SPADE: Self-Play Framework Generates Adaptive Environments for Continuous Agent Improvement
_AndrewZhao · x · 2026-08-21
The paper introduces SPADE (Self-Play in Adaptive Synthetic Executable Environments), a framework enabling continuous self-improvement for LLMs via self-play. The LLM acts as both an 'Environment Designer,' writing executable training code with OpenAI Gym-style interfaces, state transitions, and reward functions, and a 'Reasoning Agent' that acts within these environments. By optimizing the Agent's 'regret' signal—the gap between performance with and without privileged hints—the Designer learns to generate adaptive, diverse goals at the edge of the Agent's capabilities, overcoming the fixed goal distribution limitations of static environments.
Related event: SPADE: Single-Model Self-Play Writes Its Own Training Environments(13 posts)→
More from coding & agent
- Apple code signing drags for hours; official advice: just wait a week — pjausovec · 2026-10-16
- Hermes: a 4B model with 8K context got worse with more test-time compute, pointing to contextual reasoning — mangahomanga · 2026-10-16
- Flux 3 Klein's Default Templates Are Broken Without Its Prompt Upscaler — StrictMove5839 · 2026-10-16
- 21-person team ships hundreds of PRs weekly with AI writing 97% of the code — rseroter · 2026-10-16
- ClosedMCP: Set Up MCP Servers Once, Share Them Across Your Whole Team — Ok_Sprinkles1301 · 2026-10-16
- Heisentools: let an agent rewrite a mock.py until it mirrors any real tool's API — kieranklaassen · 2026-10-16