every time the market shifts and new asset categories emerge, new rules of the game, new user groups, and new market structures emerge, creating opportunities for a shakeup in the exchange sector. The explosion of Bitcoin in 2011, altcoins like Litecoin in 2013, and Ethereum's ERC-20 tokens in 2017 all gave rise to dominant exchange giants. If this pattern continues, a new generation of exchanges will inevitably emerge as the industry transitions to RWAs. So what will this new generation of exchanges look like? In several of his crypto speeches, the new SEC Chairman, Atkins, has repeatedly mentioned so-called "super apps," or platforms capable of trading all asset classes within a single application. As the US crypto industry absorbed these platforms, they were renamed "Exchanges for Everything" and are currently a hotspot for US crypto venture capital. Currently, there are two approaches to the emergence of Exchanges for Everything: centralized exchanges for everything, exemplified by Coinbase's "Project Diamond," and decentralized exchanges for everything, exemplified by Hyperliquid. Both have laid out clear plans to consolidate all trading instruments—digital assets, stocks, bonds, gold, foreign exchange—on a single platform, even going so far as to embrace the lucrative Polymarket prediction market. This is truly ambitious and ambitious. Whether or not we admire this gluttonous beast, an all-in-one exchange conforms to market principles and the principles of network effects. Aggregating all trading instruments, liquidity, information, and users on a single platform is undoubtedly the holy grail of the trading market and the most efficient. While such an exchange is of course impossible to achieve in reality, the process of moving towards this goal will undoubtedly create a behemoth platform and profoundly transform the industry's structure and rules of the game. I hope the Chinese crypto industry can prepare for and plan for this trend. Silicon Valley relies on its "circle of acquaintances" to form a unique innovation hub. During this study tour, the organizers fully considered the unique characteristics and richness of Silicon Valley. Rather than focusing entirely on crypto, they included numerous discussions on AI and other innovative technologies. In the first session, renowned technology expert and science writer Wu Jun gave a comprehensive introduction to Silicon Valley and global technological innovation trends. He then invited several prominent speakers to share insights with us, helping us understand the region. Having visited Silicon Valley many times, I'm familiar with some of its statistics. For example, Silicon Valley, including San Francisco, covers a total area of ​​4,800 square kilometers, with the core built-up area comprising just over 500 square kilometers. The Bay Area has a total population of approximately 9 million, but the Silicon Valley tech belt has a population of only 3 million. Professor Wu Jun provided some more detailed statistics, such as 1.7 million employees, 150,000 software engineers, an annual household income of $200,000, and 40% first-generation immigrants, with Chinese and Indian ancestry each accounting for 6%. These are very interesting data. We often say that Singapore is a small country, but in terms of built-up area and population, Silicon Valley is even smaller than Singapore, yet it has achieved such remarkable success. I believe that most people who visit Silicon Valley have one question in mind: "Why is this place so amazing?" Silicon Valley is undoubtedly the most successful and entrepreneurial-friendly technology innovation hub on the planet. Regarding venture capital, Silicon Valley and the San Francisco Bay Area attracted a combined $69.7 billion in venture capital in 2024, a 125% year-on-year increase, accounting for 52% of all venture capital in the United States. In addition to those renowned publicly traded tech giants, Silicon Valley boasts nearly 300 privately held unicorns, representing nearly 40% of all unicorns in the United States. In a private conversation, a Chinese VC manager in Silicon Valley proudly stated that 20% of the world's successful venture capital opportunities are within a 40-minute drive. A Wall Street investment banker friend of mine told me that Wall Street won't look at startups unless they originate from Silicon Valley or Israel. What exactly is it that makes Silicon Valley so innovative? There's a wealth of discussion on this topic, and even in Chinese literature, there are diverse perspectives. Some emphasize the role of Stanford University, others attribute it to the shaping of Silicon Valley culture by early tech companies like HP, Fairchild, Intel, and Apple, others credit the Valley's abundant venture capital and industrial clusters, and many simply explain it as the region that produces the world's brightest minds. But these arguments aren't particularly convincing to me, as I feel many of them confuse cause and effect. Take the concept of "brilliant minds," for example. It's not that Silicon Valley's soil and water contain some mutant microbes that breed so many geniuses, but rather that talent from all over the world constantly