😸 Your plane's been talking

gm legends, happy Monday.

One note from Friday: Arena Agent Mode is free. Our take suggested otherwise.

Your flight has been leaking data the whole time, someone made their broken keyboard your problem, and Google Labs wrote you a story from your own inbox.

CabinLink reads the flight data your airline's WiFi has been broadcasting all along: position, altitude, speed, ETA, and destination weather. Vishrut Jha, an ASU computer science student, noticed the manifest was sitting there for any app to parse and built native layouts for iPhone, iPad, Mac, and Vision Pro — not a blown-up phone view.

🔥 Our Take: Airlines have served this data to their own seat-back maps for years. The WiFi manifest was always accessible to anything on the network. Jha just built the native app nobody else did, and it works on most major carriers without a paid WiFi pass. That's a lot of polish for something most people would have left as a browser trick.

QWERTYS is a puzzle game where you stack keyboard keys back onto their correct positions before the pile overflows — Tetris logic, keyboard layout rules, built by Diego Dotta.

🔥 Our Take: The full pitch is the tagline: "My keyboard fell apart. Now it's your problem." Honest, self-aware, and accurate. Most products could learn something from it.

Full disclosure: Wispr Flow is the AI dictation tool most of us at Product Hunt (use we still have a few holdout typers, what romantics). Hold a key, talk, and clean text drops straight into whatever app you're already in — Slack, email, Notion, your IDE, wherever your cursor lives. No switching windows. No copy-paste ritual. Just say the thing – yes, you can whisper it – and even your most run-on sentences will be turned into polished writing at 4x the speed of typing.

Dreambeans is a Google Labs experiment that generates a short personalized AI story each day drawn from your connected Google apps.

🔥 Our Take: Google has been reading your email for years. Dreambeans just makes it visible. Whether a daily story pulled from your inbox and calendar feels charming or uncomfortable tells you something about how you already feel about Google having the data.

Nikolas Dimitroulakis (@nikolas_dimitroulakis), who founded ApyHub, posted a pointed question in the vibecoding forum: when AI writes the code, who's actually responsible for testing it?

The thread split into two honest camps. The rigorous side asks Claude for unit and integration tests alongside every feature. The pragmatist side ships the happy path and lets real users find the edges. Neither is wrong, and most people reading this are probably somewhere in between and not admitting it.

"Building something is only half the job; testing is what actually tells me if it works." — Deangelo Hinkle, landing the obvious thing that still needed saying.