# Xinwei Xiong (cubxxw) - AI, Open Source & Nomad Blog > Tech blog by Xinwei Xiong — AI Builder, open source contributor and digital nomad sharing Kubernetes, Go, AI projects and travel. Author: Xinwei Xiong (熊鑫伟, cubxxw) — AI Builder, open-source contributor, digital nomad. Primary domain: https://cubxxw.com/ Available languages: English (https://cubxxw.com/), Simplified Chinese (https://cubxxw.com/zh/). ## About - [About / Contact](https://cubxxw.com/about/): Personal profile and ways to reach me (WeChat, email, GitHub, socials). ## AI Agent ### [GEO Measurement & Tools: How to Know If AI Actually Cites You (with a DIY Monitor)](https://cubxxw.com/ai-agent/posts/geo-measurement-and-tools/) > GEO's core metrics aren't rank and clicks but citation rate and share of voice - because most value happens in AI answers the user never clicks through to. The answer first: swap your metrics, stop staring at rank and clicks In the GEO era, classic “rank + click” fails systematically — because most value happens in AI answers the user never clicks through to. Swap in two core metrics: citation rate (how often AI answers cite you) and share of voice (your mentions vs competitors). This chapter gives you a workable measurement system, from … ### [GEO Blog Rebuild Case Study: Running the Five-Layer Model on Real Data](https://cubxxw.com/ai-agent/posts/geo-blog-rebuild-case-study/) > Real baseline (old domain, last 3 months) - 852 clicks, 878K impressions, 0.1% CTR, average position 13.2. The technical base scores a perfect Lighthouse SEO; the bottleneck is ranking and clicks, not tech. The answer first: perfect tech, stuck at “impressions without being chosen” I dug through cubxxw.com’s real data, and the conclusion is blunt: technical SEO is near-perfect (Lighthouse SEO 100), yet the last 3 months turned 878K impressions into only 852 clicks — a 0.1% CTR at average position 13.2 (page two). The problem isn’t the technical layer; it’s … ### [GEO Trust & Endorsement: Why Reddit and Wikipedia Make Up Half of AI Citations](https://cubxxw.com/ai-agent/posts/geo-trust-and-endorsement/) > Retrieved and top-ranked still don't mean cited - citation hinges on trust, and most trust comes from how others talk about you, not how you talk about yourself. The answer first: the final gate is trust, not content You can max out your technical base and turn every paragraph into extractable LEGO — and AI may still not cite you. Because the final citation gate hinges on trust, and most trust comes from off-site: how others talk about you, where you’re mentioned on high-credibility platforms. This chapter turns that invisible trust into operable … ### [GEO Structured Tactics: Writing "Worth Citing" Into Every Paragraph (Answer-First, Schema, llms.txt)](https://cubxxw.com/ai-agent/posts/geo-structured-content-tactics/) > Answer-First paragraphs follow a formula - one standalone conclusion + 40-100 words of expansion + at least one piece of evidence (number/source/quote). That's the block AI lifts verbatim. The answer first: structure isn’t tidiness, it’s citability engineering Getting cited by AI isn’t about prose; it’s about extractability — whether each of your paragraphs can be lifted out and used as an answer on its own. This chapter hands you four copyable crafts: the Answer-First paragraph formula, a question-heading system, AI-facing Schema, and the right way to do … ### [GEO Mechanics: How AI Retrieves, Re-ranks, and Cites You](https://cubxxw.com/ai-agent/posts/geo-how-ai-retrieves-and-cites/) > AI search is RAG (retrieval-augmented generation). You must clear three gates — get retrieved, rank high in re-ranking, and get chosen as a citation at generation. Each gate rewards different signals. The answer first: you’re winning three gates, not one ranking AI search runs on RAG (Retrieval-Augmented Generation). A question comes in; the system first retrieves a set of candidate content, then re-ranks it by score, then has the LLM generate an answer and cite sources. GEO means clearing three gates: get retrieved (enter the candidate pool), score high in re-ranking (land near the top), … ### [GEO: The Complete Guide to Generative Engine Optimization (When Search Stops Giving Links and Starts Giving Answers)](https://cubxxw.com/ai-agent/posts/geo-generative-engine-optimization-guide/) > GEO doesn't replace SEO; it adds a layer. SEO fights for links and rankings; GEO fights for being cited when AI generates an answer. Both draw on the same content