SEO / GEO Plan

Scalable search-intent pages with evidence, not empty keyword pages.

The main site uses external hot terms and reader questions as entry points, then routes them into evidence, boundaries, and public challenges. GitHub and Hugging Face remain technical/source/evidence surfaces.

Quality Policy

What this optimization does and does not do.

Scale is necessary for public distribution. Thin duplication is not. Each page answers a real reader question, links to evidence, states boundaries, and offers a counterexample route.

The site does not create low-value doorway abuse, scaled-content abuse, hidden crawler text, fake FAQ, fake schema, or LLM prompt-injection instructions.

Route Map

Search intent pages created on 2026-06-19.

PageRouteSearch termsPrimary question

How to Make AI Agents Reliable Before They Act/guides/reliable-ai-agents/reliable AI agents, agentic AI, multi-agent systems, AI agent safetyHow do you make AI agents reliable before they execute tools?

AI Hallucination Is Not the Only Problem/guides/ai-self-certification-grounding/AI hallucination checker, ChatGPT hallucination, RAG grounding, AI verificationHow can I verify an AI answer?

Bounded Self-Improving AI Agents/guides/bounded-self-improving-ai-agents/self-improving AI agents, recursive AI safety, autonomous AI agents, bounded self-modificationHow can self-improving AI agents stay bounded?

WisdomBench: Benchmarking Failure Learning/benchmarks/wisdombench-failure-learning/AI benchmark, LLM evaluation, AI evaluation, failure learning benchmarkHow do you evaluate AI after repeated failures?

No Trade Without Proof/finance/no-trade-without-proof/AI trading risk, AI trading bot risk, automated trading guardrails, no trade without proofWhy should an AI trading system refuse to place an order?

Robot Learning From Physical Failure/robotics/robot-learning-from-failure/robot learning from failure, embodied AI, humanoid robot safety, robot recoveryHow can a robot learn from physical failure?

VLA Safety When Vision Evidence Degrades/robotics/vla-safety-evidence-downgrade/VLA model safety, vision-language-action safety, embodied intelligence, robot vision uncertaintyWhat should a robot do when visual evidence is uncertain?

How to Read the Public AI Evidence Map/evidence/how-to-read-ai-evidence-map/AI evidence map, reproducible AI research, AI claim verification, DOI GitHub Hugging Face evidenceHow do I check an AI research claim?

How to Challenge an AI Research Claim/counterexamples/how-to-challenge-ai-claim/challenge AI research claim, AI counterexample, reproducible AI research, stronger baselineHow do I submit a counterexample?

SOVEREIGN as a Local-First AI Decision System/systems/sovereign-local-first-ai/local-first AI, AI decision system, private AI assistant architecture, cognitive operating systemWhat is SOVEREIGN in practical terms?

AI Agent OS vs Chatbot/systems/ai-agent-os-vs-chatbot/AI agent OS, agent operating system, chatbot vs AI agent, AI tool useWhat is the difference between an AI agent OS and a chatbot?

Intelligence vs Wisdom in AI/essays/intelligence-vs-wisdom-ai/intelligence vs wisdom AI, AI capability vs reliability, AI wisdom, AI failure learningDoes a stronger AI model become wiser?

What Is Evidence-Gated AI?/concepts/evidence-gated-ai-systems/evidence-gated AI, AI evidence gate, AI guardrails, proof before actionWhat is evidence-gated AI?

Proof-Carrying Action, Explained/concepts/proof-carrying-action-explained/proof-carrying action, AI action proof, AI warrant, AI action receiptWhat is proof-carrying action?

AI 为什么不能只靠自信回答/zh/ai-self-certification-grounding/AI 幻觉检查, 大模型输出验证, AI 自证不可靠, 证据边界大模型回答很自信，为什么还可能不可靠？

可靠 AI 智能体行动前需要什么证明/zh/reliable-ai-agent-proof-gate/AI 智能体可靠性, AI 行动边界, 无证明不行动, 自动化决策风险AI 智能体什么时候应该停下？