Some thoughts:

Some thoughts:
1. In this round of the bull market, I see many people guessing the top, but in fact, 4000, 5000, and 6000 are not meaningful; what matters is the real estate market. In the past, every bull market has been accompanied by a surge in housing prices, with funds moving out and being drained. If this wave of the real estate market continues like this, then the bull market will not reach its peak, reshaping a generation's asset faith. Conversely, the opposite is also true, so run quickly, as history is still repeating.
2. This time, Chuan really did something impressive; the EU, Japan and South Korea all surrendered, and funds significantly flowed back to the United States, benefiting the Nasdaq and AI infrastructure. In fact, analyzing various assets requires looking at the direction of the money flow.
3. Anti-involution must be combined with demand-side policies. The past success of the supply side actually also relies on the demand side. For example, the beer industry currently has no involution on the supply side, but it's still not doing well due to the impact of deflation.
4. Will the fertility subsidy also be "layered on" in the future? Having a baby in the future will be a blessing. I have even thought that having children in the future might be like the current technology subsidies, for example, many so-called tech innovation companies are not making any profits, yet the country, the province, and various districts are all providing subsidies, resulting in an overflow of subsidies. If zc can really shift from the supply side to the demand side, then this phenomenon could occur.
5. It's still that saying, the 14th Five-Year Plan determines the direction of 💰, and various asset analyses cannot be separated from this.
In terms of AI:
6. The "underperformance" of GPT5 was actually disclosed five days ago (I even think that OpenAI intentionally released the news early to manage expectations). The real reason behind this is that Silicon Valley has now reached a new consensus: that is, no longer pursuing the groundbreaking capabilities of the model itself, but rather focusing on the practicality of the model (OpenAI now has 700 million users globally, and is no longer a research institution pursuing AGI; instead, they have chosen the opposite route). Now, the judgment of AI in Silicon Valley and Wall Street has shifted to an "economic Turing test," meaning that when an AI completes a task, you cannot tell whether it was a human or a machine behind it; in other words, as long as the AI can effectively improve productivity (regardless of whether it is AGI), the model is considered successful.
7. Following the above, when your user base reaches the scale of 1 billion, the importance of practicality increases significantly because even if production efficiency improves by just a thousandth, the increase in GDP is a frightening number. Therefore, the cutting-edge world model that Google recently released, which allows ordinary people to "mine slots," is not something OpenAI cannot achieve, but rather a strategic choice. Thus, GPT-5's "underwhelming performance" has long been anticipated on Wall Street, and the US stock market's AI hardware has been soaring these past few days.
8. The AI capex spending in the United States is expected to account for 25% of the actual GDP growth in the U.S. by 2025. The infrastructure obsession is well-deserved; I have always said that historically, the United States has been number one in infrastructure, with railway capex historically accounting for 6% of the total GDP (.