AI on mantou's blog

AI is more than a story

Sat, 25 Jul 2026 00:00:00 +0800

After learning a lesson, the last step is to do it as quickly as possible. This has led to substantive changes in their own concept sheets, relationship tables and flowsheets, so that the meaning of learning can be truly realized.

Maybe it’s an environmental effect! In our community, a whole bunch of people are learning to practice and use. A slight disregard for the content of the course will quickly open a wide gap between the students. Nor is such a state of affairs an example of alarm, as it happens almost every other time. In contrast to the role models themselves, it is inevitable that there will be some resentment and remorse. As a result, in a small environment in which alarms ring from time to time, each day they are forced to learn and grow. It is not simply a question of learning to be good, but of beginning to be indifferent from the heart. The time of the AI, it’s more important to catch up.

Unconventional day

Thu, 23 Jul 2026 00:00:00 +0800

When the sub-room came to my house, it took me to test Mac’s computer, and I had a visual feeling of that “flow.” According to the laughing teacher, Linux and Mac are essentially managing a pile of documents. In a system like that, there is no mouse by default. It’s so hard not to think that Windows is a little behind when you feel that kind of system. Because we humans are the biggest obstacle in the Windows system, and it takes a lot of time to get there. The former, on the other hand, saves these steps and inadvertently increases the efficiency of their work.

Making good use of AI

Wed, 15 Jul 2026 00:00:00 +0800

It is true that some people do things without the most basic standards of judgement and principles, but only when they see the opportunity to make money, whether it be a pit or not, and simply fall into it with a brain. As for the immeasurable consequences that could ensue, that would be absolutely and completely ignored. It was not clear at first that, after two three observations, it was as if he had brought a clear lens. While it is the same thing before and after, it is the mirror that makes itself more visible. I’ve seen a lot of possible dangers!

Another day of exploiting AI

Tue, 02 Jun 2026 00:00:00 +0800

I did some fairly in-depth investment research today. It has always been an interest of mine, so I use several of the major AI tools for it: Gemini, ChatGPT, and Claude. In terms of overall ability, Claude is second to none.At the same time, its token costs also seem to be the highest. For this reason, I went to investigate again.It was quite complicated, and I did not really understand it, so I will not offer a vague explanation here.

A day renovating the blog

Mon, 01 Jun 2026 00:00:00 +0800

The dream I had last night seemed pretty good, and it made me think about a lot of interesting things. What exactly it is, I really can’t remember. After getting up, the passionate emotions in the early morning seemed to be calmed down in time after a short rest at night. Yes, it is impossible for oneself to make wiser choices if he is too excited. It is easier to be a smart person, but it is more difficult to be a wise person. But the latter state is always worth learning from.

It takes time, fix it as soon as possible!

Fri, 29 May 2026 00:00:00 +0800

Today’s investment research and AI interaction work is very intensive and requires a lot of work. It can be considered the right thing to do, because there is a saying that if the workload increases after using AI, then it is used correctly. Otherwise, there is a high probability that it is used incorrectly. How to understand this matter? Using AI correctly means that the intensity of your work content has increased a lot, and it also corresponds to more workload. Because AI is a tool that you use correctly, more of its value is tapped by you. And if you use AI to do things that you can and should do yourself, such as reading and summarizing the summary and central idea of a book. Then it is taken advantage of by AI. So AI was not utilized well. In this state, AI does not exert its due value. I’m lucky that I didn’t let myself fall into this state. The workload has also increased a lot and is very challenging. I feel full of joy at the end of the afternoon!

Claude

Tue, 19 May 2026 00:00:00 +0800

Today I did some vibe coding and completely repaired the missing files on my blog. So from now on, my blog will be successfully iteratively upgraded. As a result, all the articles that had been lost before were found. At this point, my blog articles were finally connected. It seems that only after experiencing it in person can you realize the charm of top-notch production tools.

