AI makes creation easier and the world more vibrant.
💬 Deep Dive
In my daily work I often need to answer product questions from overseas users. My usual approach is to first write in Chinese, then use ChatGPT for translation, and finally copy and paste the result into the chat box to send to the user—a rather tedious process.
This made me wonder: could I simply type in Chinese in the chat box and use a plugin to automatically translate and replace the original text, eliminating the need to switch back and forth?
With that idea in mind, I immediately tried creating a plugin with Trae. The result isn’t perfect, but it works.
For someone with no coding experience at all, relying entirely on AI-generated code and piecing it together step by step, it was very rewarding to create a Chrome plugin in just half an hour.
However, while it translates and replaces text properly in the chat boxes of tools like Grok, Lovable, and X, it doesn't accurately recognize text in the Reddit comment section or the Google search bar, so I’m still fine-tuning it and have not yet published it on the Chrome Web Store.
In the process of optimizing this plugin, I encountered a few issues and gleaned some insights:
1. AI Programming Lowers the Barrier, But Beginners Still Encounter Hurdles
Developing a plugin might be the simplest entry point into AI programming, but even if AI can help you write code, complete beginners will still run into various pitfalls—especially when you try to improve functionality, as AI may disrupt the original features.
For example, I wanted the plugin to use shortcut keys in the Reddit comment area and in Notion to translate and replace text. However, the new version generated by AI included not only the translated text but also retained the original, and it even reverted the previously set styling back to the initial unattractive version.
This forced me to roll back the changes. Later, I learned from tutorials that developers typically use Git for version control to avoid endless modifications affecting the original version—a concept I had never encountered before.
Moreover, AI programming isn’t very stable. With the same requirement, restarting from scratch produced very different versions. I attempted to recreate my preferred version three times from zero, but it never replicated the one I was most satisfied with.
So, while AI can indeed help someone with no experience write code, if you don’t understand development and debugging, you will still get stuck.
2. Recognizing the Need is More Important Than Rushing to Develop
Last year, I jotted down some development ideas. At that time AI programming wasn’t as advanced as it is now, and I didn’t want to spend too much time learning JavaScript from scratch, so many ideas were simply noted down.
It wasn’t until later in the year, when tools like Cursor and Windsurf appeared, that I felt I could give it a try. Even though AI had evolved, my attempts from December last year to January this year produced only average results, so I set the project aside for a while.
In hindsight, for a complete beginner this was a good strategy: record your ideas and needs, let the AI evolve further, and focus on more important matters in the meantime. Perhaps by mid-year many development challenges will have better solutions.
Recognizing the need is more important than hurrying into development. Needs emerge from everyday details; if you notice a tool is inconvenient, clunky, or requires constant switching, it indicates there's a need. For instance, my translation plugin was born out of my overseas customer support work.
Many independent developers now post products on Xiaohongshu, but truly impressive products are rare. Many still focus on the familiar trio of bookkeeping, journaling, and to-do lists, with most UI designs being unattractive—seemingly developed just for the sake of development rather than to serve a meaningful purpose.
3. Taste, Personality, and Values Are Harder to Replicate Than Code
Lenny asked Ivan, aside from the vision hidden behind productivity tools, what factors contributed to Notion’s eventual success. Ivan answered:
First, a product must have users and generate revenue—that’s the basic goal. But more importantly, it must be created based on your own values, meaning build what you believe the world needs. Everyone has their own taste and aesthetic, and ultimately you need to create a product that resonates with that.
Craft’s founder Balint expressed a similar sentiment:
Creating functional software has become increasingly easy. In the realm of personal software, what sets a product apart is its personality and uniqueness.
AI may eventually excel at handling functionality, but taste, personality, and values are elements that AI finds hard to replicate. The recent popularity of 61’s vinyl design is proof of that—its creator is a passionate vinyl enthusiast, which is reflected in the design.
A product is a manifestation of the creator’s own life experiences and perceptions. This is the core of product differentiation and is far harder to replicate than code. You must first establish your own vibe for Vibe Coding to truly have its own vibe. 🤔
Final
AI makes the world come alive, and the increasing availability of useful AI tools is something to be grateful for. Plugin development is just the first step; what I’m most looking forward to is refining my own app with AI’s help in the next year or two.
Reference:
Notion's lost years, near collapse during COVID, staying small to move fast, building horizontal
Design-first software engineering: Craft – with Balint Orosz
💎 Curated Gems
1. Personality is a Spectrum
In the past two or three years, MBTI has become a popular topic of conversation. It’s simple and direct—applying a label to quickly define yourself and gain social recognition. However, it emphasizes group commonalities while neglecting individual differences, and its simplicity may lead people to assume their personality is fixed.
In contrast, understanding personality through traits is more realistic and scientific. Traits aren’t fixed types but exist on a spectrum; everyone occupies a different position. For example, some individuals are more extroverted and enjoy social interactions, while others are higher in conscientiousness and lean towards planning and self-discipline. When traits—extroversion, conscientiousness, agreeableness, neuroticism, and openness (the Big Five personality traits)—combine in varying degrees, they create uniquely diverse individuals.
Furthermore, traits are not immutable, so personality isn’t a static label or category but rather an adaptable spectrum.
Reference:
We Are So Different: 40 Lectures on Personality Psychology
2. There Is No Standard Formula for a Successful Founder
Jessica Livingston, the co-founder of YC and Paul Graham’s wife, is YC’s only non-technical founder.
She is outspoken, nurturing, and empathetic—qualities that differ significantly from the typical image of a startup founder. Her distinctive traits have helped YC feel like a family, with her acting as the nurturing figure who cares about the founders’ well-being, ensuring they aren’t overwhelmed, take care of themselves, and offering advice on handling the strains of entrepreneurship. She patiently listens and helps resolve disputes and breakups with co-founders.
So if you’re considering a startup, I suggest you reflect on what makes you unique. What unique combination of abilities and interests do you possess?
And don’t change your answer, because, as my example shows, sometimes the most unexpected ingredient can be the key to success.
Reference:
3. The Nature of Work
At its essence, any job is about solving problems and delivering results, not merely completing tasks.
Wang Shimu once said that being a product manager requires creatively solving problems.
How can you solve problems creatively? By thinking analogically—for instance, imagining vinyl records integrated into the NetEase Cloud Music interface. And how do you form such analogies? It takes more than just practice:
- Observe daily life and be mindful of the problems around you.
- Master the fundamentals, such as data analysis, user research, and market analysis.
- Synthesize and deduce by gathering your observations and integrating them into product design.
Reference:
🧵 One More Thing
When you feel like doing nothing at all, transcribing with your keyboard can be quite fitting. Since transcribing lengthy content can be tedious, I opt to transcribe Luo Pang’s 60 Seconds: A Decade Collection, an experiment I started in 2023.
It allows me to learn new ideas and writing techniques (without the pressure to memorize everything) while also serving as a practice to improve my typing speed.