AINEWS
The Browser Sim That Outlived Every Vibecoded Clone
A 14 year old free basketball game keeps growing in 2026 — and its creator says that's exactly what AI competition can't touch.
A 14 year old free basketball game keeps growing in 2026 — and its creator says that's exactly what AI competition can't touch.
ZenGM has been making a free basketball simulation game for 14 years. Not the play-the-game kind fans of NBA 2K recognize, but a text-and-stats strategy game where you draft players, set lineups, and watch simulated seasons play out. Now AI coding tools have made it cheap for almost anyone to ship a similar browser game in a weekend. In a recent blog post, ZenGM — who publishes under the pen name ZenGM — argues that is not the threat it looks like.
"Vibecoded" is the new shorthand for software, usually games, assembled quickly with AI coding assistants, often by people without engineering backgrounds, prioritizing speed over craft. ZenGM wrote that he sees a new entry land "almost every day," most of them vibecoded, and most of them on the web because AI coding tools default to HTML, CSS, and JavaScript. The wave has not yet reached his game's audience. Basketball GM's traffic and revenue are, in his telling, higher in 2026 than at any point in its history. He turned the project into his full-time job in 2021.
The case ZenGM makes for sticking around is structural, not sentimental. A free browser sim running on ads faces a specific economic problem: a rival can match its price (zero) but cannot easily match what 14 years of iteration has produced. The game has accumulated the polish that comes from a long-running project, an engine that has been refined through years of community feedback, and a player base that already knows where to find it. A clone, even a competent one, starts at zero on each of those axes. ZenGM estimates that advantage could last a year, maybe two, before vibecoded competition catches up, and he flags that estimate as a guess, not a forecast, invoking Yogi Berra's line about predictions being hard, especially about the future.
The honest read is that the moat is real but not airtight. Vibecoded entrants can iterate quickly, copy features in days rather than months, and reach the same web distribution channels without asking permission from a platform holder. If one of them hits on a feature Basketball GM lacks, like better mobile play, faster sim speed, or a fresher UI, the cost of switching for a casual user is low. A 14-year head start is a moat against direct clones, not against a genuinely better alternative.
A second factor is helping in 2026 that has nothing to do with AI. ZenGM credits part of Basketball GM's best-ever revenue to an improving ad market for niche web games, with programmatic rates and direct sponsorships both stronger than the 2021 trough, per his own accounting in the same post. That is a tide lifting all small browser games, not a defense specific to his title.
The broader question his post raises is not really about basketball. It is about what happens to a category of small, free, hobby-funded software, the kind a single person can run as a side project for years, when the cost of building a credible copy drops to near zero. The author's answer is that the moat is depth, not price. The open question is whether depth, accumulated over a decade and a half, holds against a generation of rivals that did not have to learn the hard way.