Eugenia Kuyda shares predictions about consumer AI that challenge the viability of several current trends. First, she argues that screenless AI devices are unlikely to succeed. People primarily use phones for passive activities like scrolling through feeds, making voice interactions impractical as a main interface. You can't casually glance at a voice-activated device like you can with a screen. Similarly, she believes "always listening" devices will fail because our daily lives lack sufficient content to justify constant recording. Significant moments—like therapy sessions or job interviews—are often private and would not be captured by such technology. Kuyda sees potential in mini-apps, which she believes could revolutionize software development by making it easier for users to create personalized tools without the complexity of full applications. This shift may lead to the emergence of the first true consumer super-app in the U.S. Looking ahead, she predicts that by 2030, there will be two dominant AI chatbots: one focused on knowledge and assistance, and another designed to foster relationships and personal growth. The former must be reliable, while the latter should offer an engaging experience. She also claims that performance marketing for apps is dying. Many studios rely on copycat strategies, which drive down margins and lead to unsustainable business practices. In a surprising twist, Kuyda suggests that an AI-generated webcam girl could become the fastest consumer product to reach $1 billion in annual recurring revenue, thanks to real-time video technology and constant user engagement. Finally, she emphasizes the need for companies to create better discovery methods for AI applications, moving beyond basic text prompts to unlock a wider range of functionalities. The ability to demonstrate AI's potential could shape the next major consumer platform.