One year ago, if you invested 1 RMB and 10,000 RMB respectively into a money market fund (where product data shows a past-year yield of 1.19%), how much would you have earned by now?

You might think this is just a simple elementary school math problem:

• 1 RMB Yield = 1 * 1.19% = 0.0119 RMB (Since the smallest unit of RMB is the fen, meaning up to two decimal places, the third decimal is rounded, resulting in 0.01 RMB).
• 10,000 RMB Yield = 10,000 * 1.19% = 119 RMB.

But the actual yield is: 0 RMB earned from the 1 RMB investment, and 119 RMB earned from the 10,000 RMB investment.

What if you only invested for a single day (assuming a 7-day annualized yield of 1.038% and a yield per 10,000 shares of 0.2804)? The answer is: 0 RMB earned from the 1 RMB investment, and 0.28 RMB earned from the 10,000 RMB investment.

Why is the yield on a 1 RMB money market fund always zero? Where did the yield that should have been there go?

I discovered this counterintuitive anomaly while researching money market funds recently. A 1.24 RMB money market fund investment I made a year ago still shows a yield of exactly zero. Meanwhile, another money market fund where I put over 200 RMB generates 0.01 RMB of yield every day. To figure out the root cause, I carefully read fund prospectuses, consulted fund management companies, and scrutinized the income distribution mechanisms of money market wealth management and fund products.

Almost every money market fund and wealth management product prospectus contains an “Income Distribution” section. However, most of them do not explain the exact mathematical formulas for distribution in detail, including how accounting rounding errors (tail fractions) are handled.

Before diving deeper, let’s go over some basic concepts.

1 Basic Knowledge

According to industry conventions, money market products maintain a constant Net Asset Value (NAV) of 1.00 RMB, with the smallest unit of fund shares being 0.01 shares.

For example, if you invest 200 RMB, your initial holdings are 200 shares. If your single-day earnings are 0.01 RMB, this yield is directly converted into 0.01 shares, bringing your final holdings to 200.01 shares.

2 Income Distribution Steps

1. Valuation and Reconciliation: The fund values, reconciles, and confirms its held assets daily. (The accounting valuation here is complex, so we will skip the details).
2. Calculation and Disclosure: Based on daily calculations of net asset earnings and total outstanding shares, the fund calculates the “yield per 10,000 shares” (万份收益). It also calculates and discloses the “7-day annualized yield” based on the trailing seven days of performance.
3. Daily Distribution: Income is distributed and paid daily, with the current day’s yield reinvested for compounding the next day. The daily allocation is calculated using the following formula: User Yield = Yield per 10,000 Shares * User's Daily Shares / 10000 Since the minimum unit of fund shares is 0.01, the calculated result of the above formula must undergo tail-end rounding (rounding error treatment) at the third decimal place. Common methods include: truncation (floor rounding) or rounding half-up.
4. Redistribution of Tail Fractions: The remaining undistributed balance generated by the rounding process is pooled back into the fund’s residual assets for redistribution. There are various redistribution models, but one of the most common is: sorting accounts from largest to smallest based on the third decimal place of their calculated daily yield (if the values are identical, sorting by account ID in ascending order) and allocating 0.01 RMB to each account until the residual assets are fully distributed.

3 The Siphoned Yield: How Small Accounts are “Vampired”

The vanishing yield of a 1 RMB money market fund is caused entirely by tail-end rounding errors (the discrepancies resulting from decimal rounding limits).

Given the yield per 10,000 shares (0.2804), the theoretical daily return on 1 RMB is merely 0.00002804 RMB. During the tail-end redistribution phase, because this number is so minuscule, the probability of this account ranking high enough to receive an allocated 0.01 RMB is virtually zero.

In other words, the tiny daily yield generated by your 1 RMB is systematically redirected to other accounts with larger fractional remainders. Because returns are calculated and distributed daily, these tiny yields are transferred away every single day, leaving zero opportunity to accumulate or benefit from compounding.

This explains why a 1 RMB money market fund yield remains zero indefinitely. Below, we will dive deeper into the design philosophy of these distribution rules. It gets a bit technical, so feel free to skip if you prefer.

4 Why Is It Designed This Way?

Since the minimum fund share unit is 0.01, the secondary redistribution must use 0.01 RMB as the basic unit to avoid introducing new rounding errors. (The residual assets represent the sum of decimals past the second place across all accounts. If distributed evenly, each account would receive far less than 0.01 RMB, creating another round of decimal errors and leading to a mathematical dead end.)

Of course, other redistribution rules could be designed:

• Sorting accounts by total holdings from largest to smallest;
• Sorting by subscription time from earliest to latest;
• Accumulating the residual assets to the next business day.

The exact sorting method depends on the fund issuer’s internal policy.

The fundamental design principle is a trade-off between technical efficiency and systemic fairness. While maintaining basic fairness, the rules tend to favor larger or earlier accounts. This design benefits large-scale holders at the expense of micro-holders.

4.1 Re-examining the Allocation Algorithm

From an accounting perspective, does the sum of all distributed assets perfectly equal the net asset value of the fund?

In reality, regardless of how decimals are handled, the sum of distributed assets rarely matches the actual net asset value exactly. The discrepancies arise from two factors:

1. Precision of the Yield per 10,000 Shares: The yield per 10,000 shares is typically rounded to 4 decimal places. The 5th decimal place is usually either truncated or rounded half-up. Truncation leads to accumulated undistributed assets, while rounding half-up can lead to over-allocation. The higher the decimal precision, the smaller the error.
2. Double-Calculation Precision Loss: Using the yield per 10,000 shares to compute individual user yields introduces a second round of division and decimal rounding, amplifying the precision error.