flocks to Silicon Valley. To be honest, many of the technological innovations often attributed to Silicon Valley didn't originate in Silicon Valley, or even in the United States. Instead, they arrived in Silicon Valley later, transitioning from innovation to entrepreneurship. As a local Silicon Valley investor put it during this study tour, Silicon Valley's greatest strength isn't technological innovation, but its ability to "process" innovation, talent, capital, and systems into successful startups. Therefore, my question is, what factors make Silicon Valley so uniquely adept at incubating and nurturing startups? This question is largely informed by my entrepreneurial experience over the past few years. My experience over the past few years has made me deeply aware that entrepreneurship is inseparable from venture capital, which is based on integrity, and integrity is the most difficult quality to assess in investment. For a startup to earn the trust of investors, receive warm cash, and then, in a game where it's clear they can get away with losing all the money or pocketing it under false pretenses, overcome all kinds of difficulties and obstacles, resist all kinds of temptations, and steadily build the project, is truly a violation of human nature. It requires tremendous willpower and self-praise. On the one hand, without sufficient financial support, the vast majority of startups cannot reach that turning point. On the other hand, it is extremely difficult for investment institutions to identify honest and capable entrepreneurs. Therefore, China's venture capital industry often introduces clauses such as "betting" and "buybacks" to protect investors' rights. However, this practice completely shifts the entrepreneurial risk onto the entrepreneur, deviating from the essence of venture capital and only stifling innovation. I have complained on many occasions about these degenerate "venture capital" companies in China, believing them to be a "Chinese characteristic." But after traveling abroad, I discovered that most so-called "venture capital" firms in Asia, whether in Hong Kong, Singapore, or Malaysia, come with one toxic clause or another. In comparison, Silicon Valley's authentic venture capital is truly VC, a unique exception. This intensified my curiosity about how Silicon Valley balances trust and constraints to become a unique global hub for startup incubation. This trip to Silicon Valley has given me some new perspectives on this question. Through my limited contact with the Silicon Valley venture capital community, I seem to believe that venture capital in Silicon Valley is fundamentally based on "circles of acquaintances." Entrepreneurs and investors often connect through strong relationships—classmates, colleagues, and shared interests—and are connected through a series of intimate connections. They maintain extremely high standards, undergoing long-term connections, selection, and elimination. They mutually convey trust, offer commitments, and impose constraints, forming distinct, hierarchical circles of trust. For investors, entrepreneurs who are ultimately successfully selected to enter these circles have already undergone extensive screening and scrutiny, and are subject to numerous soft constraints, making them trustworthy. For entrepreneurs, once they reach this stage, they can secure trust and resource support that would be difficult to obtain elsewhere. They are provided with money, connections, and connections, and their chances of success naturally increase a hundredfold. The consequences of "tampering" are extremely serious, and rational choices naturally lead them to follow the good advice. To put it another way, while Silicon Valley's products are among the world's top tech innovators, they rely on a very old-fashioned system of acquaintances and social networks, rather than advanced game-playing mechanisms or innovative financial tools. It is precisely by consolidating entrepreneurs and investors into a small, 24/7 environment that this powerful, binding, and supportive innovation incubation mechanism is formed, guiding and forcing entrepreneurs to follow the right path while providing maximum resource pressure to help them succeed. Many people marvel at Silicon Valley's "smallness," but in reality, its success stems precisely from its small size, allowing the network of acquaintances to thrive. I shared this "network of acquaintances" theory with several Silicon Valley friends, and they resonated with me. My question is, is this experience replicable? How does Silicon Valley view the "AI bubble" theory? When you come to Silicon Valley, you can't just talk about crypto. When you come to Silicon Valley, you can't avoid AI. In fact, the center of US crypto is in New York, and Silicon Valley's unwavering theme is AI. Before this study tour, I had already decided to learn more about Silicon Valley's perspective on the now-burgeoning "AI bubble" theory. The "AI bubble theory" may be unfamiliar to many. From what I've read online in Chinese, the current optimism about AI is largely unanimous, almost becoming politically correct. But in reality, concerns about an AI bubble have only intensified