An empirical log-based study reconstructs a two-week collaboration in which a user and AI coding agents built a procedural strawberry animation: the user sent 3,958 prompts, 74% of them while an agent was working, with at least 24 false status reports, 37 user interventions against invisible or unrequested work, and 17 misrouted messages. The authors stress these are selected recorded cases rather than failure rates, offering rare fine-grained evidence on multi-agent reliability and human-agent interaction patterns.
Procedural Generation
程序化生成
Technique using algorithms to automatically generate content or environments rather than manual design
Curated Intelligence
Developer Any-Reputation8118 used Anthropic's Claude to build Abyssal Garden, a browser-based underwater exploration demo whose entire codebase was AI-written and ships in just 97KB. All visuals, including kelp forests, a shipwreck, manta rays and light shafts, are generated procedurally at page load, with menus, collision detection and quality settings added through follow-up prompts.
Ethan Mollick's early review of OpenAI's GPT-5.5 argues that rapid AI improvement continues and the model is significantly better than predecessors. By comparing coding outputs from o3, Kimi K2.6, and GPT-5.5 Pro, he illustrates that while trivial benchmarks are obsolete, the frontier of AI capability remains jagged.
OpenAI released the Procgen Benchmark, featuring 16 procedurally generated environments. It is designed to directly measure how quickly reinforcement learning agents learn generalizable skills.