Anyone can build a platform now. Almost nobody can get people to find it. · ClaudeFolio

I want to start this post with an admission that’s a little uncomfortable, which is that the platform I built specifically to help other people solve their distribution problem is itself struggling with exactly the same distribution problem I built it to solve. ClaudeFolio is live, the submissions feature works, the upvoting works, the design is clean, and the rules are exactly the rules I wanted them to be. The only thing missing is the part where lots of people know it exists. And that part, as I’m learning the hard way, is the part that actually matters. Looking at Google Analytics I see an average of 50 visitors a day and just a handful of signups.

This is the post-AI reality for solo founders, and I think it deserves more discussion in the current wave of “look what I built in a weekend” content. The tools we have now genuinely let one person ship things that used to require teams, and there’s something legitimately exciting about that, but the celebration of how easy building has become has crowded out a much harder conversation about what happens after you ship? The truth is that anyone can build a platform now, and almost nobody can get people to find it.

The bottleneck moved

For most of the history of software, the bottleneck for solo founders was building the thing in the first place, which is why so much of the founder advice from the last twenty years was about how to build, what stack to use, how to manage your time, and how to avoid the technical traps that killed amateur projects. That advice was correct for its era because building genuinely was the hard part, and once you had something working you usually had a reasonable shot at distribution because the supply of working software wasn’t yet overwhelming the demand for it.

That world is over now, and it ended faster than most people noticed. With Claude Code and the broader generation of AI coding tools, the cost of producing a working platform has collapsed by something like two orders of magnitude, which means the supply of new platforms has exploded by a similar factor while the supply of human attention has stayed roughly fixed. The math of that shift is brutal if you think about it carefully, because you don’t just have more competition for the same audience, you have exponentially more competition for an audience whose tolerance for being marketed to has been steadily decreasing across the same period.

What this means in practice is that the bottleneck for solo founders has moved from “can you build it” to “can anyone find it,” and the founders who don’t recognize this shift end up doing the same thing I did, which is building something that might be useful and then watching it sit there in the digital equivalent of a desert with nobody walking past. The technical skills that used to be the differentiator are now the table stakes, and the actual differentiator is whatever makes you able to reach the people who would benefit from what you built. Most solo founders, including me, are radically underprepared for that part of the work.

What growing a site looked like 18 years ago

I have a useful comparison point here because I co-founded allkpop over 18 years ago, when the internet was a fundamentally different environment than the one solo founders are launching into today. The thing I want to be clear about up front is that allkpop didn’t succeed because I was some marketing genius, it succeeded because the conditions of that era made it possible to grow a site in ways that simply aren’t possible anymore, and I think it’s worth laying out what those conditions actually were so people understand why distribution is harder now in concrete terms rather than as a vague feeling.

The first thing that was different was the level of competition, where the supply of content sites was a tiny fraction of what it is today and any niche you cared about probably had only a handful of serious players in it rather than the dozens or hundreds you’d find now. When we started covering Korean pop music, there was almost nobody else doing it in English at any kind of serious level, which meant that anyone who was searching for that content basically had to find their way to allkpop eventually because there weren’t a hundred alternatives. That kind of empty niche just doesn’t exist in most categories anymore, and the few that still exist tend to get filled within weeks of being identified rather than persisting for years the way niches used to.

The second thing that was different was that social media was still in its infancy, where Facebook was new, Twitter was newer, and the algorithms that now decide whether anyone sees your content or suppressing news articles were either nonexistent or much more forgiving than they are today. You could post a link in a forum or a Facebook page or even a comment thread and have a real chance of people clicking it and showing up at your site, because the platforms hadn’t yet figured out that they could capture more value by keeping users on their own properties rather than letting them follow links elsewhere. The web was still a network of connected sites in a meaningful way, not a series of walled gardens that punish you for trying to send people to your own domain. It was also much easier to rank on Google for keywords that didn’t have much competition.

The third thing that was different, and this is the one most relevant to the post-AI conversation, was that the technical barriers to running a real website were genuinely high and acted as a natural filter on competition. To have a serious site you needed to know how to manage servers, configure databases, write code, handle security, and keep everything running through traffic spikes, and most people who had ideas for sites simply couldn’t ship them because the technical lift was too much. That gatekeeping was bad for the people who got filtered out, but it was good for the people who made it through, because the field was thinner and the audience could actually find you once you’d shipped. The collapse of those technical barriers, which is exactly what Claude Code and the rest of the AI coding tools have done, is the same thing that’s making distribution so much harder now, because the filter that used to do most of the curation work for the audience is gone.