A life that unites knowledge and action is the most cost-effective! Didn’t I know that Claude was the best AI tool before? Especially within the field of programming. But I just knew it and didn’t do it, so this “easy” problem was dragged around without iteration. Fortunately, this is just a gadget of mine, not a commercial work, otherwise I would have been eliminated countless times.

Monday passed by in a hurry...

Mon, 13 Apr 2026 00:00:00 +0800

Yesterday, when I was chatting with a programmer about the functions of two programming AIs, Cloud and codex AI, the other person said an interesting point: When comparing codex and Cloud, I found that there is not much difference in the effectiveness of these two paid AIs of the same level when programming the same project. When I communicated with him, I strongly expressed that I wanted to use Claude for vibe coding. The other party calmly analyzed it for me: In fact, on the basis that the efficacy is basically not much different, deliberately looking for Claude to use it, the price/performance ratio is not that high. At least not for my current needs, I agree. And if necessary, you can also experience it yourself so that you can feel the difference more intuitively.

Keep iterating...

Thu, 09 Apr 2026 00:00:00 +0800

Today’s workload continues, pushing my work to a new level.

The main construction work is in the two aspects of “body construction” and “trading system”. At the level of body shaping, I iterated on a Gem specifically targeted at fitness and fat loss based on the new prompt words. Compared with the previous one, which was limited to a single chat box (it is easy to forget the context after a long time) and a relatively weak AI assistant, this improvement can almost be said to be a “Davis double-click” iteration. Regarding the trading system, his work is aimed at the two fields of US stocks and cryptocurrency. I also conducted in-depth discussions, exchanges and learning about specific strategies with AI. In the U.S. stock/cryptocurrency market, targeted investment paths were discussed regarding specific positions and capital volumes. At the same time, some of my original currencies were screened with the most stringent criteria based on the minimum necessary screening criteria. Therefore, when investing and trading under such circumstances, I no longer wander around blindly without a map as I did in the past.

Continuous iteration

Wed, 08 Apr 2026 17:00:00 +0800

Today’s work is actually directed towards the questions raised in yesterday’s article. Under the new reminder structure, priority reflection work was carried out on those key items. To be honest, today’s workload is much greater than yesterday’s, but this is mainly reflected in the magnitude, because I have already completed most of the quality-related parts in yesterday’s work. Today, we rely on what seems to be “batch operations” to do higher-quality things based on higher-quality gems.

Making the most of the opportunities AI brings

Tue, 07 Apr 2026 19:56:00 +0800

Today’s update is divided into two parts:

Regarding the problems that appeared in yesterday’s blog, I carried out extremely thorough optimization and rectification work this morning. In terms of asking AI questions in the future and letting AI do research work, there is already a set of relatively optimal solutions;
I used it to practice on some important issues that came to my mind yesterday. By using the latest AI instructions, the answers to those questions have reached a new level of quality.

So in the next work, I will deal with my previous trading assistant in the most rational way: FIRE! It will be replaced by the new GEM I set up today. It is crucial to let more professional GEM do professional things.

An urgent, anxious moment

Thu, 02 Apr 2026 16:57:00 +0800

I have been conceiving some very interesting ideas recently, and it can be said that many of my product prototypes are being fabricated. With the assistance of AI, the implementation and implementation of these gadgets is actually much faster. You can know what your idea is at the lowest possible cost and in the shortest possible time, which can also quickly accumulate your experience in doing things. In fact, at this moment when we are tightening our belts, what is more important to us? Launch some products as quickly as possible, leave as many traces on your products as possible, and try to accumulate your own unfair advantages in making products…

3.26

Thu, 26 Mar 2026 00:00:00 +0800

There was a time when I was a bit stupid in vibe coding. Everything stems from my lack of familiarity and proficiency in using GitHub, so I failed to make many interesting gadgets. It’s a pity, but isn’t this inevitable?

