Every App Is Becoming Every Other App: Wispr Flow and Granola Just Swapped Jobs

• Book: Ship It — the pragmatic startup tech stack
• Also by me: Thinking in Go (2-book series) — Complete Guide to Go Programming + Hexagonal Architecture in Go
• My project: Hermes IDE | GitHub — an IDE for developers who ship with Claude Code and other AI coding tools
• Me: xgabriel.com | GitHub

In less than three months, two apps I'd have put in different categories ended up doing the same thing.

Wispr Flow is a dictation app. You hold a key, speak, and the text lands in whatever field your cursor is in. On August 5 it shipped Notetaker, which sits in your meetings and produces the summary and the action items.

Granola is a meeting-notes app. It listens to your calls and writes the notes. On October 29 it shipped Voice Typing: hold a key, speak, and the text lands wherever your cursor is. That was Wispr's whole product.

A few days later I found a games button on Medium.

"Play Games", right next to "Write"

Medium's header now has a "Play Games" button, and it sits beside "Write". It opens Good Words, a word game with three-minute rounds, a leaderboard, and a row of avatars at the top telling you who else is playing right now.

Medium's Good Words. Player names and photos changed.

The same move, everywhere

LinkedIn started shipping daily puzzles in 2024 and hasn't stopped: Queens, Tango, Zip, Mini Sudoku, Patches, Wend. It hasn't published how many people play, only how many come back: in 2025 it reported that 86% of players return the next day and 82% are still playing a week later. That's a habit metric, and it's the one LinkedIn chose to publish.

LinkedIn Games. Connection photos and counts changed.

Netflix added games to its app, and this year video podcasts. In October 2024, co-CEO Ted Sarandos put it plainly to investors: "we compete directly with YouTube for people's time."

YouTube answered from the other side. It has hosted more than 75 games under the name Playables since May 2024, and this year it brought back direct messages, a feature it had switched off in 2019.

Apple News, a news app, has a Puzzles section: a crossword, Sudoku, a word game called Quartiles, and since July 2025 a daily Emoji Game.

Spotify took on audiobooks, then video podcasts, then in August 2025 direct messages. It had removed messaging in 2017 because few people used it.

Duolingo teaches math, music and chess on top of languages, and all of it counts toward the same streak. Chess arrived in April 2025 and crossed a million daily users on iOS within a quarter, faster than any subject Duolingo had launched before.

OpenAI merged its Atlas browser and the Codex app into the ChatGPT desktop app over the summer.

Notion started with notes and docs. It now ships Mail and Calendar, and since May 2025 AI meeting notes, which is Granola's home turf.

None of these additions is what the company was founded to sell. Each one gives you something to open the app for on a day you otherwise wouldn't.

Why it's happening now

My best guess is that AI made features cheap to build.

Wispr already ran speech-to-text in production. So did Granola. Going from dictation to meeting notes, or the other way round, stopped being a big project for either of them, so both made the trip.

When a competitor can copy your feature within a quarter, the feature alone won't keep anyone. What keeps people is having your app already open when the need shows up. Daily opens and minutes in the app turn into the numbers that count, and a puzzle or a second product pushes both up.

Where this helps you, and where it costs you

Part of this works in your favour. When your meeting notes and your dictation share one vocabulary, your colleagues' names come out spelled right in both, which is the pitch Granola makes for Voice Typing. Wispr also lets Claude, ChatGPT and Gemini read your meeting notes over MCP, which pushes against lock-in. More apps should do that.

The other part is slower and easier to miss. You chose a tool because it did one job well. Then its team spreads across three jobs, and the one you care about gets a thinner slice of the roadmap. Meanwhile your data and your routines collect in one place, and every month it costs a little more to walk away.

I don't see this reversing. What you can do is read a tool's last few release notes before you commit to it. If most of them are about something other than the job you need done, that tells you where it's heading.

Has a tool you rely on drifted far enough from its original job that you left? Tell me which one in the comments.

If you're on the other side of this, building the product instead of picking it, the pull toward the next feature is strong. Ship It is my guide to the stack a small team needs between localhost and production: hosting, payments, auth, CI/CD and monitoring. The less time that stack eats, the more goes into the one job your users chose you for.