A hobbyist's air-traffic-control browser game had its leaderboard hacked within hours
A solo developer built a browser homage to the 2009 iPhone hit Flight Control, complete with hand-drawn flight paths and a live leaderboard. Within hours of it hitting Hacker News, someone found the scoring API and started posting ten thousand landings in under a second.
You drag each color-coded plane's path by hand to the runway threshold that matches it, landing and launching before two aircraft collide. There's no tutorial: players spent their first minutes inventing their own routing rules, then their next minutes crashing anyway.
A Principal Engineer at AWS who works on EKS showed up with a detailed list of interface fixes. Someone else had already worked out you could just POST a fake score straight to the API, and the board filled with ten-thousand-landing runs that never happened.
What you actually do in Airport Simulator: drag lines, match colors, don't crash
The rules aren't posted anywhere on the page. Drag a line from a color-coded plane to the runway threshold in the same color, get the geometry close enough, and the plane lands, taxis off, and clears the airspace for the next one. Miss, and it keeps flying in whatever direction you last pointed it, straight through anything in its path. Departures complicate things further: they crawl across the runway right when you're least watching for them.
Commenters converged fast on calling it a spiritual sequel to Flight Control, the 2009 iPhone game that popularized this exact drag-to-land mechanic before quietly disappearing from every app store. Airport Simulator runs in the browser instead, built solo, with a blocky low-poly style commenters compared to OpenTTD and the Commodore 64-era Kennedy Approach. Its own high-score board, visible mid-game and occluding a chunk of the map by several accounts, turned it into a competitive obsession almost immediately: one player posted a video of an 81-landing run in under seven minutes, daring "a professional SC: Brood War player" to try to beat it.
The leaderboard got gamed before the bug reports finished coming in
The 142-comment thread split cleanly between people improving the game with feedback and people breaking it for fun.
"I notice people have realized they can set their own score with an HTTP request (I mean, it's Hacker News), maybe put in some sort of mitigation?"
— a commenter who'd spotted the exploit early
By the next day, the board was full of runs nobody could have actually played.
"Should the leader board be cleared of the folks with 10,000 landings in 1 second? It looks like the true leader right now is 90 landings in 7:29."
— a player tracking the real leaderboard by hand
On the other side of the same thread, the feedback was constructive enough to double as a product spec.
"A slider to control game speed... A clearer arrow over the landing zone. It took me quite a while to realize how to land a plane... A timer over each runway to next takeoff."
— a Principal Engineer at AWS who works on EKS
One exchange nailed why the genre keeps getting rebuilt from scratch: the difficulty never comes from any single hard thing, it comes from doing too many easy things at once, all in the thread.
Someone's already imagining the next version: AI pilots on the radio
"A few months ago I built a piloting harness for LLMs in X-Plane... I imagined what it might be like to scale it up and make an ATC game out of it, talking to dozens of AI pilots. The most open-ended ATC sim ever…"
— a developer who built an LLM piloting harness for X-Plane
The genre Flight Control walked away from in 2010 turns out to still have room left in it, whether the next version gets drawn by hand or dictated over a fake radio to a dozen AI pilots at once.
Get breakdowns like this every day
The top Hacker News stories, unpacked — by email, Telegram, or RSS.
A blogger's proof that AI can't understand context failed the moment HN tested it
Rachel, who blogs at rachelbythebay.com, says her server logs are full of hits to posts that were never published, proof, she argues, that readers trust AI summaries of her writing without checking them. Her test for whether a model understands anything, rather than pattern-matches it: why is driving through San Francisco's Chinatown early in the year a bad idea? Commenters ran that exact question through ChatGPT, and it named Lunar New Year immediately.