From Zero to One Million Downloads: Logan’s Journey as a Solo Developer
Guest | Logan Yang (Copilot for Obsidian)
Interview | Xiao He
From Big Tech Engineer to Going Solo
Xiao He:
Logan, thank you so much for joining Mother of Success. Your product, Copilot for Obsidian, has already reached over one million downloads — that’s truly incredible. I’d love to use this opportunity to talk about your day-to-day life as a solo developer, and how it feels to transition from being a Machine Learning Engineer at a big company to becoming an independent builder.
Logan Yang:
Sure. I’ll probably just talk as things come to mind.
I think, at a very fundamental level, I’m just not well-suited for working at big companies. I’ve never really cared about job levels or promotions. Every time I see career ladders or performance rubrics, I’m just completely confused — I genuinely don’t understand what they’re talking about.
What I can feel very clearly is how users experience the things I build. That, to me, feels real.
So from the beginning, I’ve preferred making things on my own and having direct communication with users.
At the same time, I’m not a traditionally trained engineer. I studied physics in undergrad and only later transitioned into engineering. A lot of what I learned was self-taught after coming to the U.S. Compared to people with a CS background who grew up in the Bay Area system, I’ve always been very aware that I have gaps.
And the longer you work, the more obvious those gaps become.
For example, when it comes to promotions, companies care a lot about your understanding of systems, your grasp of low-level fundamentals, whether you can clearly explain complex technical concepts. Going from entry level to L4 or L5, you’re expected to write a lot of documentation and deeply understand the systems you’ve worked on.
But that’s not me. I never took most undergraduate CS courses. What I learned was mostly in service of building a specific feature, not understanding technology for its own sake. Things like operating systems — I honestly don’t understand them that well.
So to be honest, I often felt that I wasn’t a very “qualified” engineer, especially at a place like Google. And the longer you stay, the higher those expectations become.
Considering Other Paths (and Why They Didn’t Fit)
I did explore other options, like switching into product management.
But when I looked at what PMs at big companies actually do, it didn’t feel like what I wanted either. They spend most of their time in meetings and need extremely strong communication skills, emotional intelligence, and what people call “influence without authority.” You’re not anyone’s manager, but you still have to push things forward through persuasion.
And PMs at big companies focus heavily on metrics. That skill set is very different from what’s needed at a small company, as an indie developer, or as a research-driven startup founder.
Gradually, I developed a vague but increasingly clear sense that big-company life just wasn’t for me. I didn’t care much about promotions, and I didn’t want to go back and fill in all those core CS fundamentals. I didn’t want to be a scientist or a hardcore engineer.
What I really wanted was to build things that worked and that people genuinely enjoyed using. That reward mattered most to me.
Burnout, Quitting Without a Plan, and Discovering GPT
By early 2023, partly by chance, I was completely exhausted at my previous job — truly burned out.
When I quit, I had no plan at all. I just wanted to rest.
Around that time, GPT4 came out. I started playing with it and was completely blown away. Its ability to write code was incredible. I began using it to build small projects purely for fun.
I’ve always been someone who loves taking notes. I used Apple Notes at first, then Roam Research, and later Notion. But Notion became very slow once I had too many notes, since it’s web-based and always needs time to load.
Then I discovered Obsidian by chance. It’s local, basically a Markdown editor, and extremely fast. I migrated all my Notion notes into Obsidian.
Because I was constantly chatting with GPT, I found myself switching back and forth between Obsidian and ChatGPT all the time — which was annoying. I wondered: could GPT live inside Obsidian, maybe as a sidebar?
The problem was, I’m not a frontend engineer. I had no idea how to build something like that.
So I just asked GPT.
“GPT Basically Wrote the Plugin”
At the time, I didn’t even know Obsidian had a plugin system. Later I learned that anyone can write a plugin and submit it to the community plugin store. Once it passes review, it can go live.
This plugin was basically written by GPT.