To minimize this discrepancy, some fund issuers maintain the yield per 10,000 shares to 10 decimal places in their database systems. For a fund with tens of billions in assets under management (AUM), this keeps the total error within single-digit cents.

A more precise approach would bypass the “yield per 10,000 shares” metric entirely and calculate individual yields directly: User Yield = User's Shares * Daily Net Earnings / Total Shares Outstanding This method requires only a single division, minimizing rounding-error propagation.

4.2 Would a Variable NAV (with Constant Shares) Solve This?

No. Even with a variable NAV structure, decimal rounding issues still arise when calculating individual returns. The root cause of the “siphoning effect” is the daily mandatory distribution and payout, which prevents micro-yields from accumulating locally in your account.

4.3 Do Other Types of Funds Face the Same Issue?

Other mutual funds (such as equity, bond, and hybrid funds) use a fixed share, variable NAV structure.

While they publish their NAV daily, they do not distribute income on a daily basis. Only at the time of subscription and redemption, will share settlement (subscription) and income settlement (Redemption) be carried out according to the net value, resulting in errors only at the “transactions”, so only two interception errors will be generated in the whole holding cycle.

In short, “precision processing only occurs during transactions.” These funds do not have the daily mandatory tail-end redistribution mechanism that drains micro-yields in money market funds.

Thus, if you hold 1 RMB in an equity or bond fund, even if your daily return is too small to display on the screen, it does not vanish. As long as the fund’s NAV grows, your yield continues to accumulate locally and will eventually be paid out when you redeem.

Now that you understand these distribution rules, are you already spotting a clever way to arbitrage the system? 😉

Yesterday morning, Beijing time on June 12 (before SpaceX has been IPO), I opened YouTube and noticed a SpaceX livestream recommended on my homepage. Upon clicking in, I realized it was actually a highly realistic, well-disguised scam trap.

The stream claimed that Elon Musk was announcing a SpaceX IPO via cryptocurrency, inviting viewers to scan a QR code in the video to participate in the launch. Driven by curiosity to see what was actually going on, I scanned the QR code. It turned out the website was not for buying SpaceX stock with crypto at all, but rather a fake “SpaceX IPO Cryptocurrency Giveaway” event.

The domain name was highly deceptive: spacexholders.com (though after completing a verification step, it redirected to spacex-holders.netlify.app).

I did a quick search on Google, X (formerly Twitter), and V2EX for “buying SpaceX stock with cryptocurrency,” but found absolutely no mention of this website. Next, I ran the URL through Google AI, which immediately flagged it as a scam. A search for “SpaceX giveaway scam” turned up numerous articles detailing similar YouTube livestream scams. At this point, I was absolutely certain it was a fraudulent stream.

Looking closer, the YouTube channel was named JPMOgan.Live (ID: JPMorgan.Live-spacex), using the JPMorgan Chase logo as its avatar, yet the channel’s video history consisted of nothing but personal music uploads. There were thousands, if not tens of thousands, of active viewers, and the stream had been running for seven hours with the live chat disabled.

So I submitted a report. Within a few minutes, the channel was taken down. At the same time, I reported the phishing website to Netlify.

Around 9:00 PM, I opened YouTube again, and to my surprise, another “SpaceX IPO” livestream was recommended on my homepage. The channel was again named JPMOgan, using the JPMorgan logo, but with the ID spacex.JPMorgan. Upon clicking in, I saw the same disabled chat. Unbelievably, over 30,000 people were watching, and it had already been live for an hour. Looking through the channel’s upload history, the most recent video was from eight days ago, with content entirely in Arabic. Clearly, another hijacked channel. I immediately reported both the livestream and the channel itself.

I have to call out YouTube’s response time here—it was incredibly slow. 32 minutes after my first report, the stream was still up. I submitted another report, and it took another 20 minutes before the channel was finally banned.

Another half hour later, I received an email confirmation from Netlify stating that the phishing website had been taken down. It took nearly ten hours from my initial report to the ban, likely because their support team only started processing requests once their business hours began.

A quick search revealed that similar scams pop up every time SpaceX hosts a major live event. Articles on this date back as far as June 12, 2020, and there are multiple Reddit threads discussing these exact scams, with some users in the comments sharing that they had indeed fallen victim to them.

Below are some of the related reports and discussion links:

• Crypto Influencer Elon Musk Targeted Again! Hackers Impersonate SpaceX Channel to Steal $1.5M in Two Days
• Three YouTube Channels Hijacked for Bitcoin Scams Under the Guise of SpaceX
• Just Some Of The Crypto Scam Money Made From “SpaceX” Livestreams
• SpaceX - Musk Cryptocurrency Scam on YouTube

Meanwhile, there are already two victim reports on Chainabuse.

1 Post-Incident Reflections

The Abuse of AI Technology: These streams utilize AI-generated Deepfake videos. Today’s AI voice cloning technology is highly sophisticated; scammers clone Elon Musk’s voice and overlay it onto past event footage. Without careful scrutiny, it is incredibly difficult to realize that you are watching a Deepfake video rather than a genuine live broadcast.

