Android Faithful - Interview with Kevin Galligan of Touchlab - droidcon NYC

An interview recorded at droidcon NYC 2024 for Android Faithful, part of the Daily Tech News Show’s conference coverage. It’s the compressed version of both talks I gave that day, aimed at a general audience rather than a room full of Android developers — which meant explaining Kotlin Multiplatform to someone who doesn’t write code, always a useful exercise.

Chapter links jump into the video.

Ten years of droidcon

• 0:00 — Intro — Setting up the conversation at droidcon NYC, on the tenth anniversary of the North American event.
• 0:41 — There from the beginning — Touchlab ran the first three New York conferences, so the question is what a decade of evolution looks like from the inside.
• 1:24 — The first three years — Year one was DIY and largely unknown, with people finding out afterward that it had happened. Year two brought in people who actually knew how to run conferences. Year three was the last one Touchlab ran directly, and the only one they live streamed — a whole other level of complexity.
• 2:05 — The franchise model — droidcon started as close to a pure franchise: you’d work out a deal, get some resources, and run your own event. Early on every city had its own hand-rolled website and wildly varying attendance. That’s consolidated into four big ones — New York, San Francisco, London, and Berlin.
• 2:46 — What droidcon NYC has always been — A preference for the weird and interesting over content you could get from a tutorial or a YouTube video. The meetup runs differently again — closer to “here’s what the bytecode actually looks like.”

What changed in Android

• 4:06 — Are the old problems solved? — There’s an expectation that a field arrives somewhere and declares the problem finished. That’s not how it works — a comparison to watching the JavaScript world and wondering how it kept generating conferences. What changes is that better libraries, tools, and techniques let you stop solving the same problem and scale up to bigger ones.
• 5:27 — Why community events matter — Nothing against vendor conferences, but the talks tend to be about what the platform owner thinks you should do. At a community event it’s fine to stand up and disagree.
• 6:08 — The content provider era — Official guidance once said to build a content provider and talk to your own database through a string-based API. His reaction from the beginning was to just call SQL directly, and that message took a long time to land.
• 6:48 — Around 2016 the official docs started listening — The Android team began incorporating what the community had been saying and building. But nothing gets permanently solved — the solutions that arrive are just meaningfully better than the last ones.

Where Android development sits now

• 7:30 — Android in 2014 vs. now — In 2014 “we’re doing a startup” meant an app; you barely had to say the word. Occasionally someone would ask about PhoneGap, and that told you something about the startup.
• 8:54 — Why Touchlab started as an Android shop — Android was the platform everyone treated as an afterthought — iOS first, Android maybe if they raised money. Nobody in the co-working scene knew any Android developers, which looked like a business.
• 9:36 — Large, sophisticated apps and scaled teams — The apps became critical enough to the businesses running them that the teams grew accordingly. A large bank that had five people on its app in 2017 might have three hundred people around it now. The sophistication of the whole discipline changed.
• 10:17 — And startups today — Some go native, many don’t. Which is why the day-to-day work of a professional Android developer has trended toward large complex apps — while mobile itself stays essential.

Explaining KMP to a non-developer

• 10:59 — What Kotlin Multiplatform is — Prompted by having explained the same thing to Flutter’s founder the day before. You write common Kotlin logic that deliberately contains nothing platform-specific.
• 11:40 — What “common” excludes — No Java IO file references, nothing Android-specific, not even threading, because not every target has it. Platform pieces get represented abstractly in common and defined per platform, then it builds for Android, iOS, or wherever.
• 13:00 — It isn’t a framework — The distinction from Flutter or React Native, where the design intent is that you move into another world and build the whole app there. KMP is closer to inserting shared logic directly into the app you were already building.
• 13:42 — Sandbox containers — Underneath the platform specifics, Android and iOS are essentially sandbox containers. Permissions and camera access differ; a math equation doesn’t. To the objection that apps just call the network and put things on screen: go count how many conference talks are about architecture.
• 14:23 — The maximal KMP app — Imagine one built to share as much as genuinely makes sense, with the platform-specific parts done natively. Set that next to the same thing written twice and the win would be obvious. It’s technically possible — the question is how a team gets there.
• 15:04 — Small teams share a lot, large apps share a blip — Plenty of big apps have some KMP, and that used to be the metric worth tracking. Look at the proportion of overall code and it’s tiny. Hence the obsession: why are bigger teams struggling to scale this?
• 15:44 — Is it being adopted? — Yes. At the scale it should be is a different question. It’s stable and Google has endorsed it, so there’s somewhere to start.
• 16:25 — The risk of the story setting — People trust bigger names with larger microphones, even when those companies look nothing like their own. If the well-known teams say they looked at KMP and it didn’t really work out, it becomes another thing that was tried and abandoned. Hence the Moby Dick slide in the talk — maybe this whale is never happening, but he hasn’t cut bait yet.

Wrapping up

• 17:05 — Condensing two talks into one interview — The keynote and the second talk ran well over an hour combined, so this was the highlights pass.
• 17:46 — Where to find him — Twitter still exists but isn’t the same place, and the Android community there largely evaporated. The Kotlin Slack, LinkedIn, and the Touchlab blog are where the conversation actually happens now — and working out where everyone went is its own open question.