What Comes After Mobile? Meta’s Andrew Bosworth on AI and Consumer Tech
🎯 Summary
Podcast Summary: The Post-Phone Era, Intent-Driven Computing, and the AI Revolution
This episode of the a16z AI Revolution series features Andrew Bosworth (Bos), former Head of Reality Labs at Meta, in conversation with David George, a Growth General Partner at a16z. The core discussion revolves around the impending shift away from the current app-centric, smartphone-dominated computing paradigm toward a new, more agentic, adaptive, and immersive experience driven by AI and advanced hardware.
Key Takeaways for Technology Professionals
1. The End of the App-Centric World
The central thesis is that the last two decades of consumer technology, defined by the smartphone, touchscreen, and app model, are reaching saturation. Bos argues that if one were starting from scratch today, they likely would not build the current app-centric world. The future interface will run on intent, not taps and swipes.
2. The Convergence of AI and Hardware for New Interfaces
The transition to a post-mobile world requires breakthroughs in two areas:
- Hardware Evolution: Moving computing off the phone and onto the face (smart glasses/AR headsets) to leverage our natural input/output channels (eyes and ears). Bos details Meta’s efforts across the spectrum, from the Ray-Ban Meta AI glasses (which pivoted from smart glasses to AI glasses post-Llama 3) to high-end concepts like Orion.
- AI as the Interaction Layer: The sudden acceleration of AI capabilities (like LLMs) provides a “tremendous tailwind” for interaction design. AI agents can now better understand vague intent, context (what the user is seeing/hearing), and execute complex tasks, making the interface significantly more natural than direct manipulation (mouse/touch).
3. Intent Over Orchestration: The Inversion of the Software Model
A major “hot take” discussed is how AI will fundamentally invert the software model. Currently, users must orchestrate their needs by selecting a specific application (e.g., “I need to open Spotify”). In the future, the user will simply express intent (“Play music”), and the intelligent agent will handle the orchestration, selecting the best provider (Spotify, Tidal, etc.) based on context, quality, and availability.
4. Business Implications: Brand Moats and Marketplace Dynamics
This shift poses existential challenges to existing business models:
- Brand Abstraction: If the AI agent handles orchestration, it may abstract away brand names. Consumers may care only about performance, price, and value, rather than brand loyalty (e.g., which music service they use). This puts immense pressure on brands whose value proposition relies heavily on direct consumer attachment.
- Emerging Marketplaces: The failure points of early AI agents (where they cannot fulfill a request) will reveal the most valuable new functionality gaps. These gaps will become the basis for new developer ecosystems and marketplaces, where services (from software apps to real-world services like plumbers) can build “hooks” for AI agents to access.
- Trust in the Distributor: As the AI agent becomes the primary distributor, trust in that distributor (e.g., the OS provider or the foundational AI model) to provide the best price/performance, rather than the best-paid partner, becomes paramount.
5. Challenges and Timeline
Bos acknowledges the difficulty of displacing the smartphone, calling it an “incredible anchor device” with a massive developer ecosystem.
- 5-Year View: Tricky, as the phone is deeply entrenched. Early, less rich experiences (like always-on, low-resolution smart glasses for quick content grabs) will emerge.
- 10-Year View: Much clearer confidence in a proliferation of alternative content delivery vehicles (glasses, immersive VR/AR) that offer experiences currently impossible (e.g., feeling courtside at a game with a remote relative).
- Hardware Hurdles: The challenges of making new form factors attractive, light, affordable, and capable of all-day battery life remain significant.
6. Methodology: Problem-First Approach
Bos attributes his success (and that of his cohorts at Meta) to being deeply immersed in what people are trying to do rather than being driven solely by the technology available. This problem-first orientation allows builders to honestly assess tools and embrace the next wave when it genuinely solves user problems, which is why the current AI revolution feels so tangible and broadly applicable.
🏢 Companies Mentioned
💬 Key Insights
"This is a classic, I believe this is even going to be commodities. And you want to commoditize your complements."
"we're going to make way more progress if these models are open. Because a lot of these contributions aren't going to come from these big labs. They're going to come from these little labs."
"It abstracts away a lot of companies' brand names, which is going to be very hard for an entire generation of brands."
"That's the gold mine that you take to developers, and you're like, 'Hey, I've got 100,000 people a day trying to use your app, use your app, use your app.'... If you build these hooks, you've got 100,000 people clamoring for your service, coming in for your service."
"I am wondering if the progression of AI over the next several years doesn't turn the app model on its head."
"What we're now excited about is, okay, take all those pieces and layer on the ability to have an interactive assistant that really understands not just what's happening on your device... but also what's happening in the physical world around you and is able to connect what you need in the moment with what's happening."