Storming Kaggle Cayley puzzles 555 and 666 with Claude vs Codex
AndLukyane · x · 2026-08-23
The author tackled Kaggle Cayley puzzles 555 and 666 using Claude and Codex, and explains why they're harder:
- Larger solution space and diameter hurt: models move toward solutions iteratively, so each extra step compounds error; longer paths make correct labels scarcer, and random walks get lost or fail to reach distant points.
- The pipeline reuses earlier approaches: initial model training, then polishing and beam search; a transformer was dropped as too slow.
- After several days of training on an A100, the resulting model doesn't solve every puzzle but performs consistently.
More from coding & agent
- Developer calls MCP research paper error-riddled and partially AI-generated — benfielding · 2026-08-23
- Connect Claude Code to secure remote MCP servers on Google Cloud Run — fhinkel · 2026-08-23
- Executor Gateway Cuts Tool Context Usage from 278k to 1k Tokens — ethanniser · 2026-08-23
- Non-tech clients burned by using Claude as runtime layer — DuaneJRich · 2026-08-23
- Agent-Aware Infra: Optimizing Inference via Cache and Scheduling — _ScottCondron · 2026-08-23
- Claude auto-replies to customers: From draft to send, risks and governance — Healthy_Condition779 · 2026-08-23