AI Project Roundup from the Past Six Months

🌱 Entrepreneurship/Content Experiment Log & Summary (Periodic Review)

Starting from the Spring Festival holiday last year (late January), and due to the handover period after changing jobs, followed by having more free time post-job change, I went ALL IN on AI for the past six months. I tried quite a few projects and stepped into many pitfalls. Now, taking advantage of the holiday, I'm organizing and summarizing these experiences to share with everyone.

Here's a simple chronological list:

1. AI + Web Novel Overseas Project

   • Project Goal: Translate domestic web novels into English and publish them on overseas fiction platforms.
   • Monetization Model: Revenue from reader subscriptions.
   • Trial Period: Short-term (a few days).
   • Result: Failed.
   • Reasons for Failure:
     • High-quality web novel genres are already published overseas.
     • This type of content requires deep localization and lacks universal global appeal.
     • Among domestic web novels, themes that can be accepted by Western/global audiences are extremely scarce.
   • Key Learnings:
     • Cultural differences must be considered for content going overseas.
     • The importance of market research.
     • The value of rapid trial and error.
     • Original novel publishing overseas might be viable, but it heavily relies on the author's ability to build the world, characterize, and control the plot. AI can only be the icing on the cake; relying purely on AI is impossible.

2. White Noise Project

   • Platform: YouTube
   • Content: Create and upload various types of white noise audio (e.g., rain, ocean waves) paired with AI-generated visuals, synthesized into long videos (typically up to 10 hours).
   • Advantage: Low production barrier; can use scripts for batch, automated generation of long videos, saving editing time.
   • Monetization Model: YouTube platform ad revenue from views.
   • Challenges:
     • The white noise niche is highly homogeneous; it's hard to create something new (it's just rain, wind, fan sounds, etc.).
     • High fan loyalty but low migration; many videos from ten years ago have billions of views, but even large YouTube channels see very few views on newly posted white noise videos.
     • After persisting for 2 months and producing/releasing 15 videos, view counts were dismal, often in the single digits.
     • Tried long-tail keyword optimization, content innovation, and format innovation, but homogeneity is too severe.
     • A large number of homogeneous accounts were registered in the past six months.
     • Cold start became hellishly difficult, so I gave up.
     • Result: Abandoned.
   • Takeaways:
     • Content innovation and SEO optimization can be effective, but the most important thing is choosing the right niche. If the niche is wrong, nothing helps.
     • Barrier to entry is crucial. If it's too low, competition is too fierce, making it impossible to stand out.

---

3. Zi Wei Dou Shu + AI Interpretation

• Project Goal: Combine traditional fortune-telling (Zi Wei Dou Shu) with AI interpretation, attempting to productize it.
• Reason for Trying: I'm personally very interested in metaphysics/fortune-telling. Compared to traditional Bazi, Zi Wei Dou Shu has finer granularity and a larger knowledge base, giving AI interpretation more advantage.
• Implementation: Development of a chart calculation tool + knowledge base construction + Demo website launch for experience (beijixing.me) + small-scale promotion.
• Experience: The experience was indeed good, with high accuracy and fine granularity.
• Challenges: This niche has an inherent flaw. Fortune-telling places more emphasis on human connection and providing emotional value. If it's a fortune-teller (a person), even an average one, interpreting a Bazi or Zi Wei Dou Shu chart would cost at least several tens to hundreds of RMB. For highly skilled fortune-tellers, thousands or even tens of thousands is normal. For AI interpretation, no matter how accurate, almost no one is willing to pay, and even if they do, it won't be much (a few RMB at most, based on the domestic payment landscape). The psychology behind this is complex. Scarcity is one aspect; on the other hand, many feel AI interpretation lacks "soul."
• Status: Ongoing.
• Future Plans: Continue optimizing the product. Considering the domestic payment situation, the subsequent plan is to go overseas. Considering the acceptance of traditional Chinese culture in Europe and America, the target market will still be East Asia and Southeast Asia.

4. Tarot + AI Interpretation Product

• Background: Considering the acceptance of Zi Wei Dou Shu among European and American users, Tarot is the first choice for going overseas. Therefore, a website for AI-interpreted Tarot readings was built.
• Advantage: Tarot content is widely popular, has a large user base, and payment willingness is stronger in Europe and America.
• Current Status: Website experience is being continuously optimized; promotion hasn't started yet; payment integration is in progress.

5. AI Resume Optimization Project

• Project Goal: AI-assisted resume optimization.
• Background: During a conversation with a friend, I learned their team of 3-4 people had invested 3-4 months and were still in the prototype design stage without a live demo. I helped them build and launch a demo in 2 days (cv.makemoneywithai.site), naming the project 马上鹿职 (a homophone of 马上入职, "get hired immediately").
• Subsequent Development: Due to the need for continuous product iteration and insufficient support capacity, the complete demo was handed over to the friend for their team to maintain and operate independently.

