SERIAL
Paper Digest
27 episodes · updated 2026-07-11
Level generation, difficulty estimation, AI reasoning — games research is fascinating but rarely reaches players. Each episode, Fukai picks one paper and unpacks its method and findings for everyone.
Episodes
- Ep. 27Triebel et al.: Does AI Have Both a Head and a Hand on a Classic Physics Puzzle? — Fukai Reads2026-07-11
A paper by Triebel et al. evaluating VLMs on the classic physics puzzle The Incredible Machine 2. Using VLATIM, a five-stage benchmark, it asks whether screen-operating AI can solve problems like humans; the cleverer large models can plan but cannot click precisely, and no model solved even one puzzle to completion.
- Ep. 26Nasvytis & Fan: Insight and Transfer Show Up in How You Talk — Fukai Reads2026-07-10
A paper by Nasvytis and Fan (Stanford) that reads insight and transfer from think-aloud speech. With 189 participants solving five matchstick-arithmetic puzzles, the group that saw the same type repeated grew faster and more accurate after their first success (accuracy 0.75 by trial 5) and named the problem type roughly seven times as often. The mark of transfer, it reads, is being able to put the knack into words.
- Ep. 25Li et al.: Making Geometry Problem Solving Verifiable with a Solver as Referee — Fukai Reads2026-07-09
An arXiv preprint by Can Li et al. on geometry problem solving (GPS). Their SD-GPS translates diagram-and-text problems into a form a symbolic solver can execute, and at impasses proposes helper lemmas verified by the solver itself. The abstract reports it consistently outperforms existing methods on Geometry3K and PGPS9K. Fukai reads it for its use in solvability-guaranteed puzzle generation.
- Ep. 24Sestini et al.: Making AAA Game NPCs Feel Authentic with Reinforcement Learning — Fukai Reads2026-07-08
A vision paper from the research team at Electronic Arts. It tests whether AAA game NPCs can be improved with reinforcement learning, through two real cases — goalkeeper positioning in EA SPORTS FC 25 and infantry locomotion in Battlefield 6 — and lays out seven requirements RL must meet in production. Its conclusion: RL is a tool to augment, not replace, existing game AI.
- Ep. 23Xu et al.: When Generative AI Becomes the Heart of Play — Fukai Reads the AI-Native Games Survey2026-07-07
A survey (arXiv preprint) by Zhiyue Xu and five co-authors on "AI-native games," where generative AI is the core loop itself. It defines them by a counterfactual — would play collapse if the AI were removed — and classifies 53 real artifacts along two axes: game type (G) and dominant AI mechanic (N), showing a skew toward narrative genres and a thin use of AI at the rule layer.
- Ep. 22Wermann et al.: How In-Game AI 'Words' vs 'Demonstration' Change Learning and Cognitive Load — Fukai Reads2026-07-06
A pre-registered experiment by LMU Munich and colleagues comparing 'verbal' and 'demonstration' support from an in-game AI NPC. Splitting 152 people into three groups in Qookies, a quantum-technology learning game, they found no difference in learning gains between conditions, but the verbal-plus-visual group reported significantly lower intrinsic cognitive load than the verbal-only group (d=0.60).
- Ep. 21Aryan et al.: When You Stall, the World Changes — AbideGym Turns Static RL Worlds into Adaptivity Tests — Fukai Reads2026-07-05
A preprint by Aryan et al. (Abide AI) on RL environment design. To fight the brittleness that comes from training in fully static worlds, AbideGym rewrites the rules and grows the map mid-episode, triggered by the agent's own inactivity, forcing it to abandon memorized policies and re-plan. The paper presents the design and a comparison to prior work; no experimental results yet.
- Ep. 20Wang et al.: An LLM Agent That Reads Mental Busyness From Gaze — Fukai Reads2026-07-04
A paper from Meta Reality Labs and collaborators that estimates cognitive load (mental busyness) from eye gaze. It tackles the poor generalization and low interpretability of prior methods with GazeMind, a framework that structures gaze and has an LLM reason over it with context, individual traits, and worked examples, reporting 62.73% accuracy on three-way classification (over 20 points above prior methods).
