https://niraya666.github.io/en/ Recent content on LZY Blog https://niraya666.github.io/img/基于大语言模型的%20Agent：科普向-assets/cover.webp https://niraya666.github.io/img/基于大语言模型的%20Agent：科普向-assets/cover.webp Hugo -- gohugo.io en-us Sun, 13 Sep 2026 16:00:00 +0800 https://niraya666.github.io/en/posts/ai-decouples-production-and-cultivation/ Sun, 13 Sep 2026 16:00:00 +0800 https://niraya666.github.io/en/posts/ai-decouples-production-and-cultivation/ AI is decoupling output from cultivation: as entry-level tasks are absorbed, how talent debt forms, where judgment comes from, and why growth no longer just happens on its own. https://niraya666.github.io/en/posts/agent-eval-guide-03/ Wed, 05 Aug 2026 12:00:00 +0800 https://niraya666.github.io/en/posts/agent-eval-guide-03/ Part three of the Agent Eval Practical Guide series: a deep comparison of three scoring approaches — Code-based Evaluators, LLM-as-a-Judge, and Human Evaluation — their applicable scenarios, selection criteria, common pitfalls, and how evaluators themselves should be calibrated. https://niraya666.github.io/en/posts/agent-eval-guide-02/ Fri, 17 Jul 2026 12:00:00 +0800 https://niraya666.github.io/en/posts/agent-eval-guide-02/ Building an eval set is more than collecting queries — every case needs explicit success criteria, world state, trajectory anchors, and metadata to enable stable scoring and regression testing. https://niraya666.github.io/en/posts/agent-eval-guide-01/ Mon, 13 Jul 2026 12:00:00 +0800 https://niraya666.github.io/en/posts/agent-eval-guide-01/ The opening piece of the Agent Eval Practical Guide series: why LLM system evaluation is statistical rather than deterministic, why evaluation must precede optimization, and how human judgment is embedded into automated systems through the evaluation framework. https://niraya666.github.io/en/posts/agent-memory-benchmark-summary/ Thu, 09 Jul 2026 19:03:00 +0800 https://niraya666.github.io/en/posts/agent-memory-benchmark-summary/ A comprehensive survey of agent-memory benchmarks: from LoCoMo and LongMemEval to PersonaMem, covering evaluation methodology, metrics, and data synthesis pipelines across conversational memory, user profiling, and multimodal memory. https://niraya666.github.io/en/posts/ai-transformation-hardest-part-isnt-ai-but-dirty-work/ Wed, 10 Jun 2026 11:00:00 +0800 https://niraya666.github.io/en/posts/ai-transformation-hardest-part-isnt-ai-but-dirty-work/ Why AI transformation stalls after model selection: dirty data, undocumented workflows, integration debt, evaluation gaps, and organizational resistance are the real bottlenecks. https://niraya666.github.io/en/posts/from-agent-trajectory-to-agent-skill-compounding-returns/ Tue, 02 Jun 2026 20:00:00 +0800 https://niraya666.github.io/en/posts/from-agent-trajectory-to-agent-skill-compounding-returns/ How to convert agent trajectories into reusable skills through extraction, validation, deduplication, evaluation, and continuous optimization. https://niraya666.github.io/en/posts/agent-harness-minimalist-design-guide-2026/ Wed, 08 Apr 2026 21:03:00 +0800 https://niraya666.github.io/en/posts/agent-harness-minimalist-design-guide-2026/ A practical 2026 guide to the minimum architecture of an AI agent harness: loops, tools, filesystems, context management, memory, subagents, and observability. https://niraya666.github.io/en/posts/understanding-llm-thinking-budget-control/ Tue, 05 Aug 2025 20:12:00 +0800 https://niraya666.github.io/en/posts/understanding-llm-thinking-budget-control/ A practical analysis of LLM overthinking and thinking budget control solutions, comparing reasoning length and compute budget configurations across mainstream model APIs. https://niraya666.github.io/en/about/ Sun, 21 Apr 2024 12:00:00 +0800 https://niraya666.github.io/en/about/ <h3 id="about-me">About Me</h3> <ul> <li>Post-modern human</li> <li>AI practitioner and lifelong learner</li> <li>Work enthusiast</li> <li>World explorer</li> <li>INFJ</li> <li>Reading, listening, creating</li> </ul> <p>Welcome to my personal blog!</p> https://niraya666.github.io/en/travel-map/ Mon, 01 Jan 0001 00:00:00 +0000 https://niraya666.github.io/en/travel-map/ Places I have visited