Sloppy Web Development: The phishing website looked like it was thrown together quickly via “Vibe Coding.” The design was instantly recognizable as AI-generated and had glaring flaws. For instance, scanning the QR code took users to a CAPTCHA verification page, which redirected to a Netlify subdomain after verification. This sudden domain change exposed the fact that the project was hosted on Netlify. Furthermore, visiting that Netlify domain directly bypassed the verification entirely.

Classic Scam Tactics: At its core, it relies on the same old formula: “send a coin, get double back for free.” How could such a deal exist? Not to mention that blockchain transactions are irreversible; once you transfer your funds, there is no such thing as a refund. This setup reminds me of the scam ads printed on the back pages of magazines like Reader or Yilin during my childhood: “Solve a puzzle to win a prize” (they would ask a ridiculously simple question like 1+1=?, and tell you to mail in your answer along with a 10-cent coin, promising a 50 RMB reward if you got it right).

Human greed remains unchanged through the ages.

I recently encountered an issue using a coupon on Taobao Flash Sale. It took six rounds of back-and-forth communication with customer service to finally resolve it. Through this after-sales experience, I want to discuss the systemic issues revealed by Taobao Flash Sale’s resolution process.

1 The Origin: A Non-Functional Coupon

One day, I noticed a “2 RMB Flash Sale Payment Coupon” available for 99 points in the Alipay Points Center. Thinking I needed groceries, I redeemed it, planning to place an order at the “Hema Chao He Suan NB” store the next day (Thursday).

To my surprise, the coupon failed to apply after I placed the order. I first contacted Alipay customer service, who directed me to Taobao Flash Sale customer service.

After I provided feedback to the online agent, a specialist called me back. They asked me to first reproduce the issue and then attempt to use the coupon across different channels (App, Alipay Mini-program, Taobao) and various stores. The final conclusion was that the coupon only worked within the Alipay Mini-program, but specifically excluded Hema stores.

Consequently, I requested a refund of the 2 RMB difference I had to pay out of pocket because the coupon failed.

2 Rights Protection: From Refusal to a 12315 Complaint

During the second call from the specialist, they refused the refund, arguing that since the coupon worked in other stores, it was not technically “invalid” even if it failed at Hema.

I was furious. The Hema store page clearly indicated the coupon was applicable, yet it failed during actual payment. This was a blatant platform bug that resulted in a loss of consumer benefits, yet they still refused a refund.

After hanging up, I collected my evidence and filed a complaint on the 12315 platform (China’s national consumer rights system). Shortly after, I received a text message to rate the service, implying my support ticket had been closed. I contacted the online agent again to ask why the ticket was closed and provided a screenshot showing that the Hema page supported the coupon. However, the new agent was completely unaware of my case history, forcing me to describe the entire problem and my demands all over again.

On the third call with a specialist, I stood my ground. In addition to the screenshots, I pointed out that the coupon had already been removed from the Alipay Member Center (it had a redemption limit of about 400; its removal likely indicated a known rule flaw). Only then did the agent admit that the issue might be related to Hema and required internal confirmation.

The following day (Saturday), I received a fourth call. The agent provided a progress update, stating they were expediting the matter.

3 The Resolution: Refund

On Monday, a representative specifically handling 12315 complaints called to re-confirm the coupon’s effective range and asked for additional information. Surprisingly, this information was already in the PDF documentation I had uploaded to the 12315 platform. It turned out their internal system did not support viewing PDFs from the 12315 platform, so they sent me a link to re-upload. It was clear that their internal work order platform was unable to capture or display detailed external case data.

Eventually, the agent called back with a conclusion: the platform issue had been reported to developers for a fix, and 2 RMB was refunded to my “Taobao Flash Sale Wallet.”

4 Retrospective: A Microscopic View of Big Company Issues

As an internet industry professional, I identified several underlying problems through these details:

4.1 1. Operational Incident

This was likely an operational mishap. Although the scale was small (around 400 redemptions), it reflects a failure in internal coordination within the Alibaba ecosystem. This was a joint marketing activity between Alipay and Taobao Flash Sale, yet the coupon failed within 48 hours of launch. Furthermore, I learned from the agent that Hema stores within Taobao Flash Sale use a “customized” setup.

4.2 2. Root Cause Analysis

Taobao Flash Sale partnered with Alipay for marketing by listing the coupon in the Alipay Member Center—a win-win intended to increase Alipay’s market share and Taobao Flash Sale’s order volume. However, because Hema stores are customized with unique payment logic, it caused a code incompatibility: the front-end displayed support, but the payment gateway did not. It is likely that Hema’s code was not synchronized when the new coupon type was introduced.

4.3 3. Information Silos in Customer Service

The level of information sharing between different agents is remarkably low. Every time a new agent took over, I had to restate the problem. The internal work order system either lacked detailed records, failed to log phone calls, or the agents lacked the authorization to view previous interactions. This is a scenario where AI could be utilized for call analysis and summarization to sync with work orders. However, management likely prioritizes ROI (Return on Investment), choosing to sacrifice user experience to balance efficiency and cost.

4.4 4. Lack of a Professional Coupon Management Backend

With a vast variety of coupons, agents cannot quickly determine a coupon’s rules based on a user’s account. Instead, they rely on users to perform “manual testing.” This suggests that their coupon operation backend lacks professional diagnostic tools and intuitive rule-validation features.

