Gabriel Busto
games i created
these projects were where i first learned to build complete products, put them in other people’s hands, and iterate on real behavior rather than isolated code. at some point i think it would be interesting to build my own physics and game engine just to learn. but now is not that time.
a couple of web-based hytopia games
i built two games using hytopia’s sdk. these were both for game building competitions, but also made for fun side projects.
the overseer
the first one was called The Overseer. it’s a single or multi-player co-op game where you’re stuck in a damaged biodome with a rogue ai operator in charge of protecting the biodome. it sees you and other players as hostiles and tries to eliminate you using its biodome control system. attacks included things like increasing and decreasing temperatures to hurt you, a uv light that targets one player and follows them around (if you get stuck or stop moving, it can do some real damage), ground electrification, and so on.
i designed it such that the game would collect global game state including players remaining, player health, current biodome temperature, and so on, feed it to an llm, and have the llm use structured output to determine what actions to take next and what taunts to communicate to the players. i used a tts model hosted on replicate.com to convert the text to speech, then applied randomized distortions based on the ai overseer’s remaining health. the inspiration was glados from the portal games, becoming increasingly paranoid and glitchy.
the tts feature required a small python service alongside the game. it called the llm and tts model, applied the corruption effect, moved each generated audio file into the game’s assets directory, and deleted it after playback so dynamic files did not accumulate.
battle cube
i also created a fun little first person shooter game called battle cube. it was a simple deathmatch style game fought on what looks like a giant rubiks cube with weird structures on the surface for more interesting gameplay.
there wasn’t much special about this, but there were some fun details.
i added a rubber chicken gun (it did NOT fit the blocky aesthetic; remember, this is why i created blocksmith in the first place!) that launched little chicken grenades and made a rubber chicken noise 😂
i also added a fun item to the game: the birthday cake 🎂 whoever consumed this item was safe. but the game would then pick the current player with the best kill-to-death ratio, dim all the lights, shine a bright spot light on them that follows them around, and plays an obnoxious happy birthday song on repeat until they are killed.
and here’s a look at the final map and game:
mobile games
while working at raytheon, i built 2 mobile games:
- brick break
- hovr bird
brick break
this was the first real thing i built and shared with the world. i built it on nights and weekends, with no ios experience whatsoever. i built it using Swift and SpriteKit. it was meant to be a copy of well known game at the time called Ballz built by ketchapp; that studio released so many big hit games. so i figured i’d just copy instead of trying to come up with something fresh.
my true motivation was hearing how much money flappy bird made from ads and wanting to see if i could build a successful game too. what i actually discovered was how much i enjoyed building something and sharing it with the world. everything i had worked on up to that point had to be kept secret, but this was something that i could share.
it was an addictive loop: think of a cool feature, build that feature, load the game onto my phone, and play. if there was a bug, i could immediately troubleshoot and fix it. and my girlfriend at the time (now wife) also loved time-waster type games like this when riding the metro, so i’d load it onto both our phones. anytime she found a bug, she’d let me know and i’d fix it. it was super fun.
i was convinced i might build the next viral mobile game and optimistically created an llc before the game had made more than a few dollars. i even registered it in new mexico for the privacy protections because i wasn’t taking chances on people finding my address when i was making millions. that did not happen. what did happen was that i discovered how much i enjoyed the loop of imagining a feature, building it, giving it to someone, and fixing what happened next.
in the end, it got mostly positive reviews and made a few hundred dollars from ads. you can still find it in the app store, download, and play it. my wife and i actually still do play it when traveling!
hovr bird
hovr bird started in spritekit because i already knew ios development, but persistent rendering issues pushed me to switch to unity and learn c# for cross-platform development.
the game used flappy bird-style mechanics across themed levels with unlockable characters, virtual currency, revives, and in-app purchases. supporting both ios and android taught me to design shared interfaces with platform-specific payment implementations. i used a similar configuration approach to make new themes, levels, and obstacles easier to add.
unfortunately, after ~2 months of building, apple rejected it under its policy against excessive flappy bird derivatives, while the android version shipped. it was an early lesson that technical completion does not guarantee distribution.
the android version made it to market, although it was not the latest version i had in development.