This episode unpacks Meta CTO Andrew Bosworth's strategic vision for a post-mobile future, detailing how the convergence of AI and new hardware like AR glasses aims to fundamentally reshape human-computer interaction and create entirely new platform dynamics.
The Future of Content Consumption: Beyond the Phone
- Andrew Bosworth (Boz) outlines a 10-year vision where content consumption moves beyond the smartphone screen.
- He anticipates augmented reality (AR) glasses—devices overlaying digital information onto the real world—becoming commonplace, alongside more immersive, social experiences akin to Las Vegas's Sphere but accessible anywhere.
- The 5-year outlook is less certain, predicting a spectrum from high-end, expensive AR/VR devices to more accessible, lower-resolution smart glasses suitable for quick information retrieval.
- Boz notes the challenge lies in creating experiences rich enough to potentially replace current devices, acknowledging an initial period of uneven distribution based on cost and capability.
Combining Tech Trends: From Newsfeed to AI Interfaces
- Boz reflects on Meta's historical strength in combining emerging technologies to solve user problems, citing the creation of the Facebook Newsfeed as an early example merging social, mobile, and early AI.
- He emphasizes a problem-first approach: "We really immersed in like what the problem was... you are going to reach for whatever tool is available." This user-centric view, he argues, allows for better identification and adoption of genuinely useful tech trends.
- Contrasting with past tech waves that felt forced, Boz highlights the current AI revolution's tangible feel and broad applicability across numerous domains, despite present challenges like factuality and compute costs.
- He notes AI's unique potential: "This kind of feels like, oh, everything's going to get better... every single interface that I interact with... are going to be made easier by virtue of this new technology."
Meta's Vision: AI Meets New Computing Interfaces
- Meta anticipated the saturation of the mobile phone form factor around 2015, leading to a long-term bet on new interfaces.
- The core idea was that future computing needed more natural input/output, requiring face-worn hardware for visual/auditory delivery and neural interfaces (systems translating brain signals or nerve activity into commands) for control without keyboards or touchscreens.
- Boz describes this hardware and interaction design challenge as immense, requiring building entirely new modalities beyond the decades-old direct manipulation interfaces (like mouse or touchscreen).
- The recent surge in AI capabilities arrived sooner than Meta expected, providing a "wonderful tailwind" by offering a more intuitive way for users to express intent and for devices to understand context (what the user sees/hears).
Reality Labs Deep Dive: Quest, Smart Glasses, and Orion
- Boz discusses Meta's current hardware efforts under Reality Labs.
- The Ray-Ban Meta smart glasses evolved into "AI glasses" mid-development following breakthroughs like Llama 3, adding conversational AI capabilities and "Live AI" (real-time visual understanding).
- He contrasts the incremental hardware changes from Ray-Ban Stories with the significantly richer interactions enabled by AI.
- Orion represents Meta's high-end AR glasses prototype. Experiencing Orion, Boz suggests, offers a glimpse into a plausible "post-phone world," especially when combined with AI capabilities like the demonstrated breakfast recipe example (where the glasses identified ingredients and suggested recipes).
The Post-Phone World: Challenges and Convergence
- Replacing the deeply integrated smartphone presents enormous hurdles.
- Hardware challenges (attractiveness, weight, cost, battery life) remain significant.
- The phone's established ecosystem and user familiarity are major barriers. Boz questions whether simply porting phone apps to AR/VR is the right approach, drawing parallels to early mobile failures where web experiences were merely shrunk down.
- "What we found is the apps want to be different when they're not controlled via touchscreen," Boz states, highlighting the need for native interaction designs suited to AR/VR inputs (like hand gestures or voice).
- While partnerships with phone ecosystems aren't ruled out, Boz sees the phone platform as both a potential hardware anchor and a software "anchor to drag" due to its non-native app structures.
AI's Disruptive Potential: Inverting the App Model
- Boz presents a "hot take" that AI could fundamentally invert the traditional application model, even on existing devices like phones.