assets. The answer, first (this paragraph is GEO itself) GEO (Generative Engine Optimization) is the systematic optimization of your content, structure, and off-site reputation so that generative AI systems — ChatGPT, Perplexity, Google AI Overviews, Doubao, DeepSeek — can more easily understand, trust, and cite you when answering a user’s question. It doesn’t replace SEO; it layers on top of … ### [Dissecting open-lovable: An App Generator That Tames the Raw API Without an Agent Framework](https://cubxxw.com/ai-agent/posts/dissecting-open-lovable/) > open-lovable is Firecrawl's flagship open-source demo. It dresses up a dull middleware ("scraping") as a viral, tangible app — and those 27k stars convert naturally into Firecrawl API users. A textbook "open source as a growth funnel." Paste a URL, and in seconds an AI rebuilds it into a running, previewable, chat-editable modern React app. That’s the first impression of firecrawl/open-lovable — 27k stars, 5.2k forks, 94.9% TypeScript, a flagship open-source example built by the Firecrawl team. It targets the commercial product Lovable.dev (the README says outright “for a complete cloud solution, use … ### [Building a Production-Grade AI Agent System from Scratch: A Full Architecture Breakdown of Relay](https://cubxxw.com/ai-agent/posts/relay-agent-architecture-design/) > Multi-agent does not mean parallelism. The real reasons to split are different trigger modes, model-tier requirements, data flywheel independence, and divergent prompt evolution cadences. “Most Agent projects die in the unmapped wilderness between PoC and production.” I wrote that line while reading through the Relay project documentation. Relay is an open-source AI Agent system for job searching — not a demo built on three lines of LangChain plus GPT-4, but a project with complete architectural documentation, 172 engineering tasks, a hybrid tech stack, and explicit … ### [Context Is Not Prompt: Why Context Engineering Is Becoming AI's New Foundation](https://cubxxw.com/ai-agent/posts/context-engineering-the-new-foundation/) > Prompt engineering is writing one instruction well. Context engineering is deciding, on every inference call, what goes into the whole window, in what order, and what gets evicted. The center of gravity moved from wording to wiring. “We are not really writing prompts. We are furnishing a room for the model — deciding what gets carried in, where it sits, when it gets moved out. The wording is just a sticky note on the desk. What we are actually doing is the interior work.” If you had asked me in 2024 “how do I use AI well,” I would most likely have talked to you about prompts: how to phrase … ### [The Agent Engineering Map: Where Does That 98.4% of the Work Actually Live?](https://cubxxw.com/ai-agent/posts/agent-engineering-the-98-percent-harness/) > The agent loop is 10 lines of code; agent engineering is 100,000 lines. A paper reverse-engineering Claude Code v2.1.88 gives the narrative anchor: the core is a simple while loop, and almost all the code lives in the infrastructure around it — a permission system, a five-layer compaction pipeline, subagent isolation, session storage. “The agent loop is 10 lines of code. Agent engineering is 100,000 lines of code.” The first time I read that, I paused — and the more I sat with it, the sharper it cut. It punctures the single biggest illusion in this whole field: people think building an agent means writing a good prompt and wiring up an LLM API. But the actual work of pushing a demo to production — of running … ### [Agent Identity: From Locke to OpenClaw](https://cubxxw.com/ai-agent/posts/agent-identity-from-locke-to-openclaw/) > Locke's 1689 personal identity theory directly translates to engineering specifications: agents need persistent memory, self-referential capability, and continuity verification mechanisms to form trustworthy identity. Agent Identity: From Locke to OpenClaw On the engineering practice and philosophical framework of AI agent identity continuity 0. Introduction: An Engineering Problem Misdiagnosed as Philosophy The cost of agent amnesia is systematically underestimated. Not because users are annoyed—though they are. But because statelessness breaks the foundation of the trust account. Every session, the agent … ### [LangChain: Building LLM Applications](https://cubxxw.com/ai-agent/posts/harnessing-language-model-applications-with-langchain-a-developer-is-guide/) > LangChain provides a framework simplifying language