Perhaps it was because I was a boy, but I was active from a young age, constantly running and jumping, never stopping for a day. As my grandma would say, it was as if I had eaten “monkey meat” and couldn’t stop moving. Perhaps this childhood activity laid a good foundation for me; my athletic ability in later studies was quite good. Later, when I played basketball, my physique gradually developed into what it is today. The story begins with my experiences after college.

This blog was built from scratch in a few days. Previously, Next.js was used, and it was built with AI at that time. I basically don’t know front-end development. AI did it all. I just made some simple adjustments, but after a while, I forgot how to launch it. Coupled with Next.js’s security issues some time ago, I simply switched to Hugo. The most important thing is that Cloudflare’s CDN supports Hugo, so no extra configuration is needed; you can directly upload the entire project. It can be said to be very convenient. However, the problem was that I hadn’t played much with Hugo. Fortunately, there was AI, but there were also some minor hitches in between, and I have some thoughts to share.

The successive passing of my parents truly felt like a huge tumble, leaving me a bit numb. I didn’t know what sadness was, nor how to express it. Only occasionally, when thinking of them, would tears fall uncontrollably. For a long time, my emotions were at an all-time low, feeling as if I was at a critical point where I might shatter with the slightest touch. I tried to seek solace from various angles. My biggest knot was the desire for even a single word of farewell. I wasn’t by their side when they passed, and I even had unresolved emotions. Little did I know, I would never see them again. From initial sadness and regret, it finally turned into silent acceptance. Of course, there was no reason not to accept it. As someone who has read a few books, there was no reason not to accept, but behind the acceptance, there was indeed an unspoken discomfort. Many people passed away before them; this was a form of self-comfort. Just like when I didn’t want to get married, I thought there were countless people in the world without descendants, so why should I be the one to carry on the family line? Now, I can apply this again: countless people in the world have lost their parents, and I am just one of the many. They were also normal people, and reaching this point was very natural. All the realistic and self-comforts always had a loophole. No matter how reasonable, normal, or natural everything was, I always felt unwilling. The reason was that on the night my father passed away, I dreamt of my deceased mother pacing anxiously in front of me, seemingly with something difficult to say, but she never spoke it. I didn’t think much of it at the time. Reflecting on it later, I suddenly understood.

These days I have been thinking to build a great APP. Two reasons:

1. I am out of job and I need to do something to make a living.
2. I would like to build my own business.

Then the wind is AI and I would like to have it as my back. There are many great products, closed or open source. But most of them is independent of people’s daily life. They are good apps but I think one more step can be made to change something. So the mind here is: make AI invisible in your life. But still it helps you without your notice. This is the ideal AI App do.

I previously developed a chatbot for intrenal use. The mind is simple: Use IBM watson assistant as the bot and a simple UI. But it is not as good as ChatGPT nowadays. At least it is not easy to tain. The documents have been manually composed to feed it up. So this little project spreaded so many teams.

Then there are a lot of opensource LLM, like llama.cpp, alpaca.cpp, dailai, and gpt4All and I am wondering can I just make a easy-customizing AI tool with these projects. Especially help read documents. Maybe more?

Why choose begining? Obviously, AI, whether in technology or concept, did not just appear, but developed through decades of technological accumulation. However, before the launch of ChatGPT, it had remained relatively unknown to the general public, at least in terms of its relevance to their daily lives. But things changed after that. Is this the peak of AI development? Clearly not. Looking back at the development of computers, IBM created the first commercial computer, but it was not IBM that truly popularized it, it was Apple, Microsoft, Dell, and others who brought it to the level of personal computers. I personally think that these two developments are likely to be similar, this is just the beginning, and the best is yet to come. The only difference may be the speed of development. Let’s wait and see.

Up to my age, I have got through a lot. And I am supposed to understand a lot. To some extent, I do. But due to my condition, I am nothing compared to other elite. I am from grassroot. I am an ordinary dog.

I am not lucky because of my family education. This may not be fair to parents. They tried their best to bring me up. But this is the reality. I am not here criticize my parents. I want to show something that hidden in my life.

I have been working for years. But I did not feel better and better.

Why?

I have been working in my current company for more than 8 years. This company was a dream place for working, but not any more. And it would take much for me to go away. Not because how senior I am but I have a little got used to its pace. Maybe I should not use a little here.

Recently I have been learning a coursera course, Learn to learn. Actually I had a tried before. And I just let it go. Because I was confident. I have beening good at learning after all these years. Especially when I was a student. I was always at the top class. I could learn the stuff so quickly and remembered them for so long. I can even recall the passages I recited in my 8th grade. I was so proud of my memory, my efficiece, and my way to learning.

I try to deploy my dev env and I experienced some pitfalls. I want to share this. If this can save people’s time, that will be my happiness. And this is my first blog about ebpf. This will be a good start.

Prerequisits

1. Install latest version pahole or just build it from source Make sure pahole lib in the LD_LIBARY_PATH:

    `export LD_LIBRARY_PATH=/usr/local/lib:$LD_LIBRARY_PATH`
   
   
    Without this set, I just faild to compile the kernel with error:
   
    `FAILED: load BTF from vmlinux: No such file or directory`
   

2. Install llvm sets. Again it should be the latest versioin. I just referred the kernel doc. But there is an error when to compile the test cases of bpf: