This episode of the Jukebox podcast features Luke Carbis — a long‑time WordPress contributor, plugin author, member of the Plugin Review Team, and co‑host of the Crossword podcast. Fresh from his WordCamp Asia talk, Luke lays out a practical set of ideas for improving the wordpress.org plugin ecosystem in the face of a recent flood of submissions driven largely by AI tools.
The problem: quantity means discoverability suffers
In the last year plugin submissions to wordpress.org have increased dramatically. AI makes it trivial to scaffold and publish multiple plugin projects quickly, which has produced a deluge of similar or low‑value items. That creates two related problems: for the Plugin Review Team it raises workload (though Luke notes the team has grown and wait times are reasonable at about a week), and—more importantly—for plugin authors and users it makes discovery and standing out much harder.
Current search and ranking
WordPress’s plugin search is open source and weights many signals: keyword matches, recent reviews, support responsiveness, and especially active installs. Active installs are hard to fake and therefore a powerful ranking signal, but they favour already‑established plugins and make it difficult for genuinely useful newcomers to surface. The open nature of the ranking rules also makes the system easier to game than proprietary search engines whose algorithms are closed.
Practical experiments and proposals
Rather than grand structural changes that might be politically unrealistic, Luke proposes a set of achievable, incremental improvements that would improve discoverability, developer experience, and user trust.
1) Connect wordpress.org accounts with WordPress installs
Using the upcoming Connectors API in WordPress core, users could log in to wordpress.org from within their site. That would allow the Add Plugins screen to show each user’s favourites and make personalized suggestions, helping experienced authors and agencies install the tools they use most often without hunting around. It also lays groundwork for any future commerce features that require authentication across services.
2) Allow curated “untrusted sources” or repo lists
Luke suggests letting users store lists of Git repositories (e.g., GitHub repos) or authenticated premium plugin locations in their wordpress.org profile. From a site’s Add Plugins screen a user could enable their trusted sources and install those plugins as easily as uploading a zip. This approach supports boutique or client‑specific plugins, consolidates distribution, and wouldn’t replace the official directory—wordpress.org would retain control and could remove any source flagged as malicious.
3) Consider an official commercial marketplace (carefully)
Luke raises the idea of enabling premium plugins to be sold via the wordpress.org ecosystem, with revenue shared to sustain the project. He outlines a model of roughly an 8% platform fee (3% to cover payment processing, plus 5% “for the future” to a governing foundation). Funds could support WordCamp travel, contributor stipends, and teams like plugin review. He acknowledges the controversy: introducing money touches core open‑source values and risks centralizing incentives, but argues it could make product launches and updates easier for creators and re‑energize the ecosystem.
Tradeoffs, governance, and leadership
Bringing commerce into an open community is contentious. Luke stresses that any move like this would require strong governance—ideally transfer of wordpress.org oversight to the WordPress Foundation—transparent rules, and community buy‑in. He accepts that leadership with a clear direction can be polarizing, but argues it may be necessary: without coherent product decisions, WordPress risks stagnation at a moment when competing approaches and AI are reshaping the web.
AI: adoption, ethics, and generational attitudes
AI is a double‑edged sword. On one hand, tooling like the official AI plugin shows how AI can be an optional, useful add‑on to WordPress. On the other hand, heavy reliance on AI for contributions—auto‑generating docs, plugin code, or site content—raises ethical and participation questions. Luke points out generational differences: many Gen Z users express strong skepticism toward AI, sometimes refusing to use it for assignments or content creation. Overreliance on AI could alienate contributors and users who value human craftsmanship and accessibility.
An actionable, balanced step: AI disclosure for plugins
To bring more transparency without heavy-handed policing, Luke proposes introducing an AI disclosure field in plugin headers. Authors could voluntarily declare the degree to which AI assisted development (from minimal idea generation to fully AI authored code). Displaying this metadata on plugin pages alongside reviews and last update info would let users make informed decisions and give maintainers data about AI’s real world impact on plugin quality and update frequency. Luke acknowledges the risk of dishonesty (some authors may misreport), but thinks visibility and market pressure are a pragmatic place to start.
Why these ideas matter
Taken together these changes aim to: improve discoverability for deserving plugins; make it easier for developers to manage and install their own or premium plugins; provide sustainable funding for core community maintenance; and introduce measured transparency about AI usage. Luke frames these as realistic, incremental experiments rather than sweeping revolutions—steps the community could try, evaluate, and refine.
Closing notes and where to hear more
If you want to hear the discussion in full, Luke cohosts the Crossword podcast (crossword.fm), and this Jukebox episode goes deeper into the tradeoffs and real‑world anecdotes behind each proposal. The conversation blends product thinking, ethics, and community governance—asking not whether WordPress should change, but how it can do so in ways that are practical, transparent, and respectful of its open‑source roots.