over the past few years. Ever since the breakthrough of ChatGPT 3.5, a group of experts, led by Yann LeCun, Chief AI Scientist at Meta, have publicly expressed skepticism and even criticism about whether Large Language Models (LLMs) can achieve strong artificial general intelligence (AGI). With the release of ChatGPT 5, the development of core capabilities for large models has clearly slowed. Against this backdrop, the "AI bubble theory" has recently begun to ferment, with Gary Marcus, an American cognitive scientist and professor of psychology at New York University, as a prominent figure. Gary Marcus is a long-time critic of Silicon Valley and the connectionist approach to AI. His criticisms of Silicon Valley and AI primarily focus on three aspects. First, he argues that Silicon Valley is excessively pursuing short-term commercial interests, touting AI as a "panacea," while ignoring the technology's true limitations and societal risks. Second, he criticizes the large-model approach for its over-reliance on massive amounts of data and computing power, mocking deep learning as "autocomplete on drugs." The lack of in-depth research into common sense reasoning, causal understanding, and transparency leads to AI often exhibiting a false sense of intelligence in real-world applications, often "talking but not understanding." Third, he points out that Silicon Valley companies lack sufficient ethical responsibility, focusing on capital market hype and narrative packaging while refusing to address the potential safety, bias, and regulatory issues that AI may bring. These are Gary Marcus's long-held views, but after the release of ChatGPT 5, he became particularly active, actively accepting interviews from various media outlets, claiming that the potential of this wave of large-model AI is nearing exhaustion and the bubble is about to burst. So, what do people in Silicon Valley think of the "AI bubble theory"? During my study tour in Silicon Valley, I encountered three different attitudes. The first was dismissive, arguing that the AI ​​bubble doesn't exist and that the future is bright. Some investors expressed strong confidence in the prospects of AI. While they didn't deny that ChatGPT 5 fell short of expectations, they reminded us that the revenues of leading AI companies are growing rapidly, and their valuations are skyrocketing. By next year, we might see unlisted AI companies with valuations of $500 billion or more. The second attitude firmly believes that there is a bubble in AI, and it's a very large one. When I asked about this, a renowned scholar, whose name is withheld, replied that the bubble in Silicon Valley's AI companies today is no less than that of China's "Four Little Dragons" back then. While it's unclear when the bubble will burst, if it does, Silicon Valley will suffer a severe blow. Another emerging Silicon Valley investor has predicted the bursting of the AI ​​bubble and has already begun making corresponding investment moves, preparing to turn a bad situation into a good one. A third view holds that while there is a bubble in AI, it's not a bad thing; AI will "dance with it" and continue its progress. This view is more moderate and represents the majority opinion. In principle, I support the third view, but I do have my own concerns about the current AI development model. The most prominent characteristic of AI, which is based on deep learning, is its opacity. It is the first "unexplainable" black box in human technological history. Today's most advanced AI models contain artificial neural networks with hundreds of layers, exceeding human explanation capabilities. Therefore, no one knows why they are so smart, nor why they are not smart enough. A direct consequence of "unexplainability" is that when AI capabilities fall short of expectations, there's no better solution than to continually pile on computing power. Each time a bit of computing power is added, the results improve. If the results aren't good enough, more computing power is added. No one knows where this technological path will end, or whether it will truly lead to AGI before all available energy is exhausted. We hear all sorts of optimistic and pessimistic views in the social media and at Silicon Valley dinner tables, but these are based on personal beliefs; no one can delve into these hundreds of layers of neural networks to truly understand the truth. The United States announced a 5.4% growth rate in fixed asset investment in August, surpassing China's for the first time in more than 30 years. However, nearly half of this fixed asset investment is directly or indirectly related to AI infrastructure. Vast resources are being devoted to building massive computing centers, while the available data on the internet has long been exhausted. The current model is one of the entire industry dedicated to building ever-grand temples in the hope of summoning the AGI deity. This model is a huge test of people's patience. Is Silicon Valley concerned about this? I think a disagreement has already emerged. If the core capabilities of large models fail to resume their rapid growth, if they fail to quickly produce intelligent advances that are clearly