4.5 5. Subpar Service Standards

Compared to other major internet platforms, Taobao Flash Sale’s efficiency is notably lower. When I encountered similar coupon issues on Didi, the agent immediately issued a compensatory coupon. When a JD Laundry appointment couldn’t be scheduled, even though they took a week to call back, they offered a 10 RMB compensation. In contrast, Taobao Flash Sale’s initial response to an obvious platform bug was to deflect responsibility.

4.6 6. The “Departmental Walls” of Big Corporations

The cost of cross-departmental coordination in large companies is massive, even in the high-efficiency internet sector. One ticket can involve operations, product, customer service, and tech. Furthermore, inter-departmental interests can conflict—similar to how “Yuanbao” links are blocked by WeChat, or how Alipay directs traffic to ByteDance, Meituan, and JD, while JD Finance directs traffic to ByteDance, Baidu, and Alibaba.

Finally, this was my first time hearing of the “Taobao Flash Sale Wallet.” Why does an Alibaba entity need its own wallet when they already have Alipay? Research shows it is primarily used for “dedicated funds” for top-up discounts and offers loan services in partnership with Bank of Ningbo. Essentially, it is an attempt to create a closed-loop financial system.

5 Conclusion

A small clue can reveal a much larger picture. While my experience might be an isolated case and may not represent the entirety of Taobao Flash Sale’s service, it highlights significant room for improvement in professional service, cross-departmental coordination, and underlying technical unity.

With Gold Soaring and Plunging, Is It Still a Good Investment? Sharing My Understanding of Gold.

Gold has recently seen violent fluctuations—skyrocketing and then plummeting. Many are asking: Is it still safe to invest? Here is my take on gold investment.

The core purpose of investing in gold is to improve portfolio diversity, hedge against inflation, offset risks, and seek safety. It is crucial to be clear on one thing: Gold is for long-term asset allocation, not for short-term speculative riches.

Ray Dalio, founder of Bridgewater Associates, suggests that gold should make up 5% to 15% of an investment portfolio. particularly in the current context of ballooning global debt, he advises leaning towards the higher end, allocating 10% to 15%.

Gold is a medium for storing wealth. It cannot be replicated (unlike paper currency which can be printed infinitely), and its confiscation risk is very low—it is difficult for governments or banks to freeze it or artificially devalue it. However, gold itself does not generate value, nor does it yield interest. Its long-term returns are correlated with the money supply (M2), with a long-term real annualized compound return rate of approximately 2.0% to 3.6%.

Of course, gold prices will not rise forever; history shows periods of prolonged decline. Regarding a drop in the price of gold within my portfolio, my mindset is this: Allocating gold is like buying insurance for my investments. If it falls, I treat it as paying the insurance premium for renewal; if it keeps rising, it’s like receiving an insurance payout.

1 The Forms of Gold Investment

There are several ways to invest in gold, each with its own pros and cons:

1. Gold Accumulation Plans (Paper Gold) Suitable for those who want long-term holdings but don’t want to store it themselves. This is equivalent to lending your gold to the bank; the bank keeps it for free and may even pay you some interest. However, there is a critical prerequisite: The bank must keep its promise. There is a liquidity risk (the risk of being unable to redeem or sell), especially during violent market volatility when banks may restrict trading. Essentially, this is lending gold to a bank in exchange for interest.

2. Physical Gold This is true “hard currency.” In times of turmoil (war, economic collapse, collapse of national credit), physical gold can help you survive. The risks are that during chaotic times, the state may prohibit the possession of gold (refer to the Gold Yuan era in the Republic of China), and there is the risk of theft/robbery. Therefore, for self-custody, I suggest keeping gold that is small, easy to hide, or melted down into small nuggets or dust. Because it resists inflation and truly stores wealth, it is suitable for long-term holding.

3. Gold ETFs / Funds These offer good liquidity and are suitable for medium-to-short-term investment, primarily to hedge against “Black Swans,” war, and disasters. Since these are financial products, in extreme scenarios, they carry systemic risk (trading at a discount or premium), liquidity risk, and hidden costs (value decay, management fees, custodian fees, transaction fees). The worst-case scenario involves credit risk, where the fund is liquidated, becoming mere “paper wealth.”

4. Gold Futures High liquidity, flexible, and allows for two-way trading (long or short). Because leverage is involved, capital efficiency is amplified, but the risk is extreme. Due to the costs of rolling over positions (wear and slippage), it is not suitable for long-term investment.

2 My Gold Investment Strategy

I primarily choose Gold ETFs and Physical Gold as my investment targets.

• Gold ETFs: Convenient trading and low fees make them suitable as part of my asset allocation.
• Physical Gold: This is for “Gold in Chaotic Times”—to cope with future extremes (war, unrest, economic crisis). It serves as a survival chip and a hedge against inflation in peacetime. I plan to buy at a suitable price (I still find the current price of over 1000 RMB/g a bit expensive), prioritizing small, concealable, and portable formats.

I dynamically adjust my allocation ratio, keeping it between roughly 5% and 15%. When market risk is high, I increase the proportion of gold. My current target is 10%, with a strategy of adding to my position when prices dip.

I don’t think high gold prices are the problem. The real issue is: The risk of not allocating gold is greater than the risk of gold prices being high. Therefore, regardless of the price, I will maintain an allocation, though naturally, buying at a low point is best.

Finally, I want to say that assets are not equal to money. Until an asset is converted into real money, it is just paper wealth. In times of turmoil, all financial products and paper currency can become worthless. Only physical goods (gold, food, water, energy, materials, etc.) are hard currency, and among these, gold is more suitable than any other physical commodity to serve as a medium of exchange.