When it comes to making products, I am really a novice! After all, I claim to have been writing articles for four years, but from the perspective of communication, it is really unrecognizable. Looking back now, I can almost say that they were all self-pleasure projects, and I was often troubled by them. So based on this premise and background, the content of vibe coding that I started has actually become a traceable work content. What’s more important? That is to first come up with a good idea, then run through an MVP-level project, and then continue to iterate and optimize during the running process. In a precise sense, production is the foundation for long-term sustainable footing and growth, and other content is nonsense!

How to understand introversion in the AI era?

Thu, 19 Mar 2026 00:00:00 +0800

I had some troubles yesterday, and I tried to digest them by writing an article. This writing actually took up a lot of time for what I originally had to do, so my sleep last night was naturally delayed a lot. As a result, when I woke up the next day, I could almost be said to be slightly depressed. During the morning reading session, I can say that my efficiency was really not that high, let alone any gains. But no matter what, yesterday’s output did leave some valuable “legacy”. In line with this situation, I took advantage of the situation and chose to finish them one by one today…

Another day working alongside AI

Sun, 15 Mar 2026 00:00:00 +0800

Today’s work is mainly to review and sort out what we learned yesterday, and to conduct a more rigorous sorting and planning of the things to be done next, that is, overall planning.

According to what I learned yesterday, I made many detailed arrangements on many specific matters (for the purpose of production):

1.Writing Micron financial report;
1. Write tutorials on how to use Google and Github;
1. Train your brain (by learning four foreign languages in the next year)
1. Thoughts after writing a series of articles on logic;
1. Force the daily input amount to 3 or more than 10,000 words.

To be precise, these contents are actually a large piece of quite complex and highly complex learning tasks. Many of the contents are similar to “stir-fried old dishes”, but this time the “old” content is significantly lower. Because the tome is very serious and solemn in its focus on “production”.

I seriously suspect that I have a tendency to be a workaholic

Wed, 11 Mar 2026 00:00:00 +0800

(Let’s start at about 16:00)

After having paid for an AI, I’m almost free for a day. Every day, busy and AI researches the feasibility of their ideas, and because they have too many ideas in their minds, it leads to too many questions each day to ask AI. While many ideas are perjured, there are also many interesting ideas that have been put into practice, which have led to a much higher level of overall efficiency from the bottom.

A busy day

Tue, 10 Mar 2026 00:00:00 +0800

Today can be said to be a day of super crazy bugs and catching up on progress. As soon as I arrived at the office in the morning, I went to Gemini to ask some difficult questions about my “Operation Fasting”. In the end, I asked more and more questions, and the questions became more detailed and in-depth. In the end, not only did I refine what I wanted to eat at each meal down to what I bought in the supermarket, the number of grams and the order of eating, but I also found an alternative strategy for gym fitness, so that I finally customized a very complete MVP model of “fat loss and muscle gain” based on my personal situation. Not only that, I also conducted as comprehensive a review and forward-looking optimization work as possible on some of the operational logic and execution actions I conducted during the TACO trading process in the past three trading days. In addition, I wrote a new foreign language practice manuscript, making the most basic preparations for the morning reading materials the next day.

March 1

Sun, 01 Mar 2026 00:00:00 +0800

As of 5:46 pm, I have mainly done two things today: compiled an article about Micron Technology and uninstalled openclaw on the computer.

The former’s job is to deepen his recent impression of AI learning, and organize the reading notes he learned into a WeChat public account article and publish it. This is just my routine.

The latter’s approach may surprise and curiosity some people. Today, when openclaw is so popular, not only did it not continue to play, but it was also directly uninstalled? In fact, I had an obvious idea of this decision the day before yesterday, and it was not until today that I officially abandoned this “lobster”. The reason is simple: the current openclaw is not worth spending more time and energy on me now.

In the AI era, I have no time to miss you!

Fri, 27 Feb 2026 00:00:00 +0800

If the year 2023 is called “AI year”, then this is 2026.A critical year for moving from a “technology carnival” to a “value landing”。

I wrote an article on OpenClaw at the end of last month, when the open-source AI project was also called Clawdbot. From the very beginning, it seems to have received a great deal of attention, and the level of debauchery is almost equivalent to the attitude of the people at the beginning of 2023 with regard to Chatgpt. It’s the same heat that made me see this project. After a preliminary review, a copy of the paper was written on the micro-mail public. Then, more than 20 days later, the official deployment and use of lobster, a fire-in-the-house project, began on the local side of his computer.

