Ryan Morrison’s Post

How to make fun, mini AI games? Chris Benjaminsen wrote a simple toy game engine called 100 Boxes to explore whether FRVR Forge (his text-to-game AI) works with other game engines. With it, he created a series of mini games, with the only commonality being that the game is over if you fail. While the games were fun and successful, the process exposed some weaknesses for us to improve on: - LLM integration is really bad at math related to rotation. This was opaque as XS (FRVR’s XS game engine) supports changing the anchor point. - Good prompt engineering does not make up for bad API design and examples. - FRVR Forge really likes rewriting code unnecessarily. It will often delete 10+ lines to change a single one. Along with 100 Boxes, he wrote a small wrapper in XS. It allows the games to be played via our existing web publishing technology. He then tied them all into a single game, where the goal is to survive as many seven-second rounds of playing one mini-game in a row as possible. And the game does get faster over time! A fun AI learning vacation project, eh? 🔗 Link for the full article in comments. #ai #artificialintelligence #LLMs

Sometimes when I think of a cool demo, it doesn't actually work out well, but I still learn something. I think it's important to share failures just as much as successes. Here's an exploration into generative AI to help with gaming, in this case, blackjack. Enjoy! https://lnkd.in/g-bvuTeq #generativeai

This just in - GamesIndustry.biz presents... The Year In Numbers 2023  GamesIndustry.biz - and consider kaedim3d.com for AI that turns images into 3d game assets in minutes

Guess what? - GamesIndustry.biz presents... The Year In Numbers 2023  GamesIndustry.biz - and consider kaedim3d.com for AI that turns images into 3d game assets in minutes

How about this - GamesIndustry.biz presents... The Year In Numbers 2023  GamesIndustry.biz - and consider kaedim3d.com for AI that turns images into 3d game assets in minutes

PDF Introduction to Game AI Neil Kirby digsell https://lnkd.in/eQSgHEyD Today’s game players expect increasingly realistic interaction within games. “Introduction to Game AI” teaches readers who are new to game AI the skills they need through hands-on projects based on small, understandable games. While there are many books that take a more advanced approach to this topic, this is one of the only books that … Read More » https://lnkd.in/eTpNe5F5

In open-world #games, ensuring they are playable and bug-free is crucial, but is becoming increasingly difficult and time-consuming using manual #gametesting. Maximizing exploration using #ReinforcementLearning is extremely useful for testing games at scale, because of the wide variety of gameplay scenes the player may encounter.   We used our #opensource library #SheepRL 🐑 and our #PyTorch implementation of #DreamerV3 on #Crafter, an open-world survival game, featuring randomly generated 2D worlds, in which players have the freedom to explore a large and expansive map and need to forage for food, collect materials, build tools and find shelter.   Why is the test on Crafter so interesting for game debugging?   Because Crafter evaluates a large number of general capabilities related to the RL agent, like strong ability to generalise (new generated maps for each episode), to deal with partial observability (each input image reveals only a small part of the world) and to long-term reasoning and survival.   These abilities are very useful for testing games at scale, providing #developers with insights to optimise gameplay and player experience.   The future of #game testing is here, and it is powered by #ArtificialIntelligence! 🔥 --- ❌ Are you interested in joining the project community? Get in touch 👉https://lnkd.in/djBf-6ZG ❌ SheepRL 🐑 is open-source, fully written in PyTorch and accelerated with #LightningFabric - by Lightning AI Feel free to use it for your #AI projects, and if you want to contribute, we are more than happy to accept your pull requests! ❤️ #opensourcecommunity #gamingcommunity #WeMakeAIhappen #gaming Federico Belotti, Davide Angioni, Refik Can MALLI, Michele Milesi, Luca Antiga

Breaking News - GamesIndustry.biz presents... The Year In Numbers 2023  GamesIndustry.biz - and consider kaedim3d.com for AI that turns images into 3d game assets in minutes

🎮 Dynamic Gameplay with State Machine Behavior in Games! 🕹️ Hey, #GameDevCommunity! 👋 Ever wondered how game characters seamlessly transition between different behaviors, creating immersive experiences? Let's dive into the fascinating world of State Machine Behavior and how I implemented it in my very own BattleTank game! 🔄 Understanding State Machine Behavior: State Machine Behavior is like a conductor guiding characters' actions in a game. It allows entities to smoothly transition between various states, reacting to changing conditions and delivering a lifelike gaming adventure. Imagine your enemy tanks shifting from idle mode to patrolling, then fiercely chasing the player and ultimately launching a relentless attack! 🚀 Implementing State Machine in BattleTank: In my BattleTank game, I embraced the power of State Machine Behavior to give enemy tanks a dynamic edge. Here's how I orchestrated it: 1. Idle State: When no target is detected, enemy tanks enter an idle state. They scan their surroundings, creating a sense of realism while waiting for potential prey. 2. Patrol State: Tanks then transition to the patrol state, moving along predetermined paths. This state adds unpredictability and keeps players on their toes, expecting the unexpected! 3. Chase State: When a player enters the detection zone, enemy tanks shift gears into chase mode. They follow the player's position, adding an element of danger and excitement. 4. Attack State: Finally, when within firing range, tanks unleash their firepower in the attack state. This state showcases their aggression, challenging players to devise strategies for survival. 🌟 Advantages of State Machine Behavior: 1. 🕹️ Immersive Gameplay: Smooth transitions between states create a lifelike world where characters respond realistically to changing conditions. 2. ⚙️ Modular Design: Each state's logic is contained, making it easier to modify, expand, or add new states as the game evolves. 3. 🎯 Enhanced AI: State Machine Behavior empowers AI characters to make intelligent decisions based on context, enhancing the challenge for players. 4. 📈 Game Balancing: Fine-tuning each state's parameters allows for precise balancing, ensuring an engaging and enjoyable experience. Share your thoughts on State Machine Behavior in the comments! Have you encountered it in other games? Are you working on similar techniques? Let's chat and level up our game development skills together! 🚀🎮 Check out my DevLog: https://lnkd.in/gapAPUX3 Outscal Mayank Grover Abhishek Smith Malhar Devasthali Manas Gangwani Rohith Kannan Shubham K. Ravi Kiran Svs Shreyash patil Unity #GameDevelopment #StateMachine #IndieGameDev #BattleTankGame #GamingCommunity #gamedev #unity

Did you know - GamesIndustry.biz presents... The Year In Numbers 2023  GamesIndustry.biz - and consider kaedim3d.com for AI that turns images into 3d game assets in minutes