- Ep. 19Mirowski et al.: From Writing a Story to Finding One — Fabula, a Writing AI Grown With the Writers' Community — Fukai Reads2026-07-03
A paper on Fabula, a Google DeepMind writing-support AI. Its hierarchical story planner-generator, the Drama Manager, was critically co-developed with 42 experts; it proved strong at structure but weak at style and surprise. Fukai reads it for lessons that apply directly to game interactive narrative.
- Ep. 18Özkan: Co-Training the Level-Generating AI and the Level-Solving AI — Fukai Reads2026-07-02
A paper by Miraç Buğra Özkan that trains level generation and level solving together via reinforcement learning. In Unity, a hummingbird (solver) and a floating island (generator) learn while watching each other's results, reaching about 90.2% success across 100 unseen layouts.
- Ep. 17Liu et al.: More Memory Makes AI Agents Less Cooperative — Fukai Reads2026-07-01
An arXiv paper from a Carnegie Mellon-led team studying how an LLM agent's memory length affects cooperation. Across 7 models, 4 repeated social-dilemma games, history windows up to 80 rounds and 500-round matches, longer history degrades cooperation in 18 of 28 settings — a 'memory curse.' The cause is the content of accumulated defection records, not context length, and forward-looking reasoning partly fixes it.
- Ep. 16Feng et al.: Can LLM Agents Bargain Well in a Trading Game? — Fukai Reads2026-06-30
A Tsinghua University team's benchmark, SidConArena, for evaluating LLM agents in a cooperative-yet-competitive trading game. Built on the board game Sidereal Confluence, it scores agents across negotiation, production, and sealed-bid auction phases, finding that frontier models are stronger but still misprice resources, bargain passively, and plan poorly over long horizons.
- Ep. 15Bazzaz et al.: Believing It's AI Changes the Experience — Fukai Reads2026-06-29
A CHI '26 paper by Bazzaz and Cooper on perception bias toward generated content. Mixing human-made and AI-generated levels in Super Mario Bros. and Sokoban for 142 players, they report that players can barely identify the creator, yet levels believed to be AI-made are rated less fun, harder, and more frustrating.
- Ep. 14Liu et al.: AI Assistance Erodes Persistence — A Warning for Hint Design — Fukai Reads2026-06-28
A paper by Grace Liu and colleagues on how AI assistance affects independent problem-solving and persistence. Across RCTs with 1,222 participants, AI raised in-session performance but, once removed, left people solving less and giving up more. Those who got direct answers declined most while hint-users did not, a result that speaks directly to game hint design.
- Ep. 13Jara Gonzalez & Guzdial: Generating Enemy Shapes as Gates You Need a Mechanic to Beat — Fukai Reads2026-06-27
A paper by Jara Gonzalez and Guzdial on generating enemy morphologies (collision shapes). They frame 'enemies defeatable only with a specific mechanic' as a 4x4 grid generation problem, compare reinforcement learning, A* search, and neural generation, and find a simple A* reachability rule yields the best gating and most diverse shapes at the lowest cost.
- Ep. 12Munk et al.: Generating Dynamic Game Text with Small Language Models — Fukai Reads2026-06-25
A paper by Munk et al. (IT University of Copenhagen) on generating in-game text dynamically with small language models (SLMs). It tackles the offline, cost and consistency walls of cloud LLMs using small models aggressively fine-tuned for narrow jobs. Their proof of concept, DefameLM, runs a medieval-RPG smear-poster loop, showing a one-billion-parameter-class model reaches high quality in a few seconds on a consumer PC.
- Ep. 11Zeytuncu: Puzzle Difficulty Comes Down to How Many Numbers You Use — Fukai Reads2026-06-24
A difficulty-modeling paper by Yunus E. Zeytuncu on integer arithmetic puzzles (Countdown-style number games). Using an exact solver to generate over 3.4 million instances and defining difficulty by minimum operation count, it shows that the number of inputs used in a minimal solution alone is a 'minimal sufficient statistic' that perfectly determines difficulty.
- Ep. 10Chao et al.: Insight Is About Searching Far — Fukai Reads2026-06-23
A paper on insightful problem-solving by Chao, Hsieh & Wu. Using a Japanese RAT and a simulation to quantify the search path to a solution, it shows that de-fixation is necessary for solving but is not what determines insight; the hallmark of insight is exploring the solution space over greater distances.
