Dev Log 1: Maybe a Conversation with Myself

The other day I participated in Ark's crowdfunding beta. I counted the projects I'd written—at least thirty-plus that could run a demo at minimum. This morning I got a WeChat video-account dividend-earning notification; I looked and it was only ten cents, and it cracked me up. Not counting the cost of running large models, I doubt it covers the MacBook's electricity bill.

I told Claude that I have shipped at least five apps on the App Store — I guess that's a small accomplishment. By simple math, out of more than 30 projects, 5 have been published. This ratio is not great, but barely passing.

"You might be using fake busyness to project your own incompetence!" — that's how Claude weighed in.

The thrill of starting a new project really does beat maintaining an existing one with no clear path forward. I suppose it's the same reason I run ten kilometers every day yet never enter a race. There's a line from Encouraging Learning: "Without a thousand half-steps, one cannot reach a thousand miles." I want to use that line to justify myself, even though giving up comes from an instinctive wariness, telling myself, "Isn't this still accumulating project experience? Without starting all these one-off projects, where would all those insights, ideas, and usage tricks come from?"

Ha — what an indecisive person I am. Whenever something comes up, I always like to chat with AI, share my thoughts, try to clear the fog and see the daylight. But even if daylight is seen, it does not mean there is a road to somewhere far away beneath it. I remember last August, ChatGPT helped me do some research, saying that only 2% of indie developers barely scrape above the subsistence line; the rest have all become cannon fodder of the digital age.

I didn't believe it—I said I wanted to try. See, what a person of my own opinion I am! Even with the data in front of me, I'm still half-doubting—only a fool wouldn't fight for that 2%! "Nothing is as hard as it looks," I thought. Develop, ship, get paid—surely it's as simple as Old Wang's diner at the village entrance. So when ChatGPT made me a plan measured in years, I was almost furious; my whole head was filled with "it can't be that bad."

"Six months and we'll know the result!" — that was the goal I set for myself, with AI in my backpack.

"Nothing is as simple as it looks, either." Turns out most golden sayings have an unknown second half. Just like that, in a blink, it's almost two months away from making up a full year.

We often catch ourselves pondering "the meaning of being alive," and that very act of pondering is itself "meaningless." If you have the spare mental bandwidth to ramble like this, it means your life is actually "not all that meaningful." The normal state should be: my life is so full of meaning that I have no leftover mental space to wonder whether it's been worth it. It's like when you eat out — you only start analyzing the dish ("did they put in too much salt, was the base oil not cooked through?") when it doesn't taste good. If it's so good it blows your mind, you'll probably just think: "Holy shit, this is fucking amazing!" Or another version: "Wow, that smells incredible!" Or perhaps: "This flavor belongs in heaven!"

So, by the same logic, those who eagerly open courses claiming "master in 7 days from zero foundation" probably haven't learned that deeply themselves. Because if they had truly mastered it, they wouldn't need to teach to make money; if they've fallen to teaching to make money, there's no need to say more—there must be some constraint or predicament forcing them to abandon proper work and take a sidetrack. For me, starting to write articles is, on the one hand, an attempt to find a sidetrack and see if I can squeeze in even for a glimpse; on the other hand, it's also using the process of writing down to summarize the life that's already passed, to pull it out and reflect on it. After all, "an unexamined life is not worth living."

So inevitably, we have to talk about development tools. This is the simplest part — a matter of multi-dimensional comparison and personal user experience. It is also the hardest. "Welcoming guests from all over, catering to the tastes of all beings" — it's hard to be fact-based and impartial, free of any private bias. But this is also the most important part. Since this is a "development diary," not talking about development tools would mean not only a scattered "form," but also an unfocused "spirit," wouldn't it?

I have always prided myself on being a "literati," but I find it too embarrassing to say outright; if I switch to "cultured person," it's clearly miles off the mark, like the joke of "soy sauce isn't oil." Having watched various calligraphy and painting videos in passing, I keep remembering something called "literati painting." I might as well boldly follow the trend and call myself a "literati handle" (文人号). That way, I can smoothly avoid putting any tables, code, or things only an AI is willing to read and understand into my paragraphs. Then, turning back to talk about developer tools, the psychological burden and guilt of "not being professional enough" can reasonably vanish into thin air — or at least be hidden away, like someone with too dark a complexion and an unlovely face must wear heavy makeup to attend a party, and only dares to show her face under dim lighting.

At this point, I imagine many of you are starting to get a little hot under the collar. The patience built up by the bonds of friendship is almost gone, yet I still haven't gotten to the main topic and talked openly about development tools. "Hey, you've seen right through me!" I have to put on a thoughtful look, soften my tone — like a paper plane gliding steadily through the wind, low to the ground, then brushing along the surface — and, with no trace and no embellishment, drop a line like that.

Because I have a view.

If I state this view plainly, it'd be too flat and wouldn't show the苦心 at all; or it'd be so vague as to seem perfunctory. The first half of this view is: use whatever works.

Hold off on hitting me—let me analyze. If a tool is in heavy circulation in the market, and it is not because of its own intrinsic value or other benefits it provides, then when the masses rush to adopt it for fear of falling behind, are we supposed to conclude that "the masses are idiots"? If we can't say that, then we can say the tool has been validated by the market and already possesses the basic attributes needed to deliver its value. In other words, if a supermarket stocks five brands of soap, the reason you're torn over which to pick is not whether the soap can wash hands—since that base function is shared—but whether, beyond "washing hands," there are other distinguishing extras, and whether those extras happen to make it "different" for you.

So, the question you need to think through is not: "Which of Cursor, WinSurf, Codex, Claude Code... is actually better?", but: "Among these tools, who provides the most attractive and valuable non-development benefits for me?" and at the same time: "Within my budget, who is the cheapest?" This is the second half of my point: "Choose whichever you think is the most worth it."

By general reasoning, if a tool is meant "for writing code", what it presents to the user is by no means limited to that single function of "writing code". Because in marketing, "writing code" is a commodity — like soap being used to wash hands, it's a basic attribute; a product with only basic attributes can only be called a "generic product". On top of that basic attribute, to differentiate their own product, vendors rack their brains to tack on extra attributes that "they think the user needs" — once this layer is added, that's the user's "expected product". At this point, looking back at the makers of these "development tools", each and every one is plotted and orchestrated by a group of the world's top minds behind the scenes, so the product will extend further to become an "augmented product". And when competition heats to a white-hot stage, it can even fully evolve into the "potential product" stage.

Naturally, this is a very immature analysis after reading The Imagination of Marketing; the words used here do not carry their customary meanings. I want to use a metaphor to summarize this analysis: a colleague (product) who has passed through rounds of interviews (market screening) and finally stands before you — extreme cases aside (such as unfair competition) — must be an employee with basic on-the-job capability. As for his performance in the company, it depends on the personal strengths that set him apart from others, and on the proportion of those strengths that the company environment allows him to exercise. If the position happens to favor creative talent, and he is also an imaginative employee, they click immediately, the position and the employee help each other and grow together. But if the position demands stability and the employee loves adventure, the only outcome is a parting of ways. The employee feels the company has treated him unfairly; the company feels the employee's ability is lacking.

Does this analogy explain my "dividing into two" view?

If you hold this view, when you feel a particular tool is especially awkward to use, you should first remind yourself that it's not that you lack ability — and further, you should conclude that the marginal utility the tool provides is not the "extra property" you expected.