In a wide-ranging conversation on the Jukebox podcast, Nathan Wrigley spoke with Luke Carbis about the present and future of the WordPress plugin ecosystem. With two decades in the project—building plugins, running businesses, contributing to Core, and serving on the Plugin Review Team—Luke lays out practical problems he sees and concrete ideas for improvement, all while wrestling with the ethical and cultural questions AI introduces.
The core problem: discovery, not just review
The plugin directory is seeing an enormous influx of new submissions—driven largely by tools that make plugin creation easier, including AI. That spike has created two separate challenges. First, it increases workload for human reviewers; second, and more crucially, it makes it much harder for good plugins to be discovered among thousands of lookalike options.
Luke notes that the review team has scaled up and improved tooling (and even uses AI internally), so wait times are reasonable—around a week—despite higher queue counts. The deeper issue is product visibility: how users find the right plugin and how authors make their products stand out. The open-source nature of WordPress’ search algorithm helps transparency, but also makes it easier to game. Active installs remain the hardest-to-fake and most important ranking signal, which benefits established plugins but can make it nearly impossible for new, high-quality entries to surface.
Practical experiments to improve discovery
Luke proposes small, achievable experiments to improve the experience without a wholesale redesign. Two linked ideas stand out:
– Connect wordpress.org accounts to individual WordPress installs. With WordPress adding a Connectors API, allowing users to log in with their wordpress.org credentials would let personal preferences and favorites flow into the Add Plugins screen. This could surface a user’s curated list of plugins and improve setup speed across sites.
– Support “untrusted sources” or repo lists in a user profile. Let users register Git repositories (or other repos) in their wordpress.org profile—these might be their own private plugins, client plugins, or premium-only code. When creating a new site, users could install those repos directly from the Add Plugins interface, as if uploading a zip. This gives developers a sanctioned way to distribute custom or premium plugins without undermining the central directory, and makes managing client installs far simpler.
Rethinking commercial distribution on wordpress.org
Luke also raises the contentious idea of allowing premium plugins to be sold directly through the WordPress plugin directory. He frames this as something that would only make sense if wordpress.org were governed by the Foundation and suggests a fee model: roughly 8% (3% payment processing + 5% to “Five for the Future” or similar). Funds could support WordCamps, paid contributor hours, the plugin review team, and other community infrastructure.
There are obvious trade-offs. Introducing money into parts of the project that have long been entirely free risks community split, governance questions, and concerns about preferential treatment. Luke argues the benefit is a more vibrant product ecosystem and easier discoverability for paid offerings, which could re-energize launches and make it simpler for developers to build sustainable businesses in WordPress. He acknowledges that this is divisive and would require transparent governance and strong oversight.
Leadership, urgency, and the role of AI
The conversation turns to leadership and direction. Luke believes WordPress needs clear, decisive leadership to navigate a period of rapid change. He sees a real risk that without a strong strategy, WordPress could lose relevance as alternatives emerge—especially with AI reshaping how websites and content are built.
On AI specifically, Luke distinguishes two things: using AI as an optional feature in WordPress (e.g., plugins that add AI capabilities) versus making AI central to how the project is run (auto-generated docs, PRs, or contributor work). He supports optional, additive AI integrations but is cautious about making AI mandatory or pervasive in contributor workflows. There are ethical concerns and a generational dimension: many younger people are openly skeptical of AI and may be discouraged from participating if AI use becomes a default expectation.
Transparency: an AI disclosure for plugins
One concrete, low-friction proposal Luke offers is an AI disclosure field in plugin metadata. Authors could voluntarily declare the degree to which AI was used during development—ranging from none to minor ideation or automatic code generation. Surfacing this as plugin metadata would achieve two things: it gives users information to make informed choices, and it generates data about how AI usage correlates with plugin quality, update cadence, and support behavior.
Of course, such a system relies on honesty and could be gamed. Still, making the disclosure visible opens a market signal and lets the community observe patterns rather than guess them.
Respecting contributors and users
Underlying all of Luke’s suggestions is a consistent theme: protect the people who use and build WordPress. Whether it’s making it easier for developers to manage their own plugins across sites, funding contributor work through a shared revenue model, or giving users clear information about AI’s role in plugin creation, the aim is to maintain trust and practical utility.
Next steps and community input
Many of Luke’s ideas are presented as experiments—small changes that could be implemented and evaluated. They invite community feedback and testing rather than top-down imposition. They also hinge on governance choices and culture: how much commercialization should wordpress.org accept, how will disclosures be validated, and how will leadership balance innovation with inclusivity?
If you want to hear Luke expand on these ideas, he co-hosts the Crossword podcast (crossword.fm) where he and Jonathan Wold cover WordPress topics in depth. The discussion on Jukebox is a useful primer for anyone interested in plugins, discovery, ethics, and the practical impacts of AI on open-source ecosystems.