Apple reveals ARKit 3, which can perform real-time occlusion of people, and a bunch of tools to make AR development easier with Reality Composer and RealityKit
Jeremy Horwitz / VentureBeat :
Context & Ripple Effects
This is the third annual iteration of a platform Apple has been building since it first unveiled ARKit in 2017, when simple demos like a virtual measuring tape suggested the iPhone could reach mass-market AR faster than Google's Tango hardware approach. The 2018 release added shared experiences across four iOS devices and spectator mode in ARKit 2; ARKit 3 now attacks the two biggest credibility gaps — virtual objects passing believably behind real people, and the tooling burden of building AR apps at all.
First-order effects
- iOS developers get real-time people occlusion plus RealityKit and Reality Composer, meaning AR scenes can render humans as part of the environment rather than flat sprites pasted over them.
Second-order effects
- Google's Android-side AR tooling faces pressure to match occlusion and drag-and-drop scene authoring, since Apple is bundling rendering, animation, and audio into one framework instead of leaving developers to assemble components.
Third-order effects
- If each WWDC keeps compounding capability on the same installed base of iPhones, AR matures as a software layer on existing devices rather than waiting for dedicated glasses — a pattern that continued when Apple followed with ARKit 4's depth capabilities and Location Anchors a year later.
The trend: Mobile AR is advancing through annual software-only platform updates that widen Apple's lead over hardware-dependent rivals like Tango-era Google.