model application development through integrated components like prompt templates, models, and output parsers for structured data handling. ### [Large Language Models: How LLMs Work](https://cubxxw.com/ai-agent/posts/exploring-large-language-models-llms-pioneering-ai-understanding-generation-human-language/) > Large Language Models exhibit emergent capabilities—contextual learning, instruction-following, and step-by-step reasoning—that scale dramatically beyond traditional smaller language models. LLM’s basic learning theory [toc] Introduction to large language models Large Language Model (LLM), also known as large language model, is an artificial intelligence model designed to understand and generate human language. LLMs typically refer to language models containing tens of billions (or more) of parameters that are trained on massive amounts of text data to gain a deep understanding of … ### [Sora Ease Guide: Mastering Sora AI for Developers](https://cubxxw.com/ai-agent/posts/sora-ease-guide-mastering-sora-ai-for-developers/) > Sora AI generates realistic and imaginative videos from text instructions by understanding and simulating physical world motion. Copyright: SoraEase , please indicate the source when forwarding. This article from the https://github.com/SoraEase/sora-prompt open source project,SoraEase wants to provide automated, easy, and quick to use tools and guidelines for SORa-related developers. Sora Prompt Collection, a repository dedicated to inspiring AI-driven video creation with Sora. Structured Prompt | Sora Collection of prompt … ### [Exploring Sora Technology for Enthusiasts and Developers](https://cubxxw.com/ai-agent/posts/exploring-sora-technology-for-enthusiasts-and-developers/) > Sora generates up to 60 seconds of high-quality video from text descriptions using innovative Diffusion Transformer architecture combining diffusion and Transformer models Sora! ! ! Recently, there has been a craze about Sora on the Internet. As the latest technology launched by OpenAI, Sora gives the magic of text-generated videos, and the results it demonstrates are impressive. At present, the appeal of short videos has far exceeded traditional novels and graphic comics. Therefore, the advent of Sora may trigger a revolution in the field of video production. The … ## Engineering ### [Kubernetes Resources and Learning Path Summary](https://cubxxw.com/engineering/posts/kubernetes-learning/) > Source code reading is essential for advancing in Kubernetes; start with Docker and Linux fundamentals before proceeding. Requirements 🔥 We need to further study and advance in kubernetes, reading source code is a necessary path. 👀 More importantly, it’s the collection of kubernetes resources. ⚠️ All resources use external links, book resources may not have links, others are personal experiences and summaries. Please contact for removal if there’s copyright infringement. 🚧 ⚠️ **Note: This article will be … ### [Troubleshooting Guide for OpenIM](https://cubxxw.com/engineering/posts/troubleshooting-guide-for-openim/) > Effective troubleshooting follows a systematic process: identify the problem through logs and testing, locate the root cause by tracing execution flow, and solve based on code understanding. Translation and Enhancement of the Article on Troubleshooting Techniques Using OpenIM If you’re seeking specific answers to issues related to OpenIM, I regret to inform you that this article isn’t a collection of problems and solutions. Instead, it focuses on the troubleshooting methods and debugging techniques gleaned from development and operational experiences using OpenIM as a case … ### [Navigating the Open Source Landscape](https://cubxxw.com/engineering/posts/navigating-the-open-source-landscape/) > Open source welcomes contributors at all skill levels; basic knowledge and persistence matter far more than being an expert. Open Source Has Greatly Contributed to My Growth Open source has significantly contributed to my growth, providing numerous experiences and learning opportunities. For those interested in my journey, here is a guide based on my first open source experience: Open Source Contribution Guidelines . When I first got involved in open source shortly after starting university, I discovered that many … ### [Go Directives & Automation Tools Deep Dive](https://cubxxw.com/engineering/posts/directives-and-the-use-of-automation-tools/) > Go provides go:generate directive to automatically create repetitive