Cal.com Goes Closed Source — And Explicitly Blames AI-Assisted Cloning

Cal.com, the open-source Calendly alternative with millions of users, announced today that it is closing its codebase. The company's blog post was unusually candid: competitors are using large language models to analyze Cal.com's open source code and rapidly clone its features, eroding the competitive moat that justified the open-source bet.

"We built this in the open so the community could contribute and trust us," the post reads. "But we've watched multiple well-funded companies use our MIT-licensed code as a blueprint and ship feature-equivalent products faster than we can innovate." The shift will not affect existing deployments running the current open-source version, but future feature development will happen in a private codebase.

The move ignited a fierce debate on Hacker News (260 points, 146 comments) that went well beyond Cal.com. The core argument: AI has fundamentally changed the economics of open-source competition. In 2020, cloning an open-source product required months of engineering time. In 2026, a well-funded team can use AI coding agents to analyze a codebase, understand its architecture, and generate a feature-equivalent clone in days. The "community goodwill" advantage of open source, the argument goes, no longer offsets this asymmetric threat.

Critics pushed back strongly. Several commenters noted that closing the source rarely stops determined competitors — they still have the existing open-source version as a reference. Others argued Cal.com is diagnosing the wrong problem: the real threat isn't code cloning but distribution, sales, and brand moats that open source was never meant to provide. Developer Marcus Ellers' response post ("Cal.com Is Drawing the Wrong Lesson") argues that the right response to AI-assisted competition is to ship faster, not to hide code.

The broader implications reach every commercially-licensed open-source project. If AI lowers the cost of cloning to near zero, the MIT/Apache license ecosystem faces a structural challenge it wasn't designed for. Expect more projects to quietly add source-available licenses like BSL or SSPL in the months ahead — and expect more conversations about what "open source" even means when AI can read and reproduce it instantly.