I am one of the authors from Sakana AI and MIT. We just released this paper where we hooked up LLMs to the classic 1984 programming game Core War. For those who haven't played it, Core War involves writing assembly programs in a language called Redcode that battle for control of a virtual computer's memory. You win by crashing the opponent's process while keeping yours running. It is a Turing-complete environment where code and data share the same address space, which leads to some very chaotic self-modifying code dynamics.
We did not just ask the model to write winning code from scratch. Instead, we treated the LLM as a mutation operator within a quality-diversity algorithm called MAP-Elites. The system runs an adversarial evolutionary loop where new warriors are continually evolved to defeat the champions of all previous rounds. We call this Digital Red Queen because it mimics the biological hypothesis that species must continually adapt just to survive against changing competitors.
The most interesting result for us was observing convergent evolution. We ran independent experiments starting from completely different random seeds, yet the populations consistently gravitated toward similar behavioral phenotypes, specifically regarding memory coverage and thread spawning. It mirrors how biological species independently evolve similar traits like eyes to solve similar problems. We also found that this training loop produced generalist warriors that were robust even against human-written strategies they had never encountered during training.
We think Core War is an under-utilized sandbox for studying these kinds of adversarial dynamics. It lets us simulate how automated systems might eventually compete for computational resources in the real world, but in a totally isolated environment. The simulation code and the prompts we used are open source on GitHub.
We are the team at Sakana AI. To give some context on the difficulty here, an OpenAI agent placed 2nd in the AHC world tournament last August, so taking 1st place against 804 humans in this contest is a significant milestone for us. Our agent approached the production planning problem by running its own experiments during the contest. It independently discovered a Simulated Annealing strategy using a "virtual power" heuristic which ended up outperforming the greedy solutions that the problem setters anticipated.
We used inference-time scaling with GPT-5.2 and Gemini 3 Pro Preview to make this happen. The agent ran parallel code generation loops to iteratively refine the algorithm, costing about $1,300 in total compute for the 4 hour event. We published the full logs showing the agent's analysis and code evolution at the link in the post.
Happy to answer any questions about the architecture!