The current era of AI chatbots is the MS-DOS moment—a clunky, text-heavy command line that prevents most people from accessing the true potential of models that can already do so much more. Eugenia Kuyda, drawing from her experience pioneering the AI companion Replica, argues we are on the cusp of a “Windows or macOS moment” for AI interfaces. This new paradigm is not about voice assistants or better chatbots, but about “personal software”: a universe of small, visual, hyper-specific applications that are as easy to create, share, and remix as a YouTube video or TikTok reel. Through her new platform, Wabi, she envisions a future where software creation explodes from the domain of 20 million professional developers to 8 billion everyday creators.

The transition hinges on moving beyond the affordances of a command line. When people see a chatbot, they default to a narrow set of use cases like search or writing assistance. To unlock everything else, we need interfaces that are interactive, visual, and intuitive. Kuyda describes a future operating system where you’ll have mainstream apps alongside a feed of “mini apps” suggested by AI, built by friends, or crafted by you for a fleeting need. These are not traditional, durable apps built for scale, but ephemeral and deeply personal tools—like a puzzle game for a child featuring their favorite princesses, or a gym workout generator tailored to your specific equipment, fitness book, and goals.

This shift turns software into a form of personal expression and social currency. Just as people share video clips and playlists, they will share and tweak mini apps, building communities around niche interests. A fitness influencer might distribute their workout protocol as an app; a group of dog lovers could share an app that transforms pet photos into royal portraits. The platform becomes a framework for memory, context, and expression, learning about you as you create and use these apps. Kuyda believes this will lead to a new class of creator—not just developers, but anyone with a specific need or creative idea—and will fundamentally change our relationship with technology from one of consumption to one of active, joyful creation.

Surprising Insights

• Voice is not the primary interface of the future. Despite the pervasive “Her” movie analogy, Kuyda argues voice-only devices are a “mind trap” because they fail in crowded spaces, offices, bedrooms, and anywhere you need visual feedback or privacy. She champions screen-first, AI-native devices.
• Software will become ephemeral and hyper-niche. The future isn’t million-feature monoliths from the App Store, but tiny, disposable apps built for a single person, a specific week, or a particular context (e.g., an art show finder for your upcoming New York trip).
• The biggest constraint on AI adoption is the interface, not the model capability. The conversation reveals that even with powerful models like GPT-3, most users never move beyond basic use cases because the chatbot interface doesn’t inspire exploration of more complex possibilities.
• Apps will evolve into community starters and social objects. Beyond mere utility, mini apps can become the focal point for communities (e.g., parents at a specific preschool) and a new form of content for creators to share with fans, beyond videos and merchandise.

Practical Takeaways

• Build software for your own hyper-specific needs. Don’t search for a perfect app; use no-code or “vibe coding” platforms to quickly build a simple tool exactly for your personal workflow, whether it’s tracking a unique hobby or managing a family routine.
• Treat software as malleable. Adopt a mindset where any app you use can and should be tweaked and improved over time based on your evolving needs, rather than accepting a static, one-size-fits-all product.
• Prioritize visual and interactive design over complex prompts. When sharing an AI-powered idea, don’t share a long text prompt. Instead, build or share a simple, self-contained app with a graphical interface—it dramatically lowers the friction for others to experience your idea.
• Focus on human empathy and context in design. Whether building an AI product or any consumer tool, deep observation of how real people struggle with technology in their daily lives is a more reliable source of insight than abstract technical ideals.