Agentic Game Development as a Verifiable Trajectory Data Engine for Scaling World Models

  • 类型:arxiv
  • 标识:2608.25518
  • 链接:https://arxiv.org/abs/2608.25518
  • 主分类:agent
  • 形态:method
  • TLDR:A common strategy for scaling world models is to train on more crawled video with more compute. We argue that this strategy is inefficient: scaling world models also requires a recursive data engine that offers grounded reward signals. The success of code agents illustrates why this matters. As code is executable, compilers and runtimes can provide high-quality rewards for Reinforcement Learning (RL) post-training of LLMs. By contrast, spatial generation still relies largely on fuzzy proxies such as CLIP scores. These signals are fuzzy and biased, making them hard to support RL post-training.
  • 副分类:engineering
  • 待LLM分类:否
  • 来源文件
  • /inbox/tom/_candidates/2026-08-28-agent-rag-longcontext-candidates.json