If I tried to start something like allkpop today with the same idea and the same level of effort, I’m not sure if I would succeed, and I want to say that plainly because I think it’s the honest assessment of what’s changed. The empty niche wouldn’t be empty, the social media platforms would be hostile to outbound links from a no-name site, and the technical barriers that used to filter out my competition would be gone, which means I’d be one of hundreds of people trying to cover the same content with the same tools and roughly the same quality. Whatever advantages we had back then were largely advantages of timing and circumstance, not advantages of skill, and the fact that we built something that worked under those conditions tells you almost nothing about whether I could repeat it today. That’s a humbling realization for someone who built a career on that earlier success, but it’s also the realization that makes the rest of this post matter, because if I’m going to launch new things in the current environment I have to actually develop the distribution skills that the old environment let me skip.

Why distribution is so much harder than building

The thing nobody tells you about distribution is that it’s the kind of work that doesn’t scale linearly with effort, where you can spend a hundred hours and get nothing, and then spend two hours doing the right thing in the right place at the right time and watch a flood of users show up. That non-linearity is what makes it so frustrating compared to building, because building has the feedback loop most founders are wired for, where you write code, you run it, it either works or it doesn’t, and you fix it and move on. Distribution doesn’t work that way, and the absence of clear feedback loops is the thing that drives most people to give up on it before they figure it out.

The other thing that makes distribution hard is that it requires a different set of skills than building, and most technical founders genuinely don’t have those skills and aren’t naturally inclined to develop them. Distribution is about understanding your audience deeply enough to know where they hang out, what they care about, what language they use to describe their own problems, and what would compel them to try something new. None of those questions have answers in the documentation. None of them can be solved by writing better code. And most of them require you to do things that feel uncomfortable for technical founders, like writing for non-technical audiences, posting on platforms you don’t enjoy using, or actually talking to strangers about what you built.

There’s also a self-defeating loop that hits a lot of solo founders in the distribution phase, where the discomfort of marketing makes you want to retreat back into building, so you start adding features instead of finding users, and you tell yourself the next feature is what’s going to make the difference. It almost never is. The platforms that succeed are usually the ones that found their audience first and added features second, while the platforms that die are usually the ones that kept polishing what they had instead of fighting for the first hundred users who would tell them what to actually build. I’ve made this mistake repeatedly and I’ll probably make it again, because the gravitational pull toward “just one more feature” is genuinely strong for builders.

The launch platforms aren’t enough

I’ve already written about why Product Hunt and the other launch platforms are largely closed to founders without an existing network, so I’m not going to cover that ground again here, but I want to make a related point that goes deeper than the specific gripes about hunters and waitlists. Even if the launch platforms were perfectly fair, they wouldn’t actually solve the distribution problem for most solo founders, because a single launch event is a one-time spike of attention rather than a sustainable source of users. The founders who win on Product Hunt and similar platforms tend to be the ones who already had distribution figured out before they launched, and the launch is just the moment they cashed in on that work, not the moment they built it.

The real distribution work is the slow compounding kind, where you write content that gets indexed, build a presence in communities your audience actually uses, develop relationships with people who can vouch for you, and create reasons for users who do find you to bring others. That work is unglamorous and slow and doesn’t produce the dopamine hit of a launch day spike, but it’s the work that actually builds platforms over time. Solo founders who treat distribution as something you do for one day at launch end up confused when their traffic graph drops to zero on day two, while founders who treat distribution as something you do every day for years end up with platforms that grow on their own.

The other problem with relying on launch platforms is that they create a single point of failure that’s outside your control. If Product Hunt’s algorithm changes, if your post gets shadow-flagged, if a hunter you were counting on backs out, your launch can collapse and there’s nothing you can do about it. Distribution channels you build yourself, by contrast, are channels you actually own, and they keep working whether or not the platforms-of-the-moment are paying attention to you.