I figured I might as well submit it — there was no real cost. And since I was early, I named it “Copilot.” It’s a very plain, almost cheesy name, but I thought that claiming a simple name early might bring some built-in traffic.
After the plugin passed review, I noticed downloads starting to increase. People began opening issues on GitHub. Slowly, I realized that real users were actually using this thing.
That’s when I decided to keep going.
To be honest, I’m not someone who’s obsessed with clean or elegant code. Many engineers care deeply about whether code is pristine or strictly follows best practices. I don’t.
I care about one thing: does it work? Do users like it? Do I wake up tomorrow feeling less stressed?
As long as I can read it and users are happy, that’s enough for me.
From 2023 until now, I’ve written very little code by hand. I see hand-writing code as a very primitive behavior.
Back then, many engineers I knew were mocking AI-generated code, saying it was weird or stupid. But now it’s 2026 — and this is clearly the direction things are heading.
VC, the Silicon Valley Model, and Why Bootstrapping Won
Xiao He:
I’m really curious — did you ever seriously consider the classic Silicon Valley startup path? Finding a co-founder, raising venture capital, scaling aggressively?
Logan Yang:
That’s a great question — and honestly, it’s a kind of soul-searching question that almost everyone runs into at some point.
More and more, I feel that any path you choose has to come back to who you are as a person. I’ll start with the conclusion: I don’t think I’m particularly suited for the VC-funded Silicon Valley model.
When I look at people around me who raised VC and followed that path, they share one very obvious trait: extremely high energy. In Silicon Valley, this almost feels like a prerequisite. No matter how old you are, people work with the intensity of someone in their twenties — pulling all-nighters, sprinting hard, shipping things in very short timeframes, and then broadcasting it all online. That culture is highly celebrated here.
But that’s just not who I am.
For years, I’ve been quietly observing a different path on Twitter. The people on that path tend not to be high-energy. They don’t want to build something in a few months — or even a single month — and then immediately start pitching VCs like crazy.
Both paths have their advantages. But my self-awareness tells me that I don’t want to be pushed forward by external forces. I want to move at my own pace.
VC money can absolutely be a shot of adrenaline. Early on, it can help you launch faster, trade money for users or exposure. But for me, it would consume a huge amount of my time and energy — and more importantly, it would take away my freedom.
VC money is never free money. It comes with very clear expectations: extremely high returns. Ten times isn’t enough — it has to be a hundred times. Once that growth expectation exists, it inevitably shapes your product roadmap and your daily decisions. You might end up doing things that aren’t very friendly to your existing users just to reach a much larger market.
That kind of shift impacts your relationship with users and your personal brand.
So until I reach a point where I’m convinced that this product simply cannot work without VC funding, I’ll almost always prefer to bootstrap.
Bootstrapping gives me more freedom, more control over my time, full ownership, and a much more direct relationship with users.
What a Day Actually Looks Like
Xiao He:
What does your life look like now? What’s a typical day for you?
Logan Yang:
It feels kind of surreal.
When I was working a regular 9–5, I always had ideas in my head. But after a full workday, I’d get home with no energy left. Over time, I realized I’m just not the kind of person who can do a second shift at night. After ten years, that became very clear.
After quitting, I finally had the time to work on things without being constantly stressed.
That said, I’m not especially disciplined. My rhythm is more like this: I’ll suddenly feel extremely motivated for a few days, work intensely for three to five days, ship a lot of code. Once things work and user feedback looks good, I might rest for an entire week and do nothing.
That kind of rhythm was impossible before. Now, no one is telling me when something absolutely has to be finished.
So my life is basically cycles of intense sprints followed by rest — and then another round.
Building in Public: YouTube, Twitter, and the Reality of “Distribution”
Xiao He:
You post product updates, demos, and progress on YouTube and Twitter. How do you integrate media work with development? Because beyond coding, there’s also a lot of admin work.
Logan Yang:
This part is actually quite hard for me.
