Light manifesto and random honks

2025-12-29

Towards the end of 2025, with the help of my dearest friends Gemini and ChatGPT, I finally finished translating my existing learning notes from Chinese into English. All the content has now been migrated into this brand-new space, which I call PelicanFlow Space.

2025 has been a year of both the best and the worst luck. Just one week after joining a new role at Foster + Partners, someone broke into my flat and took my most precious belongings — my personal laptop, which contained all the materials I had not yet backed up for my job application. I lost a significant amount of work-in-progress code and my most up-to-date portfolio files. For a long period of time, I was busy trying to catch up on lost work, recover files, and deal with complete chaos in my personal life. It took me nearly half a year to recover, relocate, and start a new chapter. Through this experience, I truly realised the importance of properly documenting and backing up not only data, but also thoughts, on a regular basis.

Another reason for starting this blog is my growing belief in the importance of knowledge sharing between real human beings. In an era dominated by AI-generated content, it has become incredibly easy to obtain information. However, how information is digested and truly understood, how we identify fake, invented, or biased content and reflect on it critically, and how we avoid endlessly repeating shallow ideas that have already been said thousands of times — these are the questions that actually matter. AI can never give you the answers to these.

As human beings — and particularly as engineers — we need to be able to solve problems instead of simply burning resources for fun.

To achieve this, it is not only about the ability of building tools( Attention: with GPT nowadays everyone can build things), but also about connecting the dots, understanding why things are not working, and deciding the direction in which we choose to move forward.

That is why most of the knowledge I posted here is pre-AI and many of them are even like dinosaur, such as fundamental data structure that was invented when storage is ultra-limited. I find this knowledge might become sleeping in the museum and dying in human conversation soon, which could be extremely dangerous.

— The Untitled Penguin