techniques 45 records
Techniques
Every mathematical, AI, ML, and LLM technique in the set, with its maturity in games and what it does to reproducibility.
| name | category | year | maturity | determinism | mechanics |
|---|---|---|---|---|---|
| 3D Gaussian Splatting | Neural rendering & 3D generation | 2023 | Prototype | Preserves | 2 |
| Agent-based automated playtesting and balancing | Automated game design & playtesting | 2018 | Shipped (niche) | Preserves | 6 |
| Automated generation of game rules and mechanics | Automated game design & playtesting | 2024 | Prototype | Preserves | 5 |
| Continuous cellular automata and foundation-model search for artificial life | Artificial life research | 2024 | Research | Preserves | 3 |
| Deep reinforcement learning policies for NPC opponents and navigation | Reinforcement learning | 2022 | Shipped (niche) | Mitigable | 3 |
| Differentiable simulation | Differentiable & accelerated simulation | 2020 | Research | Preserves | 6 |
| Diffusion and autoregressive neural game engines | Generative world model | 2024 | Prototype | Breaks | 5 |
| Generalist instructable game agents (SIMA) | LLM agents | 2024 | Research | Mitigable | 5 |
| Generative agents (memory stream, reflection, planning) | LLM agents | 2023 | Shipped (niche) | Breaks | 9 |
| Genie foundation world models | Generative world model | 2025 | Prototype | Breaks | 7 |
| Grammar-constrained decoding for valid game actions | LLM agents | 2024 | Research | Mitigable | 5 |
| Graph network simulators and neural operators | Neural physics & surrogates | 2020 | Research | Mitigable | 6 |
| LLM explanation of simulation state | LLM narrative & dialogue | 2023 | Prototype | Preserves | 5 |
| LLM game master and narrative director | LLM narrative & dialogue | 2019 | Shipped (niche) | Breaks | 6 |
| LLM models of individual human choice | Player modeling | 2025 | Research | Preserves | 9 |
| LLM-adjudicated state transitions (LLM as game mechanic) | Generative world model | 2024 | Shipped (niche) | Mitigable | 5 |
| LLM-guided evolutionary program search | Mathematical & algorithmic discovery | 2025 | Research | Preserves | 6 |
| LLM-written skill libraries (Voyager) | LLM agents | 2023 | Research | Mitigable | 5 |
| Language models grounded in strategic planning for negotiation games | Game-theoretic solvers | 2022 | Research | Breaks | 5 |
| Learned kinematic motion synthesis | Animation & motion | 2020 | Shipped (niche) | Preserves | 2 |
| Learned level and map generation (PCGRL, MarioGPT, LLM generators) | Procedural content generation | 2023 | Research | Mitigable | 4 |
| Learned physics-based character control | Animation & motion | 2024 | Research | Mitigable | 4 |
| ML reconstruction of the planet from aerial imagery | Procedural content generation | 2020 | Shipped (niche) | Preserves | 3 |
| Machine-learned weather forecasting | Neural physics & surrogates | 2023 | Research | Mitigable | 5 |
| Many-agent LLM societies | Multi-agent social simulation | 2024 | Research | Breaks | 8 |
| Massively parallel GPU simulation | Differentiable & accelerated simulation | 2021 | Research | Preserves | 8 |
| Mean field games and mean-field learning | Game-theoretic solvers | 2021 | Research | Preserves | 6 |
| Monte Carlo tree search for game AI | Game-theoretic solvers | 2006 | Shipped (niche) | Mitigable | 5 |
| Muse World and Human Action Model (WHAM) | Generative world model | 2025 | Prototype | Breaks | 3 |
| Neural cellular automata | Artificial life research | 2020 | Research | Mitigable | 4 |
| Neural rendering in the graphics pipeline | Neural rendering & 3D generation | 2025 | Shipped (mainstream) | Preserves | 3 |
| On-device small language models for characters | Inference efficiency | 2025 | Shipped (niche) | Mitigable | 5 |
| Open real-time interactive video world models | Generative world model | 2025 | Prototype | Breaks | 6 |
| Open-ended co-generation of environments and agents | Reinforcement learning | 2024 | Research | Mitigable | 5 |
| Physical-world foundation models | Generative world model | 2025 | Research | Breaks | 3 |
| Prompt-to-persistent 3D world generation | Neural rendering & 3D generation | 2025 | Prototype | Preserves | 4 |
| Quality-diversity search (MAP-Elites) for content and balance | Procedural content generation | 2015 | Research | Preserves | 6 |
| Real-time generative restyling of game video | Neural rendering & 3D generation | 2025 | Prototype | Mitigable | 2 |
| Realtime speech-to-speech and expressive TTS for characters | Speech & audio | 2024 | Shipped (niche) | Breaks | 4 |
| Regret minimization and search for imperfect-information games | Game-theoretic solvers | 2019 | Research | Mitigable | 6 |
| Self-play planning with learned models (AlphaZero, MuZero, Dreamer) | Reinforcement learning | 2020 | Research | Mitigable | 5 |
| Simulation-based inference for calibrating simulators | Mathematical & algorithmic discovery | 2019 | Research | Preserves | 6 |
| Subspace neural physics and learned deformers | Neural physics & surrogates | 2019 | Prototype | Preserves | 2 |
| Text- and image-to-3D asset generation | Neural rendering & 3D generation | 2025 | Shipped (niche) | Mitigable | 3 |
| Wave Function Collapse and constraint-based tile generation | Procedural content generation | 2016 | Shipped (niche) | Preserves | 3 |