3 Reference

Gold as a strategic asset: 2026 edition

Gold Market Primer: Market size and structure

gold return attribution model

Historical demand and supply

Gold return

Gold prices

Investment Update - Beyond CPI: Gold as a strategic inflation hedge

The impact of inflation and deflation on the case for gold report

Should You Hold Gold In 2026?

Found some good news today: The threshold for becoming a “Traffic Master” (monetization) on WeChat Official Accounts has been lowered.

As long as your follower count is 100 or more, you can now activate most advertising slots. This includes “Comment Section Ads,” “Bottom Ads,” “In-Article Ads,” “In-Article Keyword Ads,” “Post-Video Ads,” and “Commission-based Product Promotions.”

For long-tail, small subscription accounts like ours, this is absolutely a win-win situation. Official Account authors get incentives to continue creating, while the platform gains more ad revenue and more content.

(Note: The threshold for “Mutual Selection Advertising” has not been lowered; it remains at 500 followers. The Traffic Master backend overview shows it requires >= 500 followers, but the activation page displays a threshold of >= 100. This is likely a copywriting bug.)

1 Why Lower the Threshold?

As a creator of a small subscription account (low follower count), let me discuss a few phenomena I’ve observed:

1. In 2025, acquiring traffic for subscription accounts has become increasingly difficult.

Although algorithmic recommendations can bring traffic to small accounts, the general feeling in 2025 is that recommended traffic is shrinking. It is obvious that the recommendation strategy has undergone a major overhaul, now placing greater emphasis on account weight (follower count, account activity, and output quality).

At the same time, the use of “Content Boost” (search/recommendation promotion) has become stricter. In 2024, my account could use boost coupons, but in 2025, I am surprisingly blocked from using them, with a prompt stating, “The current account cannot use the boost function temporarily.” There is no issue with the quality of my articles (all handmade originals), and the WeChat Open Community is full of similar cases. It can be inferred that restrictions are being applied based on account weight.

In 2025, I haven’t felt the platform’s traffic support for small subscription accounts; in fact, the support seems to have diminished. It is also possible that the articles I post simply don’t suit the taste of the recommendation algorithm.

2. The recommendation algorithm and account weighting are a mystery.

The backend no longer displays data for “Recommendation Impressions” or “Click-Through Rate,” only “Recommended Reads” (clicks). Creators can no longer reverse-engineer the algorithm’s preferences through data. This may be intentional on the part of the Official Account platform to prevent writing specifically to pander to the algorithm. This might be a good thing, forcing everyone to return to the content itself.

3. Growing follower counts is getting harder.

Even with a viral article getting 10,000 reads, it might only bring in about 100 new followers. With the recommendation feed, users don’t care if the content in their stream is from an account they follow or a recommendation—as long as they are interested, they will read it even without following. This lowers the impulse to follow.

Users will only generate a desire to follow if they truly recognize the value of the subscription account’s content and want to see more of its articles.

4. The Matthew Effect is intensifying.

Even after a user follows, if they don’t “Star” (mark as important) the subscription account, it is very difficult for the account to appear in their subscription feed. The lower the read rate from existing fans, the less traffic the article gets for diffusion, making it even more unlikely to generate new followers. This forms a dead loop: small accounts cannot grow, leaving only the big accounts, turning the Official Account platform into stagnant water (which could likely lead to decline and extinction).

In the era of short video, Official Account traffic is shrinking, gradually being overtaken by short video platforms. Facing the shockwave of short video, Official Accounts need fresh blood.

Lowering the threshold is intended to let small accounts earn a little income so they can continue updating. Only then will the platform have a continuous source of high-quality text-and-image content. The platform needs more quality content to retain traffic and maintain user stickiness.

At the same time, I hope the Official Account recommendation algorithm can lower the weight of the account and give higher weight to the quality of the article. Only then can small and new accounts grow, and only with fresh blood can the platform flourish. Simply put, it needs a decentralized, equal-rights recommendation algorithm like Xiaohongshu (Little Red Book).

The introduction of “paid boosting” also means that without payment, there is a high probability of not being recommended, which conflicts to some extent with the free “content boosting” in account growth. How to balance paid promotion and content-based recommendation (free of charge), that is, to balance commercial income, user experience, and support for small accounts to maintain a healthy ecosystem, is a consideration for WeChat official accounts.

2 Final Thoughts

2.1 Why do I insist on writing text-and-image content?

The information density and acquisition efficiency of reading text far exceed that of video (whether short or long).

At the same time, text is much friendlier for information retrieval/search, which is the advantage of Official Accounts. For knowledge acquisition, my personal habit is to prioritize text and images; I will only watch a video if I need to see a demonstration of an operation.

Posting on my Official Account satisfies my desire for expression, stemming from my habit of blogging. The Official Account is just something I do on the side—a byproduct.

2.2 Regarding Income

I don’t expect the Traffic Master program to bring me substantial income. My attitude is: If there is money, it’s the icing on the cake; if not, I earn influence. The easiest place to make money by posting text is the Ant Community, where a casual 50-word post or a post with 20k views can earn over 100 RMB.

I have a bottom line: Reading experience comes first.

Even if I were destitute and down and out, needing to earn these few scraps of silver, I would prioritize placing a donation link on my blog rather than ads.

