1.25M Users, A Few Hundred Paying. What's Wrong?

Key Takeaways

• A million users who pay for perks is a fan club, not a business. Charge for core value, keep the existing product free, grandfather the loyal.
• A one-person vendor doesn't win by being the best. Show the buyer a gap they didn't know they had, then be the only one who can fill it.
• Customers asking for features are describing jobs to be done in the only vocabulary they have.
• Solo founders selling to serious buyers: don't be the best, be the only. The doubt is mostly in your head.
• For a solo hardware founder, the answer to "how do I split my time?" may be: don't. Ship the mind before the body.

Listen: Apple Podcasts | Spotify

Introduction

Utkarsh Dalal's app has 1.25 million users in every country in the world. GameNative lets you play the PC games you already own, from Steam, Epic, and GOG, natively on an Android phone. It's open source. It has ten thousand GitHub stars and a Discord in the tens of thousands, all organic. And a few hundred people pay for it, at five or ten dollars a month, in exchange for a Discord role and early access to builds.

He brought that gap to Julian as a monetization problem. Julian didn't think it was one.

"1.25 million, and only a couple hundred people paying you. It shows that you haven't yet offered something that's actually worth paying for."

Utkarsh was the first of four solo founders who brought a live, unsolved problem to office hours: a gaming app, an AI-research vendor, a wholesale-ordering startup, and a robot company. Four industries, and the same diagnosis surfaced every time: each founder was building the thing they wanted to sell, not the thing the customer was already telling them they'd pay for.

The Fan Club Problem

The subscription GameNative sells today, Julian pointed out, is a Patreon. A role on Discord, priority support, early builds. People buy it to show support, not because it makes monetary sense. That is the wrong shape for a product with 1.25 million users, and it explains the number.

The fix isn't to take anything away. "It's not about taking things away that people are used to. It's about creating new things that they haven't paid for yet." Build core functionality that the hardcore users and the business cases would actually pay for, keep the current app free, grandfather the people who have been there. Keep the community edition open, and build the paid layer on top of it out of core functionality rather than merchandise.

Utkarsh pushed back on the one thing GameNative has that its closed-source competitor doesn't: trust. Users don't like the competitor's data practices; they trust GameNative because it's transparent. Julian agreed that trust is the moat, and then said the line of the episode.

"User trust doesn't matter if you don't exist."

The Only, Not The Best

Ryan Junejo's problem was a different shape. He builds benchmarks and RL environments for frontier AI labs at Crash Labs, and he was shipping a new one, Agent Mafia, later that afternoon: frontier models playing Mafia against each other, scored on deduction and on who lies best. What he wanted to talk about was being solo. He'd become a solo founder the week before, and his fear was specific: once a lab looks under the hood and finds one person, will they take him seriously?

Julian's first answer was about the week, not the labs. Being solo is hardest before you have customers, because customers are "an incredibly orienting force." Without them, building becomes the addictive default, because it always feels like progress. And you've lost the accountability buddy that made the hard days easier to start.

His second answer was Kevin Kelly's.

"You don't want to be the best. You want to be the only."

If you're competing head to head with a dozen vendors, team size matters. If you're the only person who can fill a gap the lab didn't know it had, it doesn't. Ryan's plan already had that shape: publish the benchmark so labs come to him, show them the deficiency, then sell the fix. Julian's addition was that Ryan's background, at Cohere doing evals, is itself the credential; people should want to talk to him on the basis of what he's done. The rest is "probably more in your mind than it's going to be in theirs."

Jobs To Be Done

Charlie Kahn's customers love the pitch. Brio lets a wholesale buyer or seller say what they want in plain language, and the order builds, prices, and lands in the ERP without anyone touching a back-office system. That story gets him every call. On the call, the supplier immediately starts asking for features: screens, tabs, upload buttons. His question was whether to build those or go straight at the AI layer that makes them unnecessary.

Julian's answer came by way of MagicSchool, a company he worked with before starting Solo Founders. Its founder, a former high school principal, had tried to get his teachers to use ChatGPT and watched them use it like Google. The problem wasn't the model; it was the empty chat box. So he built small, single-purpose tools inside MagicSchool instead. Critics called them wrappers. "The little wrappers were the difference between somebody not using ChatGPT and somebody becoming an hourly active user."

"They're not asking for features. They're asking for jobs to be done."

Each micro-app is a job. Constrain it, let people pick the one they need, and you get the flexibility of an LLM with the specificity of a button. Julian's other note was harder: buyers say they want the two-week ordering cycle compressed. Do they? "Sometimes things that take long take long for a reason. It's just not the reason that everybody believes." Find out, then prioritize by how often a problem occurs and how badly it hurts.

The Mind Before The Body

Andy, previously a software engineer at Zoox and Oculus, is building Infinomni: describe or draw a character, get a design and a 3D-printable robot companion kit. Software and hardware in one solo scope. He asked whether he should find a second pair of hands and split the two.

Julian asked what he'd do with a person and unlimited money. Andy said: test more hardware, faster. That told Julian the company was still in discovery, and he proposed a different sequence. Put the character on a phone in a dock on the desk first. Let people live with it. Then, when the robot ships, "it is freeing the character that you've created from the phone." Maybe the answer to "how do I split my time?" is that you don't.

The reason he gave was the most reflective moment of the hour. What makes a being a being isn't its exoskeleton; it's memory, communication, recognition. "I don't know if the body is as important in the early days as the mind."

About The Founders

Utkarsh Dalal builds GameNative, the open-source Android app for playing your PC game library natively on a phone (1.25M users, 10K+ GitHub stars). Ryan Junejo is the founder of Crash Labs, building benchmarks and RL environments for frontier labs; he previously worked on data evals at Cohere. Charlie Kahn is building Brio, wholesale ordering in plain language. Andy Wang is building Infinomni, prompt-to-robot companion kits, after engineering roles at Zoox and Oculus.

Subscribe to the Solo Founders Podcast on YouTube, Apple Podcasts, or Spotify.
Join our newsletter for weekly insights from solo founders: solofounders.beehiiv.com
Building solo and want people at your back? Cohort five of the Solo Founders Program is open: solofounders.com/program

Draft notes (re-synced 2026-09-09 to the FINAL 47:35 cut — four founders; the Swarnim/services-as-software section and the Docmost mention were removed because they aren't in the published episode. H1 = locked YouTube title B; if title A wins the A/B, the blog H1 can stay — the blog is its own SEO surface and B is the search-friendlier line. ~1,420 words of body prose; ~1,710 all-in with takeaways, bios, and CTAs): numbers used = 1.25M users, "a few hundred" paying, $5–10/mo (all Utkarsh on-mic), 10K GitHub stars (README 10.4K), "Discord in the tens of thousands" (README 35K+; Ari's on-mic 50K not used). No application-supplied numbers ($15M GMV, $1.4M ARR, pre-seed, $84K ARR). Ryan's split is stated in one clause ("become a solo founder the week before") — cut to "Ryan is a solo founder, and his fear was specific" if he doesn't OK it. The final section leans on Andy's on-air phone-dock pivot — confirm Andy is fine with it being public (he's pre-launch) in the same DM as his name/company spelling. MagicSchool founder unnamed (spelling unconfirmed); "1 in 4 kids" omitted. Kevin Kelly and Pete Kazanjy attributed. The five-year client stays unnamed. No format-flagging language.