Most self-media creators, including YouTube bloggers, only take the first step when they break into a topic area and start making money. Most channels cannot become top creators, and the traffic the platform provides is limited. To achieve stable income, one must maintain the ability to create content consistently.
My early two Chinese-language channels both passed YPP within three months, and soon earned enough to support myself. The Mr Hunzi channel passed YPP in just three weeks, started making money in the second month, and from the third month onward, monthly income exceeded 1500 USD, with some months surpassing 2500 USD. In a small city in China, this is enough to maintain a decent life.
Of course, compared to programming, this income is still relatively low. I can receive 5000 USD for writing a not-very-difficult functional software once. However, delivering software means dealing with people, and inevitably encountering individuals I dislike. Moreover, programming gigs are not always available, because I mainly take orders from foreign clients—the same overseas clients from when I ran my company. Although we no longer do hardware business, they highly value my software skills. When their friends have simple development needs that they cannot handle themselves, they introduce the work to me.
Sometimes I also share larger projects with friends. After reaching middle age, I have become quite resistant to highly scheduled work, so I am picky about clients and do not take just any money. Therefore, when I do self-media, I am relatively dedicated.
When I ran Chinese channels, the profitability was high. Later, due to political factors, I could no longer continue them, so I switched to English channels. However, because of my intense criticism and mockery of the Indian government, my channel was banned by the Indian government, which significantly affected my traffic. Therefore, to maintain the channel’s traffic and revenue, I need to keep creating consistently.
Basically, every day I post at least one original video and one secondary-creation video. Sometimes when there are no good videos to share, I post two original videos instead. Many netizens are amazed at my creation efficiency and think I must work extremely hard. In fact, being serious is necessary, but I am not particularly hardworking. On the contrary, I have been experimenting with different channels and websites, increasing the proportion of automated processes, and trying to reduce my own involvement as much as possible.
For example, I operate many websites at the same time. Some make money, some do not. Some can earn steadily for a long time but with low profit. Some have very strong explosive power for a few months, then get de-ranked by the system and stop earning. I am also trying different YouTube channels. In 2025 I created eight channels, and only three have passed YPP so far. It looks like this current channel will definitely pass YPP too, but it will require increased creation effort to achieve good income.
With so many channels, I cannot possibly pour massive energy into operating all of them. Even if one channel only updates three videos per week, that is already very intense work. Not to mention someone like me in middle age—even a small team of three to five people would find it difficult to maintain three channels plus multiple websites. Therefore, my strategy is to automate most of the work and let AI handle it.
For instance, most of my websites are built with WordPress, which has excellent API interfaces. I can use Python scripts to write programs that automatically update content. However, to do this, I must find good content anchor points to avoid meaningless duplication that leads to Google de-ranking, wasting effort and earning nothing. Of the four websites I tried in 2025, only one is stably profitable, but not much—only a little over 300 USD per month. Among the other three, one once earned over 1000 USD in a single month before being de-ranked by Google, one still has not earned anything to this day, and the last one was directly banned by Google and receives no traffic anymore.
But my 2025 was not entirely fruitless. I earned nearly 10000 USD through websites and gained one profound lesson: I should not continue wasting time on websites. Their returns are far worse than videos. However, if someone particularly enjoys building websites and can invest most of their energy—unlike me being lazy—it can still be a decent money-making business.
I will consider packaging this tool properly when I have free time and providing it to friends interested in making money with websites. It can automatically anchor content from specific websites or channels, perform secondary creation, automatically publish articles, and even automatically generate or retrieve images and insert them into articles, achieving a complete combination of text and images.
In addition, the engineering experience I gained from website automation projects allowed me to automate most of the work in video creation. If I do not pursue very high video quality, my current workflow can now take an outline, hand it to AI, and let AI directly generate speech from scratch, load a virtual character, match image materials, apply image effects, and render the video. It can also customize resolution, choose different image-generation or content-creation AIs, and release versions in multiple languages.
However, at present, the three steps of automatic writing, audio generation, and image matching still require manual confirmation, otherwise quality cannot be controlled. Therefore, I have not yet replaced my current workflow with it. In all workflows for the MrHunzi and VoidHunzi channels, I still need to manually intervene in audio to correct obvious rhythm errors, manually select appropriate images, and operate video editing software. Everything else is already fully automated. The manuscripts I write myself and those generated by AI now offer virtually the same user experience. Still, I do not dare publish without checking at all.
In other words, although I still use traditional editing software like Jianying, in reality I have already broken free from dependence on it. Generating videos automatically through program code offers extremely high customizability. I can define special effects that do not exist in editing software according to my preferences, resulting in better final video quality.
This entire process has now reached the final breakthrough stage. But up to now, I have not found better solutions for image confirmation and audio correction. It may still take a few months to fully free myself. I have also tried various online internet audio services. The old Microsoft one was quite good and could be fully automated. But recently Microsoft’s AI service has regressed severely, with audio generation quality dropping sharply and frequent abnormal pauses. Models like Alibaba perform very well in Chinese, but I feel their English is not as good as the VoxCPM I currently use. After fine-tuning, VoxCPM’s generated speech is already close to perfect, but it still has abnormal pauses between sentences that must be manually corrected. Although it is better than Microsoft’s speech and can be controlled via program code, it is still not perfect. Google’s TTS is said to be excellent, but I have not had time to test it recently.
Additionally, the current solution runs on Mac, where some steps are not efficient and require switching to PC. On PC, some parts of the program run far less efficiently than on Mac. It needs both systems working simultaneously, so it is not suitable for packaging and sharing yet. But if I spend more time on these issues and summon the spirit I had when I was young, I should be able to solve them in two weeks. It’s just that now I am a sloppy middle-aged man living a lazy life. Most of the time, I only open the computer to create when a video must be released. Without AI’s help, I could not even maintain one channel.
Regardless, in 2026 I will finally resolve these problems and share them with friends at the right time. If you are also interested in self-media or making money online, welcome to follow this channel. In the near future I will release videos completely generated by code. I have already produced quite good sample videos, but some parts still need improvement. Welcome everyone to follow along. I will definitely share my techniques with the channel’s loyal viewers, rather than sharing them just because someone writes in the comments: “Sir, can you share your tool with me?”