project-specific code during development before compilation, significantly reducing boilerplate and manual errors These go: instructions in the Go source code && go automation tools Developers have a strong tendency to automate repetitive tasks, and this also applies to writing code. Boilerplate code may include operations such as setting up a basic file structure, initializing variables, defining functions, or importing libraries or modules. In some cases, packages provide boilerplate code as a … ### [Concurrent Type Checking and Cross-Platform Development in Go](https://cubxxw.com/engineering/posts/concurrent-type-checking-and-cross-platform-development-in-go/) > Go's integer type varies by architecture—int is int32 on 32-bit systems and int64 on 64-bit systems, causing overflow bugs that require platform-specific type checking. OpenIM cross-platform source code type checking tool start question In the automation path of OpenIM, it involves more and more comprehensive automated design and testing. In the process, I encountered a problem, so I completed a full set of experience from go language type detection to integrated local and CI. The problem is this issue: https://github.com/openimsdk/open-im-server/issues/1807 Our … ### [OpenIM: Version Control & Testing Workflow](https://cubxxw.com/engineering/posts/openim-building-an-efficient-version-control-and-testing-workflow/) > OpenIM enforces strict branch management with main as stable, release branches for production versions, and PR testing branches requiring rigorous code review and automated testing before merge approval. The success of an open-source project largely depends on its quality management and collaborative processes. In the OpenIM open-source community, the standardization of project management and testing processes is crucial to ensure the quality and stability of the codebase. This document provides a brief overview of our testing strategy, branch management, quality control policies, and how they are … ### [GitOps & Kubernetes Deployment Strategies](https://cubxxw.com/engineering/posts/gitops-practice-theory-part/) > GitOps combines Git as single source of truth with Kubernetes Operators to automate deployment from version control to clusters Today we will take a look at the theory and practice of kubernetes and gitops Command introduction First, let’s take a look at the subnaming supported by kubectl for our convenience: kubectl apply - applies the definition of one or more resources. Typically used to deploy applications or update resources. kubectl get - displays information about one or more resources. This is a common command to … ### [Deployment and Design of Management Backend and Monitoring](https://cubxxw.com/engineering/posts/deployment-and-design-of-management-backend-and-monitoring/) > OpenIM supports multiple deployment strategies including Docker, source code, Kubernetes, each optimized for different environments and scalability needs. OpenIM offers various flexible deployment options to suit different environments and requirements. Here is a simplified and optimized description of these deployment options: Source Code Deployment: Regular Source Code Deployment: Deployment using the nohup method. This is a basic deployment method suitable for development and testing environments. For details, refer to the Regular Source Code … ### [Hugo Advanced Tutorial](https://cubxxw.com/engineering/posts/hugo-advanced-tutorial/) > Hugo templates leverage Go's html/template with automatic HTML escaping and pipeline syntax, where dot notation manages context through hierarchical data access. 136: Hugo Advanced Coming to the advanced part, you need to learn some advanced Hugo techniques in depth. Module Hugo modules are the core building blocks of Hugo. A module can be your main project or a smaller module that provides one or more of the 7* component types defined in Hugo: static, content, layouts, data, assets, i18n and archetypes. You can combine modules in any combination you like, … ### [Kubernetes for Kustomize Learning](https://cubxxw.com/engineering/posts/kubernetes-for-kustomize-learning/) > Kustomize enables declarative Kubernetes configuration customization through patches without modifying original YAML files, maintaining base application settings across environments. Introduction About Kustomize GitHub Repository Get Started Kustomize is an open-source configuration management tool designed specifically for Kubernetes. It helps users customize