- Instead of users selecting specific apps (like Spotify) to perform tasks (play music), they would simply state their intent to an AI assistant ("play this music").
- The AI would then orchestrate the task, selecting the best service based on availability, quality, user subscriptions, or even suggesting new services.
- This shifts focus from app providers to user intent, potentially abstracting away brand loyalty and forcing services to compete more directly on performance and value. Boz questions: "If you were building a phone today, would you build an app store the way you historically built an app store?"
Building the New Ecosystem: AI, Developers, and Marketplaces
- The transition to an AI-driven interface model requires building a new developer ecosystem.
- Boz envisions agentic AI (AI capable of taking actions and completing tasks autonomously) becoming more capable. User requests that the AI cannot fulfill represent a "gold mine" for developers.
- Meta could present this "query stream" of unmet needs to developers, showing clear demand for specific functionalities or integrations ("hooks") into existing apps or services.
- This creates a dynamic where developers build bridges for the AI to access their functionality, potentially creating new marketplaces for services (even physical ones like plumbers) discovered and orchestrated via AI.
Trust and Market Dynamics in an AI-Driven World
- This AI-mediated future raises significant questions about trust and market structure.
- Users must trust the AI intermediary not to be biased or prioritize partners over the best user outcome ("not be bought and paid for in the back end").
- Boz draws a parallel to the evolution of Google Search, where an initial phase of disruption and value creation (like in travel) eventually led to SEO "gamesmanship" and potentially degraded user experience.
- The challenge for AI platforms will be to avoid this "decaying era" where optimization for the system outweighs genuine user value. This dynamic puts pressure on brands that rely on direct customer relationships.
Meta's Open Source Strategy: Llama and Beyond
- Meta's commitment to open-sourcing AI models like Llama stems from two core beliefs.
- First, the open research culture of FAIR (Fundamental AI Research group), which believes societal progress is faster through collaboration. Boz notes that historically, major AI breakthroughs (like the Transformer paper from Google) were shared openly.
- Second, a strategic business rationale: AI models are seen as potential commodities that complement Meta's core products (social media, advertising, future interfaces). "You want to commoditize your compliments," Boz explains.
- Open-sourcing powerful models fosters a competitive ecosystem, lowers costs, benefits startups and academia, and ultimately helps Meta as a major application provider leveraging that AI. This aligns societal progress with Meta's business model.
Navigating the Road Ahead: Risks and Challenges
- Despite the potential, Boz acknowledges significant risks to realizing the AR/AI vision.
- Invention Risk: Basic technological hurdles might still prove insurmountable; the desired hardware may not yet be buildable at scale or cost.
- Adoption Risk: Social acceptability of wearing face computers and users' willingness to learn new interaction modalities are major unknowns.
- Ecosystem Risk: Developers and essential applications might not embrace the new platform quickly enough.
- Regulatory Risk: Profound questions arise around privacy and the implications of "superhuman sensing" (AI-enhanced vision, hearing, memory). Boz poses the example: "Am I allowed to use a tool to assist me [with remembering someone's name] or not?" Missteps here could derail progress significantly.
Meta's Conviction: Believing in the AR/AI Future
- Boz stresses Meta's deep conviction, driven by Mark Zuckerberg's belief that this is the necessary next computing platform.
- He cites Meta Chief Scientist Michael Abrash's concept of the "myth of technological eventualism"—progress isn't inevitable; it requires dedicated effort and investment.
- "We believe in this stuff in our cores... We will not fail for lack of effort or belief," Boz asserts, framing the work as a rare, generational opportunity to redefine human-computer interaction, comparable to the early days of personal computing.
Bosworth articulates Meta's high-stakes bet on AI-powered AR/VR fundamentally displacing the smartphone. Crypto AI investors and researchers should monitor the evolution of AI as an interface layer, the viability of open-source models in this new paradigm, and the emerging platform dynamics that could reshape digital interaction and value capture.