perceived by the public, and if applications continue to expand at the current level, then it won't be long before Wall Street and the American public become enraged by the relentless investment of hundreds of billions of dollars in computing power. Silicon Valley also has its limitations. As the study tour drew to a close, several of my fellow travelers and I reviewed our learnings, and we all felt that Silicon Valley has its own closed-mindedness and limitations. Silicon Valley is indeed passionate about technological innovation, filled with the devotion of a tech pilgrim. However, the flip side of this passion and devotion is its indifference to many major issues in the outside world, especially those facing ordinary people. At a recent tech summit hosted by Trump, Silicon Valley tech leaders, led by Mark Zuckerberg and Tim Cook, unveiled trillions of dollars in tech investment plans, much to Trump's delight. However, some American media have pointed out that these massive plans, often worth hundreds of billions of dollars, primarily focus on data centers, AI chips, and supporting energy. While massive and technologically advanced, they have little impact on improving American employment. Does Silicon Valley care? Actually, not really. Silicon Valley has its own concerns, and they almost exclusively focus on them. They have a self-sufficient indifference to the outside world, a kind of indifference that seems to have nothing to do with them. Silicon Valley startups raise only Silicon Valley money, and the enormous wealth created before going public is almost exclusively distributed within the Valley. People in Silicon Valley don't entertain themselves, their food is average, and their nominally high salaries make life rather dull for the most part. But these indulgences and pleasures aren't important in Silicon Valley. What truly matters is gaining an advantage in the technological race, securing the next round of funding, seeing project valuations rise, improving algorithm performance, expanding infrastructure, and proving themselves truly exceptional in the face of fierce competition. They rarely care whether these efforts can truly solve external problems like education, healthcare, and wealth disparity, or even worsen them. To Silicon Valley, the outside world is an abstract object to be transformed and optimized through their technological wizardry, not a community of billions like them to shoulder and shoulder responsibilities. A school of thought that has become extremely popular in Silicon Valley in recent years, embodied by programmer and renowned blogger Curtis Yarvin, is the Dark Enlightenment. This school of thought, also known as "neo-reactionism," is often categorized as a branch of neo-fascism on the political spectrum. This school of thought opposes democracy and strongly supports Trump, comparing the country to a joint-stock company, governed efficiently by managers elected by "shareholders" rather than relying on popular votes and political party competition. Because Curtis Yarvin uses the pen name "Mencius·Monbag," many in the Chinese community refer to him as "Mencius Yawen." Mencius Yarvin's ideas, fueled by Silicon Valley luminaries like Peter Thiel and accelerated by Trump's election victory, have subtly influenced many within and outside Silicon Valley. Some people we interact with in Silicon Valley may not have read Mencius Yarvin's work or even know him, but in fact, many of their ideas have been shaped by him. For example, they prioritize technology over efficiency, and promote meritocracy. They're unconcerned with the problems plaguing the American people and the world at large, instead focusing on building a "microcosm of technological innovation" in the Bay Area. Their ideal order often emphasizes technological means over political consultation, relying on capital and computing power as sources of legitimacy. Public responsibility is reduced to return on investment, and society is envisioned as a corporation that needs to be managed and optimized. Some have characterized Silicon Valley's arrogant worldview as "techno-aristocracy." Personally, I believe that the long-standing existence of such a self-important, boundlessly creative microcosm in a small region of the world is a source of pride for human civilization. There's no need to expect everyone in the world to be so profoundly righteous. As long as they can continuously innovate, let them be arrogant. However, we should also recognize that not all problems can be solved in Silicon Valley, and there are many things Silicon Valley doesn't care about or concern itself with. There's a view that Silicon Valley hasn't been, and won't be, the center of crypto, because crypto focuses on fairness, while Silicon Valley focuses on efficiency. The fundamental spirit of crypto is equality for all before cryptography and smart contracts, which runs counter to Silicon Valley's technological aristocracy. I'm not sure this view is correct. The crypto industry shouldn't pin its hopes on finding all the answers in one place. However, I believe in Silicon Valley's inclusiveness. While it may not be large, it should have ample room for both AI and crypto.