I’ve noticed that many people who write code will get excited about an idea, build it, and then stop. There’s an implicit belief that “good products will naturally be discovered.” But that’s not how reality works.
On the other extreme, some people say you shouldn’t build anything first — just make a landing page, collect emails, sell something that doesn’t exist yet. That approach is also hard for many people to accept. It can feel like you’re deceiving others.
I don’t think either extreme is ideal.
What I do now is something in the middle: the product already exists, so I use demos and videos to show it honestly. At certain stages, the value of making videos and doing distribution can actually be greater than writing code itself.
But since I’m still the main person writing code — and the few others helping are only part-time — I still have to prioritize development.
I have a Discord community for paid users. They often ask when the next video is coming out, and I sometimes feel a bit embarrassed because my update frequency really isn’t that high.
From “No One Tracking It” to One Million Downloads
A Non-Linear Growth Story
Xiao He:
Going from zero to over a million downloads must not have been linear. Looking back, how would you break down the stages? Community plugin first? Word of mouth? YouTube growth?
Logan Yang:
Honestly, if you ask me for a very precise answer, I don’t have one — because I never tracked user acquisition very carefully.
From my subjective experience, the growth definitely wasn’t linear. If you zoom into a short period, it might look linear. But if you zoom out over time, there are moments where it suddenly takes off.
In the first year, I did almost no promotion. I barely posted YouTube videos. Most of the traffic was organic. Some power users discovered the plugin and wrote about it on Medium or their blogs. I even saw platforms I’d never heard of publishing content about Copilot for Obsidian.
One important reason is that Obsidian itself is a huge river. I wasn’t digging a tiny stream from scratch — I was branching off an existing river.
For indie developers — especially those who aren’t good at promotion or storytelling — this is an excellent starting point. It solves the cold-start problem.
Even today, when I look at website referrals, the biggest source is still Obsidian’s own plugin store.
Later, when I launched the paid version, I posted more videos, and there was a noticeable boost. The slope of the growth curve became steeper.
From Free to Paid
Xiao He:
When did you start feeling that it might make sense to introduce a paid version?
Logan Yang:
This part of the story is actually pretty interesting.
At the beginning, I never thought about charging at all. There were a few reasons for that. First, most Obsidian plugins are open source. To submit a plugin, your code has to be public. There’s no backend, everything runs locally on the user’s machine. Under those constraints, how do you even charge? Technically, it’s extremely difficult.
So at first, I treated it purely as an open-source project. I had GitHub Sponsors, and people would occasionally send me some money — like pocket change. That already felt pretty good to me.
Then one day, a user called me and told me that a competing plugin had started charging. He showed me a WeChat public account article saying that the competitor made USD $70,000 in their first month.
I was completely shocked. I genuinely couldn’t imagine how an open-source, local plugin could be monetized.
I went and researched it, and eventually realized that the only reasonable way was to introduce a backend service. In other words, most features stay free, but certain features — like web search or YouTube transcription — require server support.
That was the moment I realized that charging was technically feasible.
I’m not a superstitious person, but I’m very sensitive to “signals.” When certain signals appear, it can feel like the universe is nudging you, saying: maybe you should try this.
So I decided to give it a shot.
Turning on Payments (and Going to Sleep)
I wired up the paid features, put up a landing page, and said nothing.
No tweets. No videos.
Then I went to sleep.
The next morning, when I woke up, there were already several paying users — spread across different countries. The very first paying user was from Singapore, and they immediately bought the most expensive plan.
I was honestly stunned.
There’s also a funny follow-up to this story. Later, I tried to find that original WeChat article again, wanting to verify the “$70k in the first month” claim — but I couldn’t find it anywhere on the internet. I now strongly suspect that signal was fake.
But that doesn’t really matter anymore.
What mattered was that it triggered something I already believed deep down: this was worth trying.
