158: Live From WWDC 2016 With Guests Phil Schiller and Craig Federighi

Episode Summary

Executive Summary: Live from WWDC 2016, John Gruber interviews Apple executives Phil Schiller and Craig Federighi about Apple’s push toward openness, security, and developer empowerment. They discuss iOS 10, Mac/Watch improvements, App Store changes, Siri integrations, privacy-preserving machine learning, Swift progress, and the new APFS file system, framing Apple’s strategy as careful expansion without sacrificing user trust.

Main Topics: Apple’s openness to third-party developers (Priority: 5/5): Federighi explains how iOS 10 expands system-level integrations—Messages, Siri, Maps, CallKit, CloudKit, and extensions—so apps can participate directly in core user workflows rather than remaining isolated icons. App Store policy changes and discovery (Priority: 5/5): Schiller details faster app review, search ads, and broader subscriptions, emphasizing privacy, relevance, and protections intended to help smaller developers rather than only large incumbents. Privacy and security as product architecture (Priority: 5/5): The conversation explores Touch ID, secure enclave, Auto Unlock, and differential privacy as examples of Apple designing systems that preserve convenience while limiting data exposure and attack surface. Siri, machine learning, and on-device intelligence (Priority: 4/5): Apple frames Siri’s new developer API and photo/search intelligence as domain-specific and privacy-conscious, with processing done on-device where possible and carefully bounded by categories. watchOS 3 and usability improvements (Priority: 4/5): They explain the dramatic performance gains in watchOS 3 as a result of keeping apps resident, using background updates, and rethinking the watch around glanceable, two-second interactions. Swift and platform modernization (Priority: 3/5): Schiller notes Swift’s internal adoption, source stability work, and the goal of making Apple’s APIs more native to Swift while improving long-term developer confidence. APFS and the future of storage (Priority: 3/5): The new Apple File System is presented as a major infrastructure upgrade for performance, cloning, snapshots, updates, and security across flash-based devices.

Key Arguments: Apple is intentionally opening system experiences to third-party developers so users can interact with apps where the task naturally happens, such as Messages, Siri, Maps, and the lock screen. The App Store changes were designed to improve discoverability and monetization without compromising privacy, with no minimum bid and relevance prioritized over raw ad spend. Subscription models should be available broadly, but Apple wants guardrails so users are not forced into undesirable recurring billing for trivial functionality. Apple’s machine-learning approach is fundamentally different from cloud-first competitors because it relies on on-device processing and privacy-preserving math rather than centralizing personal data. Differential privacy lets Apple learn population-level trends without being able to reconstruct individual behavior, even if company leadership changes or systems are compromised. watchOS 3’s speed improvements came from engineering decisions like keeping favorite apps resident in memory and using the battery budget more efficiently than originally expected. Apple views Swift’s source stability as the key milestone for developers, with ABI stability as a later concern; internal adoption is expanding across major OS components. APFS is positioned as a foundational storage upgrade that improves cloning, snapshots, file protection, and system updates on modern flash storage.

Data Points: WWDC live show year: 2016 - The Talk Show was recorded live from WWDC 2016. Show year for sponsor mention: Fifth year - Gruber says he has been doing the show for at least five years. Years Mezzanine hosted the event: Fourth time - Gruber notes it is the fourth time at Mezzanine. App review times: Way faster than they used to be - Schiller says Apple sped up review times to address developer pain ahead of time. App Store search ads beta charges: $0 during beta - During the beta, real downloads are delivered without marketing spend. iCloud photo analysis timing: Overnight on AC - Federighi says backlog photo analysis occurs when devices are plugged in overnight. Scene classification compute: About 11 billion calculations - Federighi describes the on-device work needed for photo scene understanding. Auto Unlock proximity defense: Speed of light time-of-flight - Apple uses peer-to-peer Wi‑Fi timing to prevent relay attacks. Secure enclave passcode tries: 10 tries - Federighi says the secure enclave only allows ten attempts for unwrapping. Messages usage claim: Most used app on iOS - Schiller says Messages is the most used iOS app, period. Emoji size increase: 3x - They reference a keynote announcement that emoji would be expanded threefold. WatchOS task target: 2 seconds - Apple aims for watch tasks to be completed end-to-end in about two seconds. Copy/paste continuity timeout: About 2 minutes - Universal Clipboard is described as time-limited to avoid surprises.

Pivotal Quotes: "giving the developers more and more opportunities to let users do what they want to do across all these experiences was a way that we could really make the platform better for all of our users" — Craig Federighi: Explaining Apple’s strategic reason for opening up system integrations in iOS 10. "we want to make sure that what we do is preserve the intelligence of your assistant" — Craig Federighi: Justifying why Siri supports only selected domains rather than a vague app handoff model. "if it works as you're describing it, it means it's not just that Apple doesn't use that information to reverse the anonymity, it's that mathematically you can't" — John Gruber: Summarizing Apple’s differential privacy approach during the privacy discussion.

Implications: Apple’s 2016 direction is clear: deepen platform capabilities while avoiding data centralization. For users, that means more seamless features; for developers, more entry points and monetization options; for the industry, a stronger argument that privacy and advanced AI can coexist.