potentially hundreds of independently behaving creatures. And all of that needs to happen in real time. It’s one thing to create a beautiful environment. It’s another to
make that environment dynamic, interactive and computationally efficient. A lot of my development effort has gone into performance
optimisation, multithreading, simulation scheduling and extending the engine to support the kind of world I want to build. I’ve had to become a game designer, engine programmer, technical artist, systems architect and occasionally my own worst critic. But that’s also what I love about independent development. You’re constantly learning, experimenting and solving problems that initially seem impossible. Interestingly, one of those problems would eventually become something much bigger than the game itself.
THE UNEXPECTED BIRTH OF LUMINIDSAI As I developed the behaviour systems behind Luminids, I kept running into a fundamental question: what actually makes an artificial character feel alive? Traditional game AI is generally very good at deciding what an agent
should do next. Find food. Complete a task. Avoid danger. Return home. But there’s a difference between an agent that performs intelligent actions and one whose decisions are shaped by its own experiences. Consider two creatures witnessing the same event. One might
interpret it as something wonderful. Another might find it frightening, based on something that happened previously. Their memories, expectations and relationships could lead them to behave entirely differently. And importantly, both reactions could make sense. That fascinated me. I started exploring how artificial agents could develop their own
perspectives, remember experiences, form beliefs and relationships, and change their behaviour over time. That research became the foundation of LuminidsAI.
CONTEXTUAL RELATIONAL INTELLIGENCE LuminidsAI is developing something I call Contextual Relational Intelligence, or CRI. The premise is that artificial intelligence needs more than information, reasoning and the ability to generate convincing responses. It needs continuity. It needs to understand not just what’s happening, but what has happened before, who was involved, how those experiences were interpreted and why they matter to the individual. CRI explores a structured approach to modelling experiences,
memories, beliefs, expectations, relationships and emotional appraisal. Rather than treating every interaction as an isolated problem, the aim is to give an artificial agent a history of its own. A history that genuinely influences its decisions. For example, imagine asking an AI character why it doesn’t trust
another character. I don’t want it to invent a plausible explanation after the fact. I want it to explain that it remembers a particular experience, formed a belief from that event and now expects a certain outcome. And if new experiences contradict that belief, I want it to be capable of changing its mind.
That’s the difference I’m interested in exploring. Not simply making AI
appear more human, but developing systems where behaviour emerges from persistent, individual context.
A MUCH BIGGER AMBITION Although the original inspiration came from Luminids, the potential applications extend well beyond video games. Imagine AI companions that develop meaningful continuity across
interactions. Virtual characters with consistent personalities and evolving relationships. Training simulations where individuals respond according to their own experiences. Or AI assistants that can interpret a person’s circumstances with greater contextual understanding, rather than relying solely on the immediate conversation. We’re entering a world where AI is becoming increasingly capable, but intelligence alone doesn’t necessarily produce understanding.
I BELIEVE AI LEARNED LANGUAGE BEFORE IT LEARNED RELATIONSHIPS We’ve become extraordinarily good at building systems that can answer questions. I’m interested in what happens when those systems can also understand context, experience, individuality and the consequences of their decisions. That’s the vision behind LuminidsAI. It’s an ambitious research and engineering challenge, and there’s a tremendous amount still to prove. But I believe it’s a direction worth exploring.
WHAT COMES NEXT? My immediate focus is bringing Luminids to players. The public demo is an important milestone, followed by our planned Early Access release on Steam in February 2027. There are still countless things I want to build, improve and discover, but I couldn’t be more excited about where the project is heading. In parallel, I’m continuing to develop LuminidsAI and the underlying
CRI technology, with the ambition of making it accessible to other developers and organisations. It’s quite extraordinary to think that a little idea about tiny creatures living in a peaceful world could lead to an entirely new direction in artificial intelligence research. But perhaps that’s what I love most about building things. You start with a question, follow your curiosity and sometimes discover something you never expected. And for me, this is still only the beginning.
November/December 2026 MCV/DEVELOP | 29
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