No bragging, no slander, objective description

Thu, 26 Feb 2026 22:03:01 +0800

After finishing today’s content course at around 2 o’clock this afternoon, I originally communicated with my AI assistant some of the inspirations I learned during the refresher training, so as to integrate and revise the current fixed investment list. As a result, I came up with an impromptu idea and asked a few questions about AI casually, and then devoted my whole afternoon to communicating with AI. The more we chat, the more in-depth we become, the more we chat, the happier we become, the more we chat, the more excited we become, and the more we chat, the more joyful we become.

This damn feeling of growing up

Mon, 23 Feb 2026 00:00:00 +0800

In these four days, what I did most was chatting with openclaw. Not only use it as a substitute for ordinary chatting, but also use it as a tool to enhance the atmosphere of the group chat. Apart from that, even chatting with Gemini and chatting about all kinds of things didn’t bother him. The other party was not only professional, but also very patient. That kind of professionalism is like a hexagonal warrior’s professionalism, and that kind of patience is also an all-round professionalism.

The twenty hours the day before yesterday, yesterday and today

Sun, 22 Feb 2026 00:00:00 +0800

Starting from the day before yesterday, since I got the GEMINI3 pro account, the time I use AI has suddenly increased a lot. Not only because this model is more advanced, but also because I no longer have to be as “timid” as before.

Due to the favorable conditions of more time and higher model authority, my own wild ideas can be implemented almost unimpeded. For example, I wanted to make a machine-intelligent dog equipped with the “GEMINI3 pro” brain. So after I shared this idea with AI, I not only learned that this project was highly feasible, but also simulated the interaction after implementation for me through the preview! If you are not satisfied, you can fine-tune it at any time without worrying about whether the other party is bored. So just like that, I designed my own “Cyberdog” prototype yesterday: the first phase of the code level was completed in a day and a half!

Wanting to write, without the energy to match

Tue, 10 Feb 2026 00:00:00 +0800

I was sitting in the office today and wanted to type some words, but my stomach just didn’t feel right and I felt a little nauseous. This state lasted for a long time and did not end. During this period, I drank tea to relieve my discomfort.

To be honest, it’s difficult to produce meaningful content in this state. Fortunately, I learned some new things by chatting with AI in the afternoon. Even though I was stuck in this disgusting state all day long, I still couldn’t accept it. Because compared to this kind of physical discomfort, the state of “doing nothing” will make me more unbearable.

After so long refusing to change, something finally clicked

Wed, 04 Feb 2026 16:00:00 +0800

In order to fulfill my commitment to product iteration, I worked hard from 7 o’clock last night until more than 1 o’clock in the morning. After the computer, mobile phone and human brain were completely exhausted, I had to take a rest to adjust my state. After waking up at 8 o’clock in the morning, I had no intention of staying in bed. My unfinished “legacy” pulled me out of bed and hurried to the computer to continue working. In the end, I relied on the “magic power” of Codex to complete the last part of all yesterday’s tasks: ** Completed the script configuration that can view the background data of this blog website in real time. **

The impact of AI cannot be underestimated

Mon, 02 Feb 2026 20:01:01 +0800

I remember that last week my focus was still on an open source AI project called Clawdbot. This week, I started to switch to the AI analysis tool that comes with WeChat Souyisou. Paying attention to the former is to catch up with trends and popularity, and the latter is because it is a practical need to use it to analyze the quality of your own WeChat public account articles. But in the final analysis, the impact of both on me is actually equally important: one makes me more clearly aware of capital’s love for AI technology, and the other makes me more clearly aware of the problems in my current article, which helps me make more targeted adjustments and optimizations to my current article from a senior and objective perspective.