Kubernetes objects and manage them declaratively without modifying the original YAML files . This means you can retain the basic settings for applications and components while overriding default settings with declarative … ### [OpenIM Use Harbor Build Enterprise Mirror Repositories](https://cubxxw.com/engineering/posts/openim-use-harbor-build-enterprise-mirror-repositories/) > Harbor enables enterprises to build private container registries with Helm and Cert-manager for secure HTTPS-based image storage and CI/CD pipeline integration. Requirements OpenIM provides various public image registry addresses, such as aliyun, github, Docker Hub, and more. Read https://github.com/openimsdk/open-im-server/blob/main/docs/conversions/images.md for more image building guidelines. Most enterprises choose to set up their own image repository using Harbor, integrating it into their CI/CD pipeline to eventually replace Docker Hub and further … ### [Learn About Automated Testing](https://cubxxw.com/engineering/posts/learn-about-automated-testing/) > Automated testing ROI increases over time as maintenance costs remain fixed while tests run repeatedly across multiple environments and browsers. Automated testing practices and strategies for GitHub open source Go projects in the cloud native field introduce As OpenIM, a popular project on Github, how to create value in the cloud native era is very important. OpenIM is a small and high-quality team, and we do not have particularly in-depth insights in automation. Continuous Integration and Continuous Delivery (CI/CD) using GitHub Actions: … ### [Deep Dive into Kubernetes CNI, CRI, CSI Components](https://cubxxw.com/engineering/posts/deep-dive-into-the-components-of-kubernetes-cni-csi-cri/) > CRI is a gRPC-based interface that allows Kubernetes to support multiple container runtimes, with containerd being the optimal choice over Docker and CRI-O. Container Runtime Container Runtime runs on every node in a Kubernetes (k8s) cluster and is responsible for the entire container lifecycle. Docker is currently the most widely used. With the development of container cloud, more and more container runtimes have emerged. To solve the integration problems between these container runtimes and Kubernetes, the community introduced CRI (Container Runtime … ### [Kubernetes Control Plane - Detailed Analysis of Kubelet](https://cubxxw.com/engineering/posts/deep-dive-into-the-components-of-kubernetes-kubectl/) > Kubelet runs on each Kubernetes node managing Pod lifecycles, handling container operations including image pulling, startup, monitoring, and status reporting to the API Server Kubelet Component Analysis Understanding kubelet The Kubelet component runs on Node nodes, maintaining running Pods and providing Kubernetes runtime environment, mainly accomplishing the following missions: Monitor Pods assigned to this Node Mount volumes required by Pods Download Pod secrets Run containers in Pods through docker/rkt Periodically execute liveness probes defined for containers in … ### [Kubernetes Control Plane - Scheduler](https://cubxxw.com/engineering/posts/deep-dive-into-the-components-of-kubernetes-scheduler/) > Kubernetes scheduler assigns pods to nodes using two phases: predicate filtering unsuitable nodes and priority ranking candidates by policies Scheduler kube-scheduler is responsible for scheduling and assigning Pods to nodes within the cluster. It listens to kube-apiserver, queries for Pods that haven’t been assigned to Nodes, and then assigns nodes to these Pods based on scheduling policies (updating the Pod’s NodeName field). The scheduler needs to fully consider many factors: Fair scheduling; Efficient resource … ## Growth & Life ### [2026 June Thought Notes: The Pushing-Away Comes Before the Reason for Pushing Away](https://cubxxw.com/growth/posts/2026-06-thought-notes/) > I finally saw the mechanism of how I push people away — the pushing-away comes first, the reason comes after. The amygdala is fighting a war that ended when I was three. Prologue: A slack surface, and high density underneath In June I came back from Laos to Shenzhen. If you only look at the state, this was a slack month. What it looked like in practice — I would only come alive after dark, stay up to two or three a.m. every night, lulled to sleep by one short video clip after another. Waking up in the day to Shenzhen’s gray-white sky, I could not even gather … ### [Ignite and Settle (Part 3): Anxious Attachment — Why We Keep Seeking Reassurance in Love](https://cubxxw.com/growth/posts/2026-06-28-ignite-and-settle-3-anxious/) > The core of anxious attachment is unstable love — love was there, but unpredictable; response was there, but hot then cold. This is Part 3 of “Ignite and Settle,” and the close of the series. Part 1, The Quality and Time of Companionship , asked why we turn away just as we come close. Part 2, Avoidant Attachment , described how the “turning away” person is made. This part describes the other half — the one that looks opposite, and bites tight to the previous one at the bottom: the anxious. 