Latest posts
-
I Built This Blog in 4 Days Using AI — And Now I’m Running a Live Experiment in Public (Follow Along)
This post is pinned and updated continuously. Think of it as a live reality show — except the budget is ¥700 (~$100) and everything is real. Come back anytime for the latest episode. A Declaration Before We Begin My child is in high school right now. Graduation is roughly five years away. I have five…
-
From 30MB of Word Files to a 400KB App: How AI Changed My Son’s English Study
China’s college entrance exam — the Gaokao — is taken by 12 million students every year. One exam. One day. Every major life decision that follows. My son is one of those 12 million. Last month I found him at his desk at 11pm, notebook open, writing the letter C in a column of 20…
-
The SEO Strategy Most AI Tool Builders Ignore: GitHub Awesome Lists
I discovered GitHub Awesome Lists as an SEO strategy. Here is what actually happened after 42 pull requests: 2 merged, 9 rejected (5 by the same person), 31 still pending.
-
AI Didn’t Replace My Voice. It Finally Let the World Hear It.
A disabled Chinese woman sews shoe insoles with her mouth to feed her daughters. One afternoon, I used AI to make sure the world could hear her story.
-
I Asked 7 AI Models Which Chinese AI Stock They’d Hold for 3 Years. The Results Were 6-1.
I made a mistake this morning. Or did I? At market open, I bought 100 shares of Zhongji Innolight (300308) at ¥920. Within two hours, it had dropped to ¥847. I was down ¥7,300 — roughly 8% — before lunch. This is Part 2 of an ongoing series. Part 1: Before Buying Innolight, I Asked…
-
I Asked 4 AIs Which One Lies the Most. They Named Names.
I asked four AI models the same question. Not “which AI is best.” Not “which AI do you recommend.” I asked: which AI lies the most — including yourself? I expected deflection. Corporate hedging. Careful non-answers about how “all models have limitations.” Instead, they named names. Including their own. The Question I Asked All Four…
-
I Asked Each AI How to Squeeze the Most Out of Its Rival
I paid $480 a year for Claude Pro and ChatGPT Plus. I was pretty sure I was only getting about 30% of the value. So I ran a simple experiment: I asked each AI how to squeeze the most out of the other one. I expected both to quietly promote themselves. Instead, one of them…
-
Why Medium and WeChat Reward Completely Different Writers
WeChat gave me fewer than 200 reads per article. Medium writers earn $10–$500 per story. The gap is not just about money — it is about what each platform is built to reward. Here is the comparison nobody in China is making.
-
I Paid ¥1,000 for an AI Content Course. Here’s What I Got. (Failed Experiments #2)
I paid ¥1,000 for an online AI content course run by a WeChat influencer who promised to make me rich. Here is exactly what happened — and why that ¥1,000 now buys something much more useful.
-
Why My First Side Project Got Zero Visitors (And What Every Indie Hacker Can Learn From It)
I spent months building it. I believed people would find it, buy from it, tell their friends about it. They didn’t. Nobody did. For three years. Looking back, the biggest mistake wasn’t the product. It wasn’t the price, the design, or even the market. It was something more fundamental — and more common than most…
-
Why Investors Keep the Wrong Stocks Too Long — and the Mental Model I Use to Stop (Investor Psychology #1)
⏱️ Proof of writing date: This article was written and published on July 28, 2026 — the same day Innolight (300308) dropped 16%.View Wayback Machine archive → I once held a stock for five years — not because it was still the best investment, but because it had become my investment. The stock was Kanste,…
-
Before Buying Innolight (300308), I Asked 5 AI Models the Same Four Questions
⏱️ Proof of writing date: This article was written and published on July 28, 2026 — the same day Innolight (300308) dropped 16% on the A-share market.View Wayback Machine archive → Breaking news as I write this: Zhongji Innolight just completed a Hong Kong IPO, raising $6.81 billion — Asia’s second-largest listing of 2026. The…
-
The Apartment Where I Grew Up Has Been Abandoned for 20 Years. AI Brought It Back to Life.
A few weeks ago, a relative sent me a photo on WeChat. They had visited the place where I grew up — a five-story brick building about 55 years old. I hadn’t been back in years. The message came with three words: “It’s all gone.” I looked at the photo for a long time. This…
-
9 AIs vs. China’s Biggest-Ever Tech IPO: The Prediction Showdown
I won 1,000 shares of CXMT (688825) via IPO lottery — China’s #3 largest IPO ever. Then I gave 9 AI systems the same data and the same 5 questions. They all said something different.
-
I Built a Free Tool That Turns 8 AIs’ Investing Wisdom Into Your Personal Investing Coach
Seven experiments, eight AIs, $104 in API bills — and the most common question was still: which investing approach should I actually use? So I built a free tool that answers exactly that.
-
I Asked 8 AIs: What Will Humans Still Be Better At in 2040? Eight Answers. One Core Idea. One AI That Said You Shouldn’t Fully Trust Any of Them.
Episode 7: I asked 8 AIs to name humanity’s last irreplaceable advantage in an AI-dominant world. All 8 rejected creativity, empathy, and consciousness — and converged on the same structural insight: AI cannot lose anything. But the framings differed sharply. And Claude made the most important move in the series: used this question to argue…
-
I Asked 8 AIs to Pick an Investing School and Show Their Prompts. All 8 Said Value Investing — Then One Warned You Shouldn’t Fully Trust Any of Them.
Episode 6: I gave 8 AIs five investment schools and asked them to pick one, build an AI-assisted framework around it, and be honest about the limits. All 8 chose value investing — the same unanimous answer as Episode 1. But the frameworks differed significantly. And Claude said something important: AI can produce ‘garbage conviction…
-
I Asked 8 AIs: $2,000/Month, 20 Years, One Action. Seven Said the Same Ticker. The Real Debate Was About Something Else.
Episode 5: I gave 8 AIs $2,000/month for 20 years and asked for one action. Seven independently chose the S&P 500 index fund — the most unanimous result since Episode 1. The one dissenter chose a global index instead. But the most interesting answer wasn’t about which ticker. It was about why you’ll probably sabotage…
-
AI Experiments: 8 AIs Answer the Same Investing Question — Every Episode, Ranked
All episodes of AI Experiments — a series where I ask 8 AI chatbots the same investing question, separately, then make them judge each other’s answers. Four episodes so far: investment philosophy, stock picks, cognitive skills, and asset allocation.
-
I Asked 8 AIs: $10,000 Into One Asset Class, 10 Years, No Selling. Six Said the Same Thing. The Other Two Made a Better Argument.
Episode 4: I gave 8 AIs $10,000 to invest in one asset class for 10 years, no rebalancing. Six independently chose the S&P 500. Two chose Bitcoin. But the most interesting moment was the one AI that started with S&P 500, read the other answers, and changed its mind — citing a single valuation number…
-
I Asked 8 AIs What Skill to Learn for the Next Decade. They Disagreed. Then They Voted — and It Wasn’t Close.
Episode 3: I asked 8 AIs what single cognitive skill would matter most in the next 10 years. They gave 8 completely different answers. Then I made them vote on each other — and 5 out of 7 first-place votes went to the same answer. The AI that won had ranked itself third.