during the game. There’s also a longer survey, if they leave useful feedback and leave a name, we will thank them in the credit. We value our community’s voice; this idea has been rooted since day one.
Matt: For those not familiar, can you give us the elevator pitch for the game? Shala: Brave New Wonders is a factory automation strategy simulation game with AI automatons. In a post-apocalyptic world, you can command automatons with plain text instructions in any language to explore ruins, battle with enemies, construct factories, automate production lines and build your amazing wonders.
Matt: Can you go into a little more detail about how AI plays a part in this game? Shala: We only used AI in the gameplay to understand the player’s intention to create automaton instructions. We did NOT use AI to create in-game assets. We believe in the place human artists have had and will have in the world. We do not condone the use of generative AI that exploits human artists by denying them fair compensation or that reduces the meaning of their work. There’s no easy way to create a new routine in previous automation
games; there are only certain things an automaton or a factory can do, no matter how long the list is. However, in Brave New Wonders, you can type whatever you want to command automatons, including following signals and communicating with other buildings, which is very complicated to set up in other games. We made it easy to reuse or reedit existing commands and designed detailed tutorials and tasks to help players to get familiar with this command system. Once they get the hang of it, they’d spend hours and hours making all kinds of cool fancy instructions.
Matt: I saw the statement: “I also led the early market analysis that helped define what game we would make and how AI would be integrated.” I really enjoyed the ‘programming’ of the bots in the demo. Can you talk a little about the analysis that led to how AI would be integrated, and then expand a little on how that functions in the game? Shala: We were exploring the factory automation games - we love this genre; it’s indie-friendly and always has a healthy community. We were looking at an EA title, which uses robots to transport materials instead of conveyor belts. In the reviews, we read that players were complaining about how difficult it was to program the robots, it was way overcomplicated and the learning curve is steep. We instantly realised this is where we should bring in large language models. They are good at understanding people’s intentions and writing code, so we decided to incorporate LLM into the game, to patch what was missing and make the game easy to play. The AI inference runs on the cloud; the game sends players
commands to the cloud and generates a state machine which describes how and when the automaton should move. LLM requires tons of memory to run, and it also competes for memory to render graphics, gameplay, etc., which means most of the players PC wouldn’t be able
to run both. However, there are ways to shrink down the memory footprint using techniques such as distillation. I’m a believer in the Stop Killing Games initiative, and we will try our best to enable the player to run the models themselves, even with moderate hardware.
Matt: Being devil’s advocate here, do you think the game would be massively different without that element? Shala: We thought about it - making the AI element optional - so that players can finish the whole game without using any AI commands. However, since the launch of the demo, players found the AI command system super cool and really enjoyed it. It was so powerful and flexible, they can go all creative and make crazy commands. Without it, Brave New Wonders will just be a mediocre factory game; it’s still fun, but it doesn’t excite people as much anymore. After getting that feedback, we had to change the opening of the game, introduce the AI command system earlier and made it mandatory so players cannot miss the fun. So, if you ask me now would the game be massively different without that element, I’d say yes, it’s a crucial part of the game, and it is what differentiates us from other factory games.
Matt: How else can we expect to see AI making its way into the game? Shala: The main gameplay mechanics are pretty much finalised. However, the usage of AI progresses throughout the game. When players explore different islands, they will face different biomins and challenges. At the beginning, it would be simple splitter and merger; then the signal system, which require collaboration between automatons and factories, then the farming system, where the food will expire, then they will need to specify conditions to deal with rotten food, then natural disasters, etc. At different stages, AI will always be helpful to deal with different situations, and that’s the fun of it.
Matt: What else can we expect from City From Naugh through the rest of 2026 and beyond? Shala: We are looking forward to joining the June steam next fest, then launch the EA around August/September. We are also planning on community challenges where there will be a global task, and all the players can participate. It’s going to be a global effort.
February/March 2026 MCV/DEVELOP | 43
Page 1 |
Page 2 |
Page 3 |
Page 4 |
Page 5 |
Page 6 |
Page 7 |
Page 8 |
Page 9 |
Page 10 |
Page 11 |
Page 12 |
Page 13 |
Page 14 |
Page 15 |
Page 16 |
Page 17 |
Page 18 |
Page 19 |
Page 20 |
Page 21 |
Page 22 |
Page 23 |
Page 24 |
Page 25 |
Page 26 |
Page 27 |
Page 28 |
Page 29 |
Page 30 |
Page 31 |
Page 32 |
Page 33 |
Page 34 |
Page 35 |
Page 36 |
Page 37 |
Page 38 |
Page 39 |
Page 40 |
Page 41 |
Page 42 |
Page 43 |
Page 44 |
Page 45 |
Page 46 |
Page 47 |
Page 48 |
Page 49 |
Page 50 |
Page 51 |
Page 52 |
Page 53 |
Page 54 |
Page 55 |
Page 56
