WWDC 2026: What They Actually Announced (From Someone Who Ships iOS Apps)
Four days ago I published a wishlist for what WWDC needed to deliver for indie developers. The keynote happened yesterday. Here's the scorecard.
Four days ago I published a wishlist for what WWDC needed to deliver for indie developers. The keynote happened yesterday — Tim Cook's last as CEO, which is its own storyline. Here's how the actual announcements stack up against what I was watching for.
The Scorecard
Background task handling for small apps: No change.
This was the item I cared most about and the one I was least hopeful about. Apple announced performance and stability improvements across iOS 27 broadly, but there's nothing in the announcements that suggests the background task API situation for small apps changed in any meaningful way. Still routing around the same constraints with the same workarounds. I'll watch the developer documentation closely, but from the keynote: miss.
SwiftUI layout reliability: Probably improved, impossible to tell yet.
iOS 27 is pitched as a stability and performance release. That's good news for SwiftUI behavior, historically. The Liquid Glass UI changes introduce new complexity at the rendering layer, which could go either way for edge cases. The honest answer is I won't know until I spend time in the beta. Provisional partial credit; revisit in August.
App Store improvements: Nothing.
Nothing on keyword space, review queue improvements, or rejection clarity. This is the one I said I'd given up on, and the keynote confirms that cynicism was warranted. The App Store didn't come up in any meaningful way. Moving on.
On-device AI without a subscription model: Complicated.
This is the most interesting result on the scorecard. Apple Intelligence is expanding significantly — deeper integration, more capabilities, the Siri AI overhaul. The on-device processing story is real and Apple's Private Cloud Compute architecture is genuinely designed to minimize data exposure.
But Siri AI now routes certain requests through Google's Gemini models. That's cloud inference, not on-device. The dream of fully local AI that doesn't require ongoing cloud costs or data exposure took a step in both directions yesterday — more capable, but also more cloud-dependent for the new capabilities. Partial credit, with a footnote.
Widget interactivity: Near miss, actually.
visionOS 27 got an extra-small widget size. That's a widget surface change, not an interactivity model change. The deeper issue I wrote about — the interaction model feeling like a workaround — didn't get addressed. I wanted to be wrong about this one. I wasn't.
The Thing I Didn't Put on the Wishlist
The macOS 27 "Golden Gate" redesigned system search index is genuinely interesting. Not because it solves a problem I named, but because a better system search means better discoverability for apps that aren't getting App Store placement. Anything that improves how users find installed apps is worth paying attention to for a studio with multiple apps in the ecosystem.
This is what I mean about the quiet announcements. Nothing from yesterday gets a stage headline for indie developers. The background improvements to search indexing, whatever stability changes landed in the SwiftUI rendering stack, the developer beta feedback loop starting today — those are where the actual value is.
The Honest Take
Yesterday was a Tim Cook keynote more than a developer keynote. The features that got stage time — Siri AI, the UI refresh, parental controls, Photos Spatial Reframing — are consumer product decisions. The parts I care about as someone who ships apps are in the developer documentation, not the keynote deck.
Developer betas are live. I'll know more in two weeks.
XenneX is a one-person studio shipping iOS apps. See what's in the lineup →