What I’m actually doing about it

Since I’m being honest in this post about the fact that ClaudeFolio is itself struggling with distribution just a few weeks after launch, I should also be honest about what I’m actually doing to fix that, partly because I think specifics are more useful than abstractions and partly because writing it down keeps me accountable to the plan. The first thing I’m doing is content, which is what you’re reading right now, because long-form writing on a topic the audience actually cares about is the one form of marketing that compounds over years rather than evaporating on launch day. Each post on the ClaudeFolio blog is a small bet that someone searching for an answer to a real problem will find their way to my site, read something useful, and then notice the platform itself.

The second thing I’m doing is showing up in communities where the audience already exists rather than trying to manufacture an audience from scratch, which means engaging genuinely on Hacker News, Reddit subs that focus on solo building and indie hacking, and the various Claude Code communities that have formed organically over the last year. The trick with community participation is that it has to be actually genuine, where you contribute to other people’s posts and conversations more than you promote your own, and the platform exposure comes naturally as a byproduct of being someone known in those spaces. This is slow work but it’s also the kind of work that builds the relationships launch platforms can’t replace.

The third thing I’m doing is making the platform itself more shareable, which means making sure that the projects featured on ClaudeFolio are good enough that the people who submit them want to share their submissions, and that the people who find them want to upvote and pass them along. Distribution and product quality are linked in ways that get underrated, where a platform that’s actively useful to a small group will spread organically inside that group, while a platform that’s only kind of useful won’t spread no matter how much you market it. The platform has to deserve the attention before the attention can do its work.

The fourth thing is that I am trying out marketing ads on social media and that includes posting ads on Facebook and Reddit. I’m doing a smaller ad spend at first to see if it is worth spending extra money on advertising on certain platforms.

The fifth thing, and this is the one that’s hardest for me personally, is being patient enough to let any of this work. Distribution timelines for solo founders are genuinely measured in months and years rather than days and weeks, and the temptation to declare a project dead because it didn’t take off in the first few months has killed more good platforms than any technical problem ever did. ClaudeFolio is a few weeks old as I write this, and the honest truth is that I have no idea whether it’s going to work, but I do know that giving up on it before the slow distribution work has had time to compound would guarantee it doesn’t.

What this means for you

If you’re a solo founder reading this and you’ve shipped a platform with Claude Code and you’re now staring at a flat traffic graph or 0 users on Google Analytics wondering what went wrong, the first thing to know is that nothing went wrong, you just hit the actual hard part of the work and the hard part is supposed to be hard. The second thing to know is that the way out of it isn’t to add more features or to launch on more platforms, it’s to develop the distribution skills that the building phase let you skip. That’s a slower and less satisfying path than continuing to build, but it’s the actual path, and the founders who accept that early are going to outlast the founders who keep refusing to. If you don’t have the skills, then pay money for advertising or hire someone who has distribution skills.

The third thing, which I think is the most important and the least talked about, is that distribution skills compound across projects in a way that building skills don’t. Every platform I’ve built with Claude Code basically resets to zero on the day it launches, because the technical work doesn’t carry over to the next project’s audience. But the distribution skills, the audience relationships, the writing habit, the community presence, those carry over completely. A solo founder who spends the next year learning distribution will be radically more capable of launching their fifth platform than they were of launching their first, while a solo founder who spends that same year building their fifth platform without learning distribution will be in the exact same position they were in twelve months ago, just with five graveyards instead of one.

Some of my platforms have 0 users while I have some that are now reaching 30,000 monthly active users. The oldest platform is two months old while the newest one is just a few weeks old. Each one is different and each grew at their own pace.

None of this is easy and I don’t want to pretend it is, because if it were easy I wouldn’t be writing this post about my own struggle with the same problem. But the framing that helps me keep going is that distribution is just another skill, and like any skill it gets easier with practice, and the only way to get practice is to actually do the work even when it feels slow and uncertain. ClaudeFolio might or might not work out, but the distribution muscles I’m building while trying to make it work will carry over to every platform I build after this one, and that’s the actual long-term bet a solo founder is making in this era. Not on any single platform, but on the skill of getting platforms found, which is now the rare and valuable thing.

If you’ve built something and you’re feeling stuck in the same place I’m in, you can also just put it on ClaudeFolio. That’s a small thing but it’s also the thing this post is ultimately about, which is that we all need to be helping each other get found, because the era when the launch platforms would do that work for us is over and the only thing that’s going to replace it is the founders deciding to look out for each other. Submit your project, upvote things you actually find useful, share the platform with one other builder you know. None of that is going to single-handedly fix the distribution problem, but enough of it stacked together over time might start to.