Eugenia Kuyda envisions a future where AI-powered personal software democratizes app creation, shifting from developer-dominated models to user-driven innovation. She critiques current command-line-style AI interfaces as outdated, comparing them to MS-DOS-era computing, and advocates for visual, mobile-first tools that make building apps as easy as editing photos. Platforms like Wabi enable users to create lightweight, shareable mini-apps that carry context and personality, fostering communities around personalized AI experiences. These apps solve the problem of sharing complex AI prompts by embedding them in interactive, visual formats—similar to TikTok filters. The discussion highlights trust, data security, and design guardrails as key to empowering non-technical users. Kuyda draws from her decade-long journey in AI, from early language models to Replika and beyond, emphasizing empathy over pure technical prowess. She dismisses voice-only devices as limiting, arguing instead for screen-based, locally powered AI that puts control in users’ hands.