四組數據分析：原生 iOS 和 Android 看來是 2026 年最弱的手機開發選擇

I spent the last week counting public data about mobile: Hacker News hiring threads, the job-seeker threads, Show HN launches and npm installs. This post gives my position on those counts. It is an opinion on partial data, and I mark where each claim comes from and where it stops. Three of the four sources are Hacker News, so they share one audience and do not confirm each other independently.

The four signals

Android moves the same way in the two hiring-side rows: mentions fell 69% and seeker posts per hiring post rose 5.9 times. On Show HN it grew faster than iOS, 4.7 times. Show HN as a whole grew 3.4 times over the same period, so iOS titles took a smaller share of the board (0.91% to 0.74%) while Android took a larger one (0.46% to 0.64%).

My position

1. Among the stacks I could compare on both sides of the HN threads, native iOS and Android look the weakest. They lost the most mentions in hiring posts (about two thirds) and gained the most seeker posts per hiring post (5.4 and 5.9 times). Full stack moved 2.9 times and backend 3.7 times. I left AI out of that comparison because its keyword grew fastest (6.2 times) and is the noisiest. Flutter's mentions in hiring posts fell about as much, from 2.8 to 0.8 per thread, but I did not count Flutter seekers, so I cannot place it. Native is one of the groups getting harder, and in this sample it is furthest along.

2. The mobile activity I can see did not disappear. Show HN titles about iOS and Android kept growing in absolute numbers. Installs of the main cross-platform toolchain rose about ten times, but react itself rose 9.6 times, so I do not read that as more builders. Whether more distinct people launch mobile apps than in 2023 is untested. What fell is the number of hiring posts that mention iOS or Android. I cannot tell whether those roles ask for native depth alone.

3. React Native sat in the least bad bucket on the two HN hiring-side measures. Its mentions in hiring posts fell 14% against 33% for the whole thread, and its seeker posts per hiring post rose 2.0 times against 5.4 for iOS. On Show HN its share of titles fell, on small counts. Expo is the part of the npm data that moved most against the control: its share of react downloads went from 2.2% to 5.2%. I take that as a reason to be able to show cross-platform work next to native depth. It does not argue for dropping native.

What would change my mind

• A larger postings dataset with a clean iOS and Android split that does not show the two-thirds drop. I could not find one for 2025 and 2026, and that gap is why I counted myself.
• October 2026 holding. It was the best month for native hiring in 2026 H2, 11 iOS and 9 Android hiring posts against 3 to 7 in July to September, though 2023 H1 averaged 19 iOS posts per thread. If November and December look like October, the ratio of iOS seeker posts to iOS hiring posts drops toward 3.4 and my 4.3 overstates the squeeze.
• Evidence that Show HN titles grow faster than the number of distinct makers. If one person ships three times as many titles, the launch signal says less than I made it say.

What I would do with it

• If your CV says iOS or Android and nothing else, add one thing you can show next to it: a cross-platform app or a backend service. The ratios above say those groups moved by smaller factors.
• Compare ratios across groups. Raw npm downloads rose ten times for react itself, so any single count mostly measures the registry.
• For one sample point: in 2026 H2 on HN, iOS-mentioning seeker posts ran at 4.3 per iOS-mentioning hiring post. That is a ratio of mentions in one corner of the market, and it is no forecast for any single opening.

Limits

• All four sources are public and partial. HN skews to startups and to US and remote roles, and three of the four datasets come from it. npm counts include CI and automated installs. A keyword in a post is not an open role.
• The periods differ: HN rows compare 2023 H1 with 2026 H2 averages, npm compares single months (January 2023 and September 2026).
• The seeker ratio is a ratio of two noisy averages. One extra iOS hiring post per thread moves it from 4.3 to 3.7. The iOS hiring average is 6.3 in the first post and 6.2 in the second.
• I did not measure salaries, seniority, location, app store data or anything outside these sources. Each earlier post lists its own limits, and the scripts are in them.
• The counts will be rerun as more months arrive.

Where does your own pipeline disagree with this reading?