For my Official Account articles, I will not enable “In-Article Ads” or “Comment Section Ads” or other forms that interfere with reading. I want to leave you with a clean reading space.

This marks my second year of surviving independently. If I were to summarize it in one sentence: A glimmer of light has emerged through the mist, though the path ahead remains hazy.

2025 was a year of diverse experiments, yet I have not quite found a direction that generates a self-sustaining positive feedback loop. However, my investment portfolio—a field I hadn’t focused on for a long time—yielded significant returns. My total asset growth rate for the year was 7.6%, with a portfolio return of 7.8%. It’s fair to say that passive income provided the “glimmer of light” this year, giving me the financial breathing room to explore new directions while my living expenses were covered.

1 2025 in Retrospect

At the beginning of the year, I explored the potential of AI-driven independent development and the “One-Person Company” model. Using AI as a coding assistant, I ventured into the unfamiliar territory of front-end development to build a tool site, monetizing it through Google AdSense. Traffic was minimal; it served primarily as a “warm-up” project.

I then delved into the AI inference market and researched startups in the field, looking for opportunities in AI infrastructure. I attempted to build a comprehensive product, managing everything from development and DevOps to product operations. This pushed me far beyond my existing capabilities. Everything was new, requiring constant learning and heavy reliance on AI assistance. The pressure was immense—I worked over 14 hours a day, and the mental stimulation was so intense that I suffered from severe insomnia, often staying awake until dawn. Ultimately, due to high operational costs and a high probability of financial loss, the project was shelved after development was complete.

By June, I began to slow down and refocus on my investment income. Noticing a recovery in the Chinese A-share market, I re-engaged with my portfolio and adjusted my holdings.

In late August, I moved to Hangzhou to lower my cost of living.

I spent the second half of the year deepening my knowledge of investing and spending my spare time consuming books and videos on entrepreneurship and economics.

2 The Shift: From Engineer to Entrepreneur

I feel like a leaf that has fallen from its branch to the earth—taking root and sprouting, weathering the storms in the hope of one day becoming a towering tree that stands tall against the elements. From a purely wealth-accumulation standpoint, this was not the “optimal” choice. On the “tree” (the corporate platform), wealth accumulates steadily; the platform shields you and amplifies your skills. On the ground, you are self-reliant, and the world is full of uncertainty. However, I feel healthier this year. My complexion has improved, and the stress-induced oiliness of my skin has faded. Work is for a better life; life is the foundation.

My greatest transformation has been moving away from the “engineer’s mindset” of obsessing over technical details. Instead, I now focus on user needs, business opportunities, and commercial viability. Technology is no longer the end goal for me. I fell into the classic trap of many technical founders: holding a hammer and looking for a nail, rather than starting from a real user need. Reading The E-Myth Revisited deepened this realization. I must transition from an engineer to a business owner—identifying scenarios, demands, and markets, rather than just “finding a job” to do for myself. Validation, business models, and operations are the real keys; technology is secondary.

3 On AI and the Future

AI represents a “lane-change” opportunity for ordinary people. It is like a new gate opening at a crowded station; those at the back of the line can suddenly surge ahead through the new path. One only needs to catch one or two such waves in a lifetime. Finding my niche within the AI ecosystem is my next goal. However, current Large Language Models (LLMs) are certainly not the final form of AI, and the “chat” interface has limited use cases. According to the Tech Hype Cycle, AI is currently in the “Peak of Inflated Expectations.” When this bubble will burst remains to be seen.

4 Investment Strategy for 2026

Investment returns are inherently uncertain, and the greatest enemy of the “snowball” (compounding) is volatility. My priority is to lower volatility, minimize drawdowns, and improve the quality of my returns. Before I invest, I ask myself: Whose money am I making? If it is a zero-sum game of “swapping pockets,” I stay away. No one wants to be the last one holding the bag. Given that A-shares are currently at a historical high where retail investors struggle to exit in time, my 2026 strategy is to reduce my exposure to A-shares. I will not add positions at these levels; instead, I will sell into the rally and reallocate to undervalued sectors or remain in cash. The core strategy is subtraction—reducing the number of holdings.

I prefer a portfolio that is simple, low-cost, long-term, and low-volatility. I spend 20% of my time on investment research to outperform 80% of people, and 80% of my time figuring out how to monetize the decade of experience I’ve accumulated. I have no desire to simply “lie flat.” I still hope to do something that impacts the world—something that might be remembered. But for now, the priority is to establish a stable, continuous cash flow before chasing higher ideals.

Reduce volatility, increase returns. Sharpen the tools, and wait for the opportunity.

I went to the Hema NB (Super Value) store near my home today. Business was booming; I even had to queue to check out.

In my neighborhood, I have access to Hema, Dingdong Maicai, and Yonghui. Previously, I always ordered Hema Super Value items via the “Taobao Instant Shopping” (Taobao Shangou) interface because of the coupons. However, the delivery was incredibly slow, usually taking over an hour (likely due to a shortage of delivery riders in my area). In contrast, Dingdong Maicai delivers much faster (typically 30 to 60 minutes), but I found that Dingdong’s prices on the Taobao interface are surprisingly higher than on their own app.

Hema NB (Super Value) gives me the impression of a fresh food supermarket that has been stripped of live seafood. It feels like a watered-down version of the classic Hema, cutting out the signature live seafood and in-store cooking services. It seems to be a continuation of the “Hema NB Pickup” model (now that the self-pickup business has been taken offline). It is cheap and offers high cost-performance, though the quality of some items is just average. In terms of quality, I’d rank them: Hema Fresh (Xiansheng) > Dingdong > Hema NB.

