Over the last month, my colleague Steph Yiu and I spent time talking to artists, technologists, and open-web builders about a question that Automattic advisor Stephen Wolfram continues to push us on—what is the future of the website?
As part of that work, we kept a running NotebookLM where we dropped our interviews, source material, and side explorations as we went. It became a kind of shared scrapbook for the project—part research archive, part idea garden, part pattern-finder.
For the final conversation, we let NotebookLM interview us (meta!).
After weeks of hearing other people’s answers, it felt useful to turn the questions back on ourselves and see what themes had accumulated. Here’s that conversation:
Are we about to enter a chaotic, glorious renaissance of delightfully weird, homemade personal websites—like the old GeoCities days—or are we just going to drown in generic AI templates?
Christie: I think it’s going to be both. There will absolutely be generic AI templates while people are onboarding to new tools and figuring out how to use them. But I also think these tools can break us out of a templated era. In some ways, websites have already been pretty uniform for a while. Shopify sites often look like Shopify sites. Squarespace sites often look like Squarespace sites. AI tools, at their best, give people more control over what they are building, and that creates room for more taste, more personal expression, and more weirdness.
Steph: There is going to be an explosion of more creative online expression, but we are also going to drown in generic templates for a while. Both things can be true. The middle will probably regress toward sameness. But I think the most reputable institutions, and the people who care deeply about their digital presence, are going to use this moment to build more distinct online experiences. That’s the optimistic version of this—AI can lower the barrier to making something while also creating new incentives to make something more personal and more memorable.
If our bots are just going to endlessly chat with your bots behind the scenes, who is the visual website actually for anymore?
Steph: One of the ideas that shaped our project came from James LePage’s writing about a split web: one layer for agents, one layer for humans. I think WordPressWordPress wordpress.org is actually well positioned for that. The people who intentionally land on your website in that world are likely to be highly intentional. They are not just wandering by. They are there because they want to engage with you, your brand, or your work in some deeper way. That makes the website more valuable, not less. The top of the funnel may increasingly happen somewhere else, like in Reddit threads, in AI search, or with AI assistants, but the site becomes the place where higher-intent visitors go to verify, and connect.
Christie: We also shouldn’t assume agents remove the need for specificity. In fact, I think they increase it. If my agent gets better at understanding my preferences, then the websites on the other end will also need to be more specific about what they are and who they are for. The “averaged” version of a brand or store will be less useful. The brands a real point of view will still need to express that both to the agent layer and to the human who lands there.
As the main internet floods with AI generated slop we’re seeing people retreat to the cozy web, to private spaces, Discord servers, and highly curated personal sites. In a world of infinite AI content, is having a clunky, 100% human-made website about to become the ultimate luxury status symbol?
Christie: I think NotebookLM’s wording is funny here because it’s implying that a human-made website would be clunky, and I don’t think that’s right. If anything, AI is often what makes things feel overcooked and clunky. What matters is not whether something is 100 percent human-made in some pure sense. It is whether human discernment is still present. Are people using AI as a tool inside a process that still has taste and judgment, or are they just letting it generate the whole thing and calling that good enough?
Steph: I had a very literal example of this. There is a clothing brand I shop from every summer, and this year they used obviously AI-generated images in their product catalog. It made me sad. It degraded the brand. It did not reflect the integrity or quality of what they actually make. That is the issue. There has to be thought attached to it. Human review. Human care. Alignment with the brand and what it is trying to communicate.
If a website is constantly shape-shifting based on who or what is looking at it, how do you even design for that?
Christie: I don’t think this is entirely new. Websites are already different for different geographies, devices, contexts, and users. In some ways, the internet archive exists because websites have always changed over time. The real difference now is speed, and perhaps the number of layers involved. But I also think it depends on the creator’s intent. Not every website is meant to be stable forever. Some people may want ephemerality. Some may want an archive. Some may want a site that changes shape constantly.
Steph: If you visit a global brand site from Europe or Asia, you already get a geolocated experience. The bigger issue is wayfinding. How do people know what is permanent, what is personalized, and what they can trust? I think websites and browsers will need clearer signals for that. The analogy I used in the conversation was infrastructure and signage at airports or train stations. Even when they differ, there is still enough consistency that you can figure out how to move through them. That is what the web will need more of if the surface becomes more adaptive. We’re already seeing this tension in media. Editorial homepages are trying to balance a stable, curated front door with growing pressure to personalize. That tension feels like one of the defining design questions ahead.
What became obvious throughout the project is that the website is not dead. If anything, websites will continue to be places where people, brands, and projects can show up with more intention than they can on more mediated platforms—and that seems to be becoming more important than ever.