The Future of Software is User-Generated: Believes the vast majority of necessary software has yet to be built and will be created by ordinary users for themselves using AI-powered, no-code platforms, rather than by professional developers.
Graphical Interfaces Trump Voice for AI: Argues that the AI industry is mistakenly fixated on voice as the ultimate interface, ignoring user preference for visual interaction and the success of simple, GUI-based AI apps. She advocates for an AI-first OS on smartphones.
Current Chatbots are in their 'DOS Era': Views today's text-based LLM interfaces like ChatGPT as primitive and analogous to the MS-DOS command line, predicting a more user-friendly 'Windows/Mac OS moment' is still to come.
AI Startups Face a Capital Dilemma: Reflects that the decision to remain scrappy and profit-focused, while viable, can cause a company to miss out on capital-intensive 'generational' opportunities like building foundational models.
Breakthrough Apps Require Native Thinking: Asserts that transformative applications emerge when developers leverage the unique capabilities of a new platform (e.g., GPS on smartphones for Uber) rather than simply porting old formats, implying the killer app for AI will be something fundamentally new.
2015
Cites a Google research paper on dialogue generation as the pivotal event that caused her to pivot Replica's entire company strategy toward building conversational AI models.
Pre-2020
Observes that major tech companies were fearful of releasing generative AI products following the public failure of Microsoft's Tay chatbot, creating an opportunity for smaller players.
2020
Replica is invited by OpenAI to be a first-launch partner for the GPT-3 API, receiving a personal demonstration from Sam Altman and Mira Murati, cementing Replica's position as an early market leader.
Post-2020
Begins to publicly reflect on Replica's strategic path, expressing that its scrappy, capital-efficient approach may have led it to miss a 'generational chance' to build foundational models.
Present
Shifts focus to her new company, Wabi, a platform for creating AI-generated mini-apps, with a product strategy focused on ease of use, social features, and avoiding technical complexity for users.
▶The Democratization of Software Creation
Kuyda argues that the current software landscape, dominated by a few professionally developed apps, is insufficient for the world's needs. She envisions a future where AI empowers ordinary users to build and share their own applications, much like Canva simplified design, predicting that the vast majority of future software will be built this way.
This positions her company, Wabi, not just as a product but as a platform play aiming to capture the 'long tail' of software needs, a potentially massive market if the barrier to creation is sufficiently lowered.
▶Evolution of AI Interfaces Beyond Text and Voice
She is critical of the current 'DOS era' of text-based chatbots and the industry's fixation on voice as the ultimate interface. Citing the success of GUI-based apps like Prisma and the prevalence of screens on devices like Alexa, she advocates for AI-first operating systems with more intuitive, graphical interfaces.
Her focus on user-friendly GUIs for AI suggests a belief that mass adoption will be driven by simplicity and visual interaction, not by users becoming prompt engineers, creating an opportunity for companies that master AI-native UX.
▶The AI Startup's Dilemma: Scrappiness vs. Foundational Ambition
Through her experience with Replica, Kuyda reflects on a key strategic crossroads for AI startups. While being nimble and profit-focused ensured survival and early market leadership, it came at the cost of not raising enough capital to pursue the 'generational chance' of building foundational models.
This highlights the immense capital requirements of foundational model development and the strategic divergence between application-layer companies and infrastructure-layer companies in the AI ecosystem.
▶From Web 1.0 to AI 1.0: Historical Parallels in Tech AdoptionApr 2026
Kuyda frequently draws parallels between the current state of AI and previous tech cycles, such as the early iOS App Store, GeoCities personal web pages, and the shift to mobile-native apps like Uber. She sees the current AI landscape as nascent and awaiting its 'Windows Mac OS moment' and breakthrough native applications.
This historical perspective suggests she believes the current AI leaders and use cases are likely transitional, and the most significant, market-defining AI applications have yet to be built, implying a long-term investment horizon is necessary.