Promotion, Channels, and Conversion Rates
Xiao He:
That’s amazing — it started completely organically, with people paying on their own. After that, did you actively promote it anywhere? Did you observe a relatively stable conversion rate?
Logan Yang:
After enabling payments, I started posting more YouTube videos. During that period, the update frequency was higher, and there was definitely a response.
YouTube views are relatively reliable. Every time I post a video, some people will watch it — it’s never a situation where no one sees it at all. I can really feel that difference.
I actually have another YouTube channel, so I know how hard YouTube traffic can be. If I post scenery videos or so-called “aesthetic organization videos,” basically no one watches them. So I know very clearly that YouTube is not a platform where everything gets views by default.
But when I post product-related videos, people do watch them. That tells me there’s already an audience base. Those videos definitely bring in additional traffic.
At this stage, I genuinely believe YouTube is an extremely important platform for solo founders and indie developers.
Twitter (now X) is also important, but it’s very different from YouTube. Content on X is short, high-volume, and consumed quickly — similar to Xiaohongshu. Traffic can spike fast, but it doesn’t always reach the right users. That depends a lot on the algorithm and whether you’re good at funneling attention.
And yes, I’ve observed a fairly stable pattern: a fixed percentage of people are willing to pay. Among website visitors, I see a relatively stable number — roughly 8 to 9 out of every 100 users convert into paying users.
Compared to many other services I’ve looked at, that’s actually quite high.
Pricing, Lifetime Plans, and Why People Pay Upfront
Xiao He:
I’m really curious about your pricing strategy. Did you research competitors, or did you just have a number in mind? Have you adjusted prices over time?
Logan Yang:
This was actually quite stressful for me at the beginning, because I had zero experience.
My initial intuition was simple: what I’m offering is, in some sense, similar to ChatGPT. For many users, the upper bound of AI usage is GPT itself. I knew very clearly that I couldn’t charge more than GPT’s subscription. So from the start, I knew my monthly price had to stay below a certain range.
At the same time, I offered a one-time payment option — a lifetime license. That was even harder to price.
Because the lifetime plan includes model tokens. In other words, you pay once, and you can keep using the model indefinitely. For many people, this is extremely counterintuitive. Models aren’t free. Tokens aren’t free. How can you allow lifetime usage without a subscription?
But I still did it.
Of course, I wasn’t completely ignoring sustainability. I chose a model with relatively low token costs — a compromise. Even if a user is still using it three or four years later, it won’t completely destroy my cash flow.
My pricing logic was actually very simple: I assumed the product’s relevance would be around three years.
Why three years? Because AI is changing too fast. I don’t really believe that, as a solo developer, I can keep a product at the core of people’s workflows for longer than that. The paradigm might shift entirely.
So I worked backward from “three years of usage cost” to set the lifetime price.
It’s a hypothesis — and a bet. I don’t know what will happen in the future. But so far, after more than a year, I think the decision has been reasonable.
Thinking Beyond the Product
Xiao He:
I really admire this — even while the product is doing well, you’re already thinking about technological shifts three years out and whether your product will still be relevant.
Logan Yang:
I struggled with this for a long time.
But I think all businesses involve bets, especially pricing. Your assumptions directly determine how you price today.
My current assumption is that this product will remain relevant for a few years. But no one knows what will happen after that. In 2023, people were still mocking AI-written code. Now, in almost every company, a huge amount of code is written by models.
The pace of change is incredibly fast.
So on one hand, I keep iterating on this product. On the other, I’m thinking about how to turn this into something more long-term.
The biggest asset this product has given me isn’t the code — it’s the user base. Users’ trust in me is something I can carry into the next product.
Users, Discord, and “Completely Beyond Expectations” Engagement
Xiao He:
How do you stay connected with your users now? Is Discord the main channel?
Logan Yang:
Yes. Paid users are mostly on Discord. Free users are more scattered across GitHub, Twitter, and YouTube.
At the beginning, I was actually a bit intimidated by Discord because I hadn’t really used it before. But looking back now, it’s probably the most important place we have.