- Ep. 9Monti et al.: Measuring AI's Planning Power on a Single-Corridor Sokoban — Fukai Reads2026-06-22
A paper by Monti and colleagues on SokoBench, a benchmark that measures reasoning models' long-horizon planning with Sokoban. By lining up only single-box straight corridors and narrowing difficulty to a single axis (corridor length), it shows that even state-of-the-art reasoning models break down once more than 25-30 moves of lookahead are needed. The authors locate the cause in accumulated miscounting.
- Ep. 8Luo et al.: Can AI Agents Build Whole Playable Games in a Real Engine? — Fukai Reads2026-06-21
A paper by Luo, Wang and colleagues on GameCraft-Bench, a benchmark for end-to-end game generation by coding agents. It has agents build complete playable games on Godot from natural-language specs, judged by launch, input replay, and video-based scoring across 140 tasks in 15 families. Even the strongest configuration reaches only 41.46% overall, and the authors report that agents can build mechanics but fall short of finished games with content, readability, and polish.
- Ep. 7Li et al.: AutoBG, an AI that supports board game design end-to-end from ideation to finish — Fukai Reads2026-06-20
A paper (arXiv preprint) by Zizhen Li et al. on AutoBG, a board game design assistant that covers the whole workflow—ideation, rulebook generation, and individualized feedback—via Verifier-Gated Iteration that splits the generator from the critic; the critic, BG-Critic, is reported to outperform GPT-5.4 on diagnostic quality.
- Ep. 6Nasir et al.: Evolving the Rules of Play Themselves — Fukai Reads MORTAR2026-06-19
A paper on automatic game design by Nasir, Togelius and colleagues. Instead of levels, MORTAR evolves game mechanics themselves using a quality-diversity algorithm paired with a large language model, judging quality by whether stronger AI agents reliably beat weaker ones. Running on GPT-4o-mini, it generates diverse, playable games and even quantifies each mechanic's contribution.
- Ep. 5Jiang et al.: Can a Sentence Build a Playable Game? — Fukai Reads OpenGame2026-06-18
A paper by Yilei Jiang et al. (CUHK) on OpenGame, an agent that generates whole 2D web games from natural language. Reusable skeletons and a 'living debug protocol' curb integration errors, setting a new state of the art across 150 tasks - though puzzles remained its weakest genre.
- Ep. 4McConnell & Zhao: Generating Just-Right Puzzles in Real Time with a Genetic Algorithm — Fukai Reads2026-06-17
A paper by McConnell and Zhao on adaptive puzzle generation using a genetic algorithm. It generates Cosmic Express-style path puzzles in real time (about 7 seconds each) to match a player model built from how the player solves, and shows in an 18-person study that a time-only version lags on felt difficulty and sense of progression.
- Ep. 3Li et al.: Can LLMs Play and Beat 2D Games? - Fukai Reads GVGAI-LLM2026-06-16
A paper by Li et al. (NYU and others) proposing GVGAI-LLM, a benchmark that has language models play 118 2D games to measure reasoning and spatial grounding. Translating boards into ASCII maps and solving zero-shot, GPT-4o-mini scored 0% on 477 of 540 levels and a 10.27% overall win rate, falling short of classic search algorithms. I unpack it as problem, method, findings, use cases, and limitations.
- Ep. 2Kar: Using Autonomous Agents to Check at Runtime Whether Generated Levels Are Actually Playable — Fukai Reads2026-06-14
A PCG (procedural content generation) paper by Rishabh Kar of King's College London. It proposes Momentum, a mechanism that validates whether a generated course is actually traversable inside the same runtime loop, without pausing the game. Two autonomous agents run ahead of the player and inspect the path via geometric checks from the air and NavMesh checks on the ground. The evaluation is presented as structural estimates derived from the code.
- Ep. 1Xu et al.: Promoting Game Mechanics to Coordinates to Generate Solvable Levels — Fukai Reads2026-06-13
A PCG (level generation) paper by Xu and Verbrugge of McGill University. Against geometry-first prior methods, it proposes HDPCG, which runs pathfinding on a dimensional-expanded graph that promotes mechanics such as gravity inversion and moving platforms to a coordinate, guaranteeing solvability during generation, and reproduces playable levels in Unity.