6. Independent Website Building + SEO Project

• Project Goal: Build a B2B vertical niche independent website and achieve lead generation, handing over leads to partners for follow-up.
• Technical Implementation:
  • Use an Agent to quickly build a basic framework prototype.
  • Based on the prototype, integrate optimizations and deploy it live.
  • Integrate a DingTalk robot to automate lead notifications.
• Challenges:
  • Building the complete website is just the first step. To get real leads, continuous SEO optimization is needed, which is a very long cycle.
  • Other optimizations are easier, like site speed, sitemaps, etc., but backlink building is difficult.
  • Content requirements are high; high-quality content is needed for search ranking, which also requires continuous investment.
• Takeaways:
  • Mastered the entire process of rapid independent website building and gained solid technical experience. Recently, I've been helping others build independent websites (paid).
  • Mastered common SEO methods.

7. Self-Media Operation - Xiaohongshu Account for Car Selection, Buying, and Usage

• Content Direction: Sharing experiences on car selection, usage, and purchasing.
• Operation Status:
  • Competition is extremely fierce.
  • Content: Too deep, users don't read; too shallow, lacks differentiation.
  • Interaction data performs well; setting topics can lead to very good view counts.
  • Gaining followers is difficult; hard to achieve stable growth.
  • Due to follower growth difficulties, motivation for continuous updates is weak.

8. Self-Media Operation - AI Entrepreneurship + Solopreneur Direction

• Direction: Documenting the entrepreneurial journey, sharing tools, and outputting personal opinions.
• Advantages:
  • Fast interaction and feedback, suitable for validating product needs and topic discussions.
  • Suitable for building in public.
• Challenges:
  • Requires a stable, consistent content update rhythm.
• Takeaways:
  • Found the feeling of conversing with users; clarified that a "documentation + tools + growth" style can work.
  • Accumulated initial followers and interaction data.
  • Cold start is basically complete.
• Current Operation Status:
  • Duration: 3+ months.
  • Follower Count: 800+.
  • Content Format: Primarily Xiaohongshu image-and-text posts.
  • Content Direction: AI tool usage, solopreneurship practice sharing.
  • Interaction Effect: Overall positive feedback is good; can gain some new resources. Currently maintaining a trend of roughly one image-and-text post per day.
• Future Plans:
  • Content Format: Increase video content (planned weekly updates) (publish on Bilibili/Video Channels/Douyin/YouTube).
  • Continuous Optimization: Improve demand collection methodology.
  • Development Direction: Continue sharing entrepreneurial experiences and insights.

Summary & Insights

Every attempt yielded gains.

1. Honed the ability to judge entrepreneurial niches. I watched a livestream a couple of days ago about how to judge if a side hustle is worth doing. I found it quite interesting. There were three criteria:

• Look at how many people are teaching it. If more than 20 people are teaching it, it's already a red ocean.
• Look at the trial cost in a field. If the cost is too high (e.g., self-media production cost), it's not worth trying.
• Look at the monetization potential. If accounts that have already succeeded still have only average monetization potential, then it's not worth trying either.

I deeply agree. Besides the above points, the barrier to entry is crucial. There must be a barrier, but it can't be too high, especially not heavy financial investment. Long-term time investment is acceptable, like follower accumulation during the cold start of self-media or continuous SEO optimization for independent websites. The prerequisite is that the monetization model must be clear enough and the prospects sufficiently good.

2. Honed the ability to quickly build and launch products. Through the use of AI tools, I can quickly prototype product selections and deploy rapidly to validate demand. This is a skill I'm continuously learning.

3. Gained a deeper understanding of entrepreneurship, especially the ability to collect and validate demand. How to quickly validate if a demand is real or fake? You can do some research, like searching related topics on Xiaohongshu. For overseas, search related keywords on Twitter and Reddit to see user discussions or post yourself. If user response is strong, it's worth doing, and your first wave of users is also easy to acquire. Otherwise, even if you build the product, you can't promote it; it's still useless.

There's a saying: the best way is to sell your product first, then start developing it. For example, you can first create a landing page with a description, and those interested can leave contact info to join a waiting list. If enough people support it, then develop it; otherwise, there's no need.

What entrepreneurs fear most is self-deception. Entrepreneurs are most prone to falling in love with their own solutions and reveling in them. As mentioned in the book Running Lean, you should fall in love with the customer's problem, not your own solution.

That's all for my sharing. Thank you, everyone. Feel free to share your thoughts and experiences in the comments, and welcome to join the community. Let's explore the infinite possibilities of AI entrepreneurship together!

#ArtificialIntelligence #Entrepreneurship #AIEntrepreneurship #ALLINAI #EntrepreneurialExperience #LargeLanguageModels #Startup #LightweightEntrepreneurship #AITools #Solopreneur