What really shocked me was how high the level of user engagement is. Sometimes we ask very open-ended questions — like whether a certain feature should be built, or which direction the product should take. Users respond incredibly quickly, and their replies are long, thoughtful, and clearly well-considered.
This level of participation is something I never experienced at a big company.
Who Are the Users?
Xiao He:
Do you have a rough sense of who these users are? Is there a clear profile?
Logan Yang:
We’re still slowly collecting data, but from observation: the largest group is in the U.S., followed by Japan, South Korea, China, Germany, the U.K., and Australia.
In terms of professions, it’s very diverse. Broadly, there are knowledge workers and students.
Among knowledge workers, there are people in finance, accounting, healthcare — including doctors and medical students. There are also many researchers, PhDs, and people who need to do a lot of reading and writing.
There’s also a group I never expected at all: older users — people in their 50s and 60s. Some of them even use it for Bible study.
If I hadn’t built this product, I would never have crossed paths with these people. But now I can clearly feel how strong their desire is to learn AI.
The sense of fulfillment I get from this far exceeds what I ever felt as a small cog inside a big company.
Small Team, Profit Sharing, and Why Not Equity
Xiao He:
Right now, the product is still mostly built by you, but you’ve mentioned that a few part-time collaborators help out. What roles do they play, and how do you work together?
Logan Yang:
We met online — through forums and communities.
They’re all engineers by background. Some worked at big companies before, some are now in startups, and they’re all interested in exploring side projects. They’re particularly drawn to web and productivity tools.
Our collaboration model is very simple: profit sharing.
Each month’s revenue is split by percentage. The more modules you’re responsible for, the higher your share. I think if you want people to truly commit, you have to share the upside.
One of my critiques of the traditional VC startup model is that early engineers take on a lot of risk, but often receive far less ownership than the founder.
From an expected return perspective, that doesn’t always make sense. At that point, you might as well build something yourself.
Of course, solo building has its own challenges. But if I collaborate, I prefer profit sharing — it’s direct and transparent.
The Hardest Part Isn’t Coding
Xiao He:
Looking back on your journey as a solo or indie developer, what were the one or two hardest parts?
Logan Yang:
The hardest part was breaking through my own weaknesses.
Writing code itself is fine for me — at least I can make things work. But promotion and sharing have always felt very unnatural. There’s often a voice in my head questioning myself: Am I exaggerating? Am I doing something weird or embarrassing?
I know that’s my own issue, and something I need to overcome.
Another challenge is scale. The idea of a “one-person company” is very popular, but once things reach a certain size, one person is never enough.
What I believe in more is a small team made up of several “super individuals.” Each person is highly capable of using AI. Three or four people can do the work that used to require more than ten.
That kind of structure is still being explored — there’s no standard answer yet.
Looking Back: The Greatest Reward
Xiao He:
If you look back on the past few years, what do you think your biggest gain has been?
Logan Yang:
Freedom.
I don’t have to work nine to five. I can rest when I need to rest. For me, that’s an incredibly luxurious thing.
I also know it might not last forever, which is why I try to cherish this phase.
Final Recommendation
Xiao He:
We usually end our interviews by asking guests to recommend a book, film, or podcast. What would you recommend?
Logan Yang:
Not something about entrepreneurship.
I’d rather recommend things that broaden perspective and relate to life. Recently, I’ve been reading The Sahara Stories by Sanmao — her life in the Sahara and in Africa, and her stories with José.
I recommend it because I think people — especially those busy with work or startups — can treat it as a form of rest. It’s a way of seeing alternative lives.
After all, there’s nothing in life that you must do. The path you choose, no matter how unconventional, doesn’t require self-doubt or guilt.
From time to time, reading about people who are even stranger, braver, and more unconventional than I am is deeply encouraging.
Xiao He:
That’s wonderful. Thank you so much, Logan, for sharing your story.