iOS 27 Shipped Siri AI and a Redrawn Liquid Glass. Here's the Redesign Tax on a Solo App
Apple released iOS 27, iPadOS 27, and macOS Golden Gate on September 14. The headline feature is Siri AI, a rebuilt version of Siri that Apple is positioning as a genuine ChatGPT-style assistant baked into the OS. The other headline is quieter: Liquid Glass, the design language Apple introduced last year, got a visible second pass, with sharper icons and more uniform refraction across the system. I run a handful of small iOS utilities as side projects, and I spent this week doing the unglamorous math on what both changes actually cost a one-person team, not the marketing math.
What actually shipped
Siri AI is the real story. It's a much more personalized, capable version of Siri that can answer general questions the way a chatbot does, not just trigger predefined actions. Alongside it, Apple added "Extend" and "Reframe" tools to Photos, photorealistic image generation inside Image Playground, and the ability to create new Shortcuts just by describing what you want in plain language. On the iPad, visual intelligence now extends to on-screen content, so you can circle something with the Apple Pencil and get context on it directly.
The Liquid Glass refinement is smaller in scope but not trivial: Apple describes it as "more uniform refraction and improved contrast," and app icons are noticeably sharper and more detailed than the version that shipped last year. If you've ever redone an icon set to match a platform refresh, you know that sentence hides real hours.
The device gate nobody puts in the headline
Here's the part that matters more for a small app than the feature list: iOS 27 runs on the same devices as iOS 26, back to the iPhone 11. But Siri AI and the rest of the Apple Intelligence feature set require an iPhone 15 Pro or newer. If your app has any hook that assumes the user's phone has Siri AI available, a meaningful slice of your install base, anyone on an 11 through a 15 non-Pro, simply won't see it. I checked my own crash analytics for one of my apps and roughly a third of active devices don't clear that bar. Building a marquee feature around Siri AI integration this cycle means building a fallback path for a third of your users on day one, not as a later nice-to-have.
The redesign tax is real, and it's not optional forever
I don't think a solo developer needs to chase pixel-perfect Liquid Glass compliance the week it ships. Nobody's five-star reviews turn into one-star reviews because your app icon has slightly less refraction than the system apps. But I've watched this pattern before with past design language shifts: skip two release cycles of visual updates and your app starts looking dated in App Store screenshots specifically, which is where a lot of install decisions actually get made. That's a conversion cost, not an aesthetics complaint, and it compounds quietly because nobody files a bug report for "this app looks like it's from three years ago."
The honest estimate, based on redoing icon assets and a few key screens for one of my own apps this week: about six hours for a small app with a simple icon and two or three screens that lean on system chrome, more if you've built custom UI that doesn't inherit system materials automatically. That is real but bounded work, which is the good news.
The part that should worry a "smart assistant" app more than a utility app
If your app's actual pitch is some flavor of "ask it anything and it answers," Siri AI just became a free competitor built into the phone. That's not a new problem exactly, Apple has always had the platform-owner advantage, but a genuinely competent, conversational Siri is a different level of threat than the old rules-based version. I'd treat this the way I treat any platform feature encroaching on a product idea: check whether your actual value is the conversational layer (which you will lose) or something adjacent to it, like domain-specific data, workflow integration, or judgment calls a general assistant won't make well. If it's the former, this is the week to start moving.
What I'd actually do
I'm prioritizing the Liquid Glass icon and key-screen updates this month, because that's bounded, cheap work with a real payoff in how the app reads in the App Store. I'm deliberately not chasing full visual intelligence or Shortcuts-by-description integration yet, because those are still narrow in scope and the API surface will likely keep shifting through at least one more point release. Where I could be wrong: if Siri AI adoption turns out to be slower than Apple's numbers suggest, because it's gated to iPhone 15 Pro and newer, the competitive pressure I'm describing might take another year to actually bite rather than showing up this quarter. I'd rather spend the cheap hours now and be a little early than spend the expensive hours later scrambling.
Author
Lukas
@lukcombinator