Interview, Fireside Chat, Conference Presentation
Seeing The Future from AI Companions to Personal Software
- Eugenia (Wabi) characterizes the current AI era as a "Microsoft DOS" phase, arguing that chatbots are the "DOS era for AI interfaces" and predicting a forthcoming "Windows/Mac OS moment" defined by visual, interactive, and user-generated applications rather than text-based command lines.
- Wabi's core product thesis posits that the future software paradigm will shift from a "20 million developer" model to a mass-market ecosystem where users build, tweak, and remix "mini-apps" for themselves and their communities, driven by deep personalization.
- The platform emphasizes "ephemeral, highly personalized software" that solves niche problems (e.g., a specific puzzle game for a child or a gym tracker based on a specific fitness book) which are too small to exist on traditional app stores but solve immediate user needs instantly.
- Wabi is introducing a "social graph" update allowing users to see who downloads specific apps, comment on apps, and "remix" or request specific tweaks from creators, transforming apps into community-building tools rather than isolated, static utilities.
- The company explicitly avoids showing code or technical details to non-technical users, aiming for a "Canva-like" visual experience where users design apps through natural language and visual controls rather than API keys or prompts.
- Eugenia critiques the industry's obsession with voice-only interfaces as a "mind trap," arguing that screens are essential for discovery, proactivity, and multi-user environments, noting that 75% of voice devices are shipped with screens even for basic use cases like timers.
- Wabi functions as an "operating system built on the platform of you," designed to unify context across apps (e.g., a nutrition app accessing fitness data from a workout app) to break down current "walled gardens" and enable true "Software 3.0" via deep personalization.
- Eugenia reflects on her 10-year journey with "Replica," noting a missed strategic opportunity to bet heavily on generative AI hardware/software earlier due to capital constraints, concluding that "sometimes you need to go big or go home" and lack of risk tolerance can lead to suffering consequences in the current environment.
- The founder attributes her predictive ability for consumer AI trends to a background in journalism and deep empathy for the human condition, contrasting her "non-technical" perspective with the "savant" profile of most AI builders who often lack understanding of everyday user friction.
- Wabi aims to monetize the creator economy by allowing influencers to distribute "mini-apps" (e.g., fitness protocols, art generators) directly to fans, offering a more functional alternative to traditional courses or videos and fostering deeper fan relationships.
- The platform addresses the "discovery problem" of AI prompts by encapsulating complex, long-form prompts into shareable, functional mini-apps, eliminating the friction of copy-pasting text and troubleshooting model compatibility.
- Eugenia predicts the future hardware landscape will be "screen-first" AI devices with locally running models, rejecting the voice-only paradigm in favor of an "AI-first operating system" that handles personalization and context more effectively than current CPU-driven smartphones.
- Historical context provided includes the 2015 breakthrough with DeepMind's dialect generation papers and the 2020 access to OpenAI's GPT-3 API, where Eugenia's team became a top partner before OpenAI publicly released ChatGPT, highlighting the early dominance of language models in the market.
- Wabi currently sees a "90/10" split in usage (consumption vs. creation), with the majority of users currently tweaking existing apps rather than building original ones, though the long-term goal is a mass-market shift toward creation.
- The team is actively developing "multiplayer" features to allow users to collaborate within apps or join community feeds (e.g., a universal dog photo feed), aiming to turn individual mini-apps into shared social experiences.
- The interview highlights a trend where "content" and "software" are merging, with creators potentially launching "professional classes" or "narrative experiences" as executable apps, moving beyond static video content to interactive, context-aware tools.
- Eugenia notes that early AI tools were often "toys" or "weird" (referencing early YouTube), expecting a similar trajectory for Wabi where the initial wave of simple, creative apps will evolve into a robust platform of community starters and functional utilities.