What’s confusing is that for Hema NB, which is supposed to be a key developmental business for Hema, the main digital entry point is a WeChat Mini Program (there isn’t even an Alipay Mini Program). Furthermore, it has no relation to the main Hema App (though you can pay with Hema’s payment code and check orders), and it has an independent membership points system. You can place orders via Taobao Instant Shopping, but those orders don’t show up in the main Hema App. This means that customer service for Taobao orders can only be handled through Taobao—the systems are independent and not fully integrated.

I completely don’t get it. Why choose the Tencent (WeChat) ecosystem as the entry point for Hema NB while abandoning their own app? And why isn’t it fully integrated with Taobao Instant Shopping?

Also, regarding Alibaba’s e-commerce integration (renaming Ele.me to Taobao Instant Shopping and merging it into the E-commerce Group), why is Hema left outside to operate independently? Perhaps it’s due to the specific nature of the business—after all, it’s physical retail involving warehousing, logistics, brick-and-mortar stores, inventory, and procurement.

Hema, to me, can be described in one word: “Chaotic.” There are too many formats: Hema Fresh, Hema Mini, Hema Outlet, Hema X Member Store, Hema Market, Hema Neighborhood, Hema NB, Super Value NB, Hema Pick’n Go, Hema Cloud, etc. (There are so many, I can’t even remember them all. I just want to buy groceries; why create so many types?). They are constantly using trial and error, launching new businesses, and often contradicting themselves—taking the X Membership system offline then back online, shifting focus from online to offline, closing Hema Neighborhood/NB Pickup (because they couldn’t beat Pinduoduo’s Duoduo Maicai), and closing X Member Stores. A company like this, with such a fast pace of launching and failing, must place huge pressure on its employees (anyone who has worked in big tech probably understands this well).

The Hema X Membership is the worst value membership I have ever purchased. First, it’s expensive. At 258 RMB, compared to JD Plus at 99 RMB or Dingdong at 108 RMB, it’s truly pricey. Second, the benefits are mediocre. Aside from free delivery, the rest isn’t great. The membership doesn’t waive shopping bag fees, the number of items with “member pricing” is scarce, and the free vegetable perk was changed from “online or offline” to “offline only.” The X Membership is even split into Gold and Diamond versions. There was also the “Yi Shan” (Moving Mountain) price war, where shelf prices were cheaper than member prices. The system was taken offline and only relaunched in 2024.

Is it time to buy Alibaba stock?

I generally don’t use Taobao anymore. My first choice is JD.com (self-operated), and occasionally Pinduoduo. I used to buy from Tmall Supermarket, but I found their packaging consisted of one giant plastic bag with the Tmall logo, which caused the items inside to get crushed. I switched to JD Supermarket because they choose packaging (boxes or bags) based on how fragile the items are. As for delivery speed, in big cities, both offer next-day delivery, and JD can often do same-day delivery.

regarding the “Delivery War,” current market share is: Meituan > Taobao Instant Shopping > JD (according to a recent JP Morgan report). However, I think Meituan and Alibaba will be fighting neck-and-neck for a while. The classic internet strategy is “fatten the pig then kill it” (gain a dominant market share/monopoly, then raise prices). It remains to be seen if the market share bought with subsidies can be sustained once the subsidies stop.

According to financial reports, Meituan’s food delivery business lost 14.1 billion, while Alibaba’s delivery business loss figures are unknown, though the Cloud Intelligence Group saw revenue growth. Alibaba is betting everything on AI—AI Intelligent Cloud, the Qwen (Qianwen) Large Model, and applying AI across various businesses. However, as infrastructure spending on AI increases, the Return on Equity (ROE) of these investments remains to be seen.

Although I am not the one making the waves, I can ride the wind. In this ever-changing era, whether in consumption or investment, maintaining independent thinking is the most important thing.

Recently, I watched the Silicon Valley 101 interview with Tian Yuandong, former Research Director of FAIR at Meta. This interview provided me with a frontline research expert’s perspective on the current development of AI and Large Language Models (LLMs). It was professional, authentic, and without agenda or bias, from which I benefited greatly.

After watching, I summarized several insights:

• Fewer people will engage in fundamental research, while more will focus on applications.
• The LLM path may not be universally correct, and the scaling law will eventually reach its limits.
• Balancing technology and business, bridging the gap between academia and engineering.
• Choosing work that is both enjoyable and has future market value is crucial.

1 Fewer People Will Engage in Fundamental Research, While More Will Focus on Applications

As model engineering technologies mature, fewer people will be engaged in fundamental model research. This is a common pattern in the technology field: a small number of people work on foundational technologies (such as Linux kernel development, cloud-native infrastructure, or basic framework development), while the vast majority focus on integrating these technologies with business needs for practical implementation.

2 The LLM Path May Not Be Universally Correct, and the Scaling Law Will Eventually Reach Its Limits

Current LLMs have parameter scales in the billions. A human can acquire at most 10 billion tokens of knowledge in a lifetime, yet models are now trained on tens of trillions or even thirty trillion tokens. Where does this high-quality data come from? Whether models trained on such a massive number of tokens can truly achieve human-level capabilities remains a big question mark.

