His project went viral this week

Author: The Next New Thing host interviewing Jonathan Wilke (Outbid, Supastarter) | Published: 2026 (episode says the launch was “this week”) | Source: transcript

Summary

A roughly 17-minute interview, sponsored by Zapier, with Jonathan Wilke, a Germany-based indie developer who built Outbid.lol in about three hours of AI-assisted coding in Cursor and saw it produce about $212,000 in revenue within days. The product is deliberately trivial: a public leaderboard where anyone can take the top spot by paying at least $1 more than the current leader. The top rank cost around $17,000 at recording time, and buyers get attention, since the site drew about 1.3 million visitors in five days. One participant reported roughly 50 demo calls from a top-10 placement.

The build story is instructive less for the code than for the workflow. Jonathan dictated a voice prompt naming the stack (Next.js, Polar for payments, shadcn/ui for components) and the one rule of the product. Cursor was configured to switch into plan mode automatically for anything non-trivial, so it asked clarifying questions before building. For more complex features he uses Matt Pocock’s “grill me” skill to force an exhaustive interrogation of requirements up front (Plan Before Build (AI-assisted development)).

The launch was a hand-driven attention loop: he stayed up until 2:30 am, posted the site on X (4.4 million impressions on the first post), then reposted every new leaderboard entry with the buyer’s link. As traffic grew, paying for rank became rational, which drew more entries and more posts (Launch Attention Loop). He rates the outcome as 99 percent luck and 1 percent execution, with the execution being a clean, non-generic interface and radical simplicity. The 85-plus clones that appeared increased hype rather than hurting it, and buyers who publicly spent money became champions of the platform because they had an incentive to make it look like a good bet.

On pricing, the host observes that most vibe-coded projects never charge because their builders feel embarrassed. Jonathan attributes his comfort to four years of selling Supastarter, a SaaS starter codebase. He launched it at roughly $29 to $49, doubting anyone would pay, and now sells it at $349 while still believing it is underpriced (Charge From Day One).

The operational half covers what breaks when something goes viral. Polar, a merchant-of-record payment provider, flagged the business as outside its fair-use policy, forcing an evaluation of Stripe and others. The host explains the merchant-of-record trade: the provider is the legal seller, handles sales tax and refunds across jurisdictions, and therefore restricts what it will sell (Merchant of Record vs. Direct Payments). A botnet hammered the site while Jonathan slept; his Vercel spend limits paused the site, Vercel restored it and refunded the overage, and he credits managed hosting for the site surviving at all. His analytics provider crashed under a 400,000-visitor Hacker News spike and he switched to DataFast. He had never heard of the Million Dollar Homepage and says he would not have built Outbid if he had.

Key Claims

• Outbid.lol took about three hours to build in Cursor and generated about $212,000 in revenue within days of launch.
• The site drew about 1.3 million visitors in five days; the top rank cost about $17,000 and had received about 30,000 clicks.
• A participant (Crowd Reply) reported around 50 demo calls booked from a top-10 placement, so buyers got real lead flow, not just vanity.
• The first launch post reached about 4.4 million impressions; reposting each new leaderboard entry with the buyer’s link drove the growth loop.
• Jonathan attributes the result to 99 percent luck, 1 percent execution (non-generic design, extreme simplicity).
• Over 85 clones appeared within days; he says copying increased hype rather than diluting it.
• Supastarter launched at about $29 to $49 and now sells at $349, and he still considers it underpriced.
• Polar’s fair-use policy did not permit this business model because a merchant of record is the legal seller and carries refund and legitimacy risk.
• Vercel spend limits paused the site during a botnet attack; Vercel restored service and refunded the attack-caused charges.
• The analytics provider crashed at roughly 400,000 visitors in an hour after a Hacker News share; he switched to DataFast and uses its live API for on-site visitor counters.

Direct quotes worth preserving

“I would say it’s 99% luck with just 1% execution here.” — Jonathan Wilke

Why citable: The builder’s own honest attribution; a guardrail against treating viral one-offs as a repeatable playbook.

“I have been very comfortable charging and learned to actually not be embarrassed of charging people through my experience with selling Superstarter over the past four years now.” — Jonathan Wilke

Why citable: Names the mechanism by which pricing confidence is built: repeated exposure to customers paying, not a mindset trick.

“If I had hosted this myself the site wouldn’t probably be online anymore because this is just something I don’t want to care about. I just want someone to take care of that.” — Jonathan Wilke

Why citable: A solo operator naming the case for managed infrastructure and spend limits over self-hosting.

“A lot of our vibecoded products are just creations that we almost feel too embarrassed to charge for.” — The Next New Thing host

Why citable: Names the common failure mode the wiki’s pricing thesis has to overcome.

Connections

Entities mentioned: The Next New Thing, Jonathan Wilke (Outbid, Supastarter) Concepts referenced: Plan Before Build (AI-assisted development), Charge From Day One, Merchant of Record vs. Direct Payments, Launch Attention Loop, Productized Services (services-first, productize later), Agentic Service Operations

Contradictions / Tensions

• Tension with the wiki’s services-first thesis (index, Productized Services (services-first, productize later)): this is a product-first, luck-driven consumer-attention play, the opposite of the vertical local services path. It does not refute the thesis; Jonathan himself says it is not repeatable. Its transferable lessons are the pricing stance, the plan-first build workflow, and the infrastructure choices.
• Tension with Agent Web: Outbid monetizes raw human attention, the model the agent-web sources say is eroding. That it still worked at scale in 2026 is a reminder that attention markets are not dead, just harder to engineer on purpose.

Notes

• Added to capture builder-side lessons on charging, planning with AI tools, and payment-provider choice for a solo operator.
• Provenance: the guest is an indie SaaS/starter-kit seller in Germany, not a services consultant, and the host is a creator-interviewer. The revenue figures are self-reported on air. Sales-tax remarks from the host are explicitly uncertain (“I don’t remember. I don’t handle that myself”).
• Transcript artifacts: “Shettsy/Chetszian UI” is shadcn/ui; “outfit.lol” is outbid.lol; “Versel/Purcell” is Vercel; “Matt PCO” is Matt Pocock; “Teemo/Tibo” is Tibo of Outrank.