Nathan Wrigley hosts Robby McCullough, co‑founder of Beaver Builder, on the Jukebox Podcast. Beaver Builder has been a staple WordPress page builder for about 12–13 years. Robby outlines his journey from running a web agency to building a product used to simplify site creation and hand sites to clients who can edit them without developer help. He notes recent major life changes—moving house and having a baby—while continuing to steer the product.
Robby explains Beaver Builder’s cautious approach to AI. Early on many products added simple “AI” features—headlines, copy tweaks—often as marketing rather than meaningful utility. Beaver Builder deliberately avoided that bandwagon. Now, however, Robby sees the technology reach a point where agentic tools and LLMs can do genuinely useful work. Behind the scenes the team is experimenting: using AI to build features they lacked resources for previously and integrating agentic coding tools into the Beaver Builder experience. He stresses they won’t announce features until they’re ready, to avoid overpromising.
They discuss how page builders changed web development by lowering barriers—no need to hand code templates, CSS, or PHP. Page builders helped broaden WordPress’s adoption by enabling nontechnical users and agencies to deliver sites faster. Robby still believes page builders are valuable: while AI can quickly generate pages, WordPress provides plumbing—drafts, featured images, metadata, extensibility—that’s important for serious, custom, or business‑critical sites. He frames the contrast: agentic “vibe coding” tools can produce static landing pages fast, but they risk hiding how a site is constructed and reduce opportunities to learn underlying technologies.
Two AI‑page builder workflows are considered. One: use AI to generate a complete site, then open it in Beaver Builder for visual edits. Two: work inside Beaver Builder and invoke AI piece by piece (e.g., generate a pricing table, update copy). Robby confirms Beaver Builder is exploring both approaches. One experimental idea is dragging an externally generated page into Beaver Builder and converting it into its interface so users can tweak it visually. Another is a chat agent inside the editor to focus on specific elements (pricing table, copy) and apply changes in context. They aim to expose front‑end code and CSS for developers who want to tinker, and to support “bring your own key/agent” models rather than closed tokenized experiences.
Nathan asks about business anxiety around AI’s rapid change. Robby admits there are existential questions—page builders faced similar threats before (Dreamweaver stigma, Gutenberg arrival)—but he remains optimistic. WordPress powers a huge part of the web; legacy sites and businesses won’t switch overnight. He imagines WordPress and page builders continuing to be relevant, especially for maintenance and customisation.
Maintenance is a recurring theme. AI can create beautiful, fast sites but the long tail—updates, seasonal changes, new content—requires editing tools and a predictable backend. Nathan suggests page builders may shift from creation tools to more of a maintenance/editing role: AI handles initial build, humans use visual editors for iterative tweaks. Robby agrees and highlights the team’s goal to blend visual editing with AI while keeping the system open and learnable.
They reflect on the loss and gain the new workflows bring. Robby feels nostalgia for handcrafting designs but finds AI empowering. He’s been learning modern CSS patterns by inspecting AI‑generated code and iterating on it. That iterative, “brute force” approach to design—knowing when something looks wrong and iterating until it looks right—fits well with AI’s rapid experimentation.
Robby shares a personal anecdote: dictation has been invaluable since he has a newborn. Talking to the computer while holding the baby has made him more productive. Nathan notes younger generations will find conversational interfaces normal; for them, the shift won’t feel strange. Robby’s partner, however, is wary of AI and prefers less exposure, illustrating how households may balance different attitudes toward these tools.
They consider social impacts. Robby worries AI’s convenience could reduce human collaboration; working with chat agents can mimic human conversation but lacks authentic human connection. Both hosts value in‑person interactions—WordCamps and community events built relationships and shared learning. They hope WordPress meetups and camps recover post‑pandemic. Nathan notes some local clubs and social activities have resurged as people seek real‑world connection, which may counterbalance increased screen time.
Robby also describes a hobby side story—tinkering with ham radios inherited from his father—and using ChatGPT to automate tedious steps (finding local repeaters, formatting CSVs). He appreciates old websites and tech as “classic cars,” valuing knowledge of how things used to work even as new tools change workflows.
They close on uncertainty and excitement. The pace of change is breathtaking; what seemed impossible a year ago is now routine. There’s both melancholy for old craft and optimism for new productivity and learning opportunities. Robby emphasises balance: make AI useful without locking users into a closed black box, keep tooling open so developers and learners can inspect and understand code, and preserve human connection and community. Nathan hopes to see Robby in person at events rather than an AI simulation—an echo of their shared belief in the continuing value of real human interaction.