From a business perspective, large-scale data and parameters imply enormous computational demands, and there’s ultimately an upper limit to computing power and energy. The scaling law indicates that exponential input yields only linear growth, making it a very inefficient endeavor in terms of input-output ratio. Whether the value generated by models produced at such high costs can exceed the investment is also questionable.

Crucially, current models lack human “insight.” They struggle to offer unique perspectives or innovative ideas when facing problems. As Tian Yuandong put it:

“This kind of high-level human insights, human knowledge, and unique perspectives on a problem, these are things that current models are lacking.”

3 Balancing Technology and Business, Bridging the Gap Between Academia and Engineering

For companies like Meta, every technological investment aims to generate business value. Meta’s relative lag in AI can be partly attributed to rushing commercialization when the technology was still immature, and instances of “outsiders directing insiders.” The cycle from academic research to engineering implementation is often long; patience is key.

4 Carefully Choosing Work That Is Both Enjoyable and Has Market Value

It’s important to clearly think about what you truly enjoy, what you can do, and what creates value. Blindly chasing hot industries is not advisable; what’s scorching hot today might be yesterday’s news tomorrow, like iOS development during the mobile internet boom.

This is my biggest realization recently: as a tech professional, if you want to make a living through technology, that technology must generate commercial value. Of course, one can also pursue technology driven by interest and passion, regardless of its market value. Ultimately, the goal of work is to live a better life.

When human beings’ seven emotions and six desires are satisfied, they can fall into the brink of nihilism. Those who have experienced life’s hardships, great ups and downs (experiencing life’s peak and trough, then returning to the starting point), will also fall into nihilism. Nihilism – a boundary incomprehensible to three-dimensional human beings. After experiencing nihilism, one often achieves self-consistency, lives with clarity, or as the saying goes, “knows the will of Heaven.”

Why do people live? What is the meaning of life? Where does consciousness come from? These are all unanswered questions for humanity.

Humans can only find their meaning through religion (Buddhism, Taoism, Christianity, etc.), or by finding their value in society.

Some people are lucky enough to quickly encounter this boundary, such as achieving financial freedom at a young age. Others must experience the sweet and sour of life, go through worldly vicissitudes (birth, old age, sickness, death, marriage, having children) to touch this boundary. Some live their whole lives and never touch it.

In high school, my deskmate quietly told me he had cancer, then turned and threw a question at me, “Tell me, what is the meaning of life?” This heavy, soul-piercing question made me feel as if time had frozen. As a high school student, I had never pondered this question; living and studying seemed so natural. That single question was like being awakened from a deep sleep.

My answer now is that life inherently has no meaning; we merely find a meaning for our own lives, whether it’s by finding our value in society (experiencing an ordinary life) or by seeking our meaning through religious faith.

Humans are three-dimensional beings, unable to comprehend higher-dimensional logic. Live in the present.

This year, I once again opted for a “no-job” mode, but to my delight, my personal assets actually saw a net increase.

So, where did this income come from? It wasn’t from my strong technical field. The two new projects I attempted this year both fell short of expectations and didn’t bring the anticipated breakthroughs. What truly made a difference was the investment logic of “money making money.”

This was somewhat unexpected, especially given the overall sluggish economy. Yet, the stock market, like fish swimming upstream, displayed unusual activity and enthusiasm. The market trend was consistently positive, as if “riding the tide.” And I, as someone not directly involved in creating social value, also benefited from it.

My fund holdings from five years ago have been “lying flat” ever since. While they haven’t fully recovered their initial value, their performance has easily surpassed the CSI 300 index, which undoubtedly affirms the active fund managers’ ability to seize market opportunities. Meanwhile, the new investments added at the end of last year have now entered a considerable profit range.

After five years of navigating bull and bear markets, I’ve seen significant growth in both my investment knowledge and mindset. Recently, I conducted a thorough review of my asset allocation and identified some areas for optimization: the proportion of underlying defensive assets is too high, while risk-averse assets like gold and overseas assets are relatively under-allocated.

I am now implementing a “barbell strategy,” aiming to balance both defense and offense. The specific plan is to increase holdings in fixed-income+ products, industry ETFs, and Nasdaq index funds, and to opportunistically allocate to gold and dividend-paying assets. In simple terms, this means moderately reducing wealth management products and increasing allocations to funds and bonds. When gold prices experience a reasonable pullback, I will consider a small additional allocation. However, since some funds have not yet matured, the current adjustments will continue for some time. I believe that after this optimization, my investment portfolio is expected to achieve its target annualized return.

Since investments can now bring considerable returns, I’ve decided to devote more energy to them. My leisure time will be spent reading entrepreneurship-related books, watching YC’s startup sharing videos, browsing platforms like Reddit and V2EX, and continuously updating my blog and Wechat official accounts. I hope to gradually shift from a technical expert’s mindset to a more holistic entrepreneur’s perspective. Of course, I will maintain my focus on cutting-edge technology and trend sensitivity, while also gaining deeper insights into market demand and potential opportunities.

My current state is somewhat akin to “Financial Independence, Retire Early” (FIRE), but I still yearn to achieve something significant. The lifestyle I envision is: a passionate career as the main thread, supplemented by stable investments as a side thread. Just as “a blessing in disguise,” my current smooth sailing might just be a result of aligning with the times. However, strictly adhering to investment discipline and overcoming inherent human weaknesses remains a formidable challenge. Therefore, I will continue to hold a respectful attitude towards the market and approach every investment decision with prudence.