1. … ### [Ignite and Settle (Part 2): Avoidant Attachment — Why We Want to Run When Someone Gets Close](https://cubxxw.com/growth/posts/2026-06-28-ignite-and-settle-2-avoidant/) > Avoidant is not coldness. It is an early-learned lesson — expressing needs doesn't work, getting close isn't safe, carrying it alone is steadiest. This is Part 2 of “Ignite and Settle.” Part 1, The Quality and Time of Companionship , described the act of “running when someone gets close.” This part focuses on its origins and repair. Part 3 will describe anxious attachment — its mirror, with which it is biting at the bottom. 1. First, get the terminology right In the Chinese context, what people often call … ### [Ignite and Settle (Part 1): The Quality and Time of Companionship — and Why We Turn Away Just as We Come Close](https://cubxxw.com/growth/posts/2026-06-28-ignite-and-settle-1-quality-and-time/) > Quality and time are not rivals. They are two semi-independent systems — responsiveness eats quality, habituation eats time. This is Part 1 of the “Ignite and Settle” series. Part 2 is on avoidant attachment — why we want to run when someone gets close. Part 3 is on anxious attachment — why we keep seeking reassurance in love. You can enter from any of the three. Read together, they form one map. Opening: A deceptively simple question Let me start with a deceptively simple question: in companionship, which … ### [The Super-Individual Stack: AI-Native Product Directions and Solo Builder Ops in 2026](https://cubxxw.com/growth/posts/super-individual-ai-product-and-solo-builder-stack/) > The super-individual's real moat is not the model—it's the harness, the schema, and distribution. The Claude Code codebase is 98.4% scaffolding and only 1.6% AI decision logic. “Software is eating the world.” — Marc Andreessen, 2011 “Now AI is eating software—and the question for the rest of us is: what’s left for one human, alone, in front of a screen?” — me, asking myself one night in 2026. Prologue: How Big Does One Person Need to Be? In February 2026, I ran my first complete overnight agent. I set a prompt, dropped it into Claude Code … ### [Seen Clearly, Loved Deeply: Five Lenses on Love, and the Buddhist Synthesis](https://cubxxw.com/growth/posts/love-seen-clearly-loved-deeply/) > Each of the five secular disciplines illuminates one facet of love yet stays blind to the rest; Buddhism's dependent origination is the meta-framework that contains them all. “From love springs grief, from love springs fear; for one freed from love there is no grief—whence, then, fear?” —Dhammapada “Just as a mother would protect her only child with her life, so let one cultivate a boundless love toward all beings.” —Metta Sutta In the same scriptural tradition, love is both the source of grief and fear, and a boundless, infinite kindness. … ### [2026 May Thought Notes: On the Edge, I Recognize Myself Through Action](https://cubxxw.com/growth/posts/2026-05-thought-notes/) > May was an experiment in intentional disconnection — quitting flomo, leaving the Shenzhen script, using action rather than meditation to know myself. Prologue: A May Without flomo For the entire month of May, I did not write a single flomo note. I barely posted to my feed either. This was not laziness. It was a conscious decision. I had already decided on April 30 that I would stop. The reason is simple: I no longer felt I needed it. Or more accurately — could it have a better form? I want to use my own product. It is not perfect yet, but I … ### [April 2026 Thought Notes](https://cubxxw.com/growth/posts/2026-04-thought-notes/) > AI era demands Harness Design and environmental/protocol thinking over traditional product design as models grow stronger than infrastructure. 2026年4月思考笔记 AI 知识时效性：从 RAG 到元认知的解决思路 2026 年 04 月 30 日 - 14:46:26 有些工具现在用的是时效性的方法 比如说是通过一些 API 或者 RAG 系统挂在最新的文档 或者是 web search 最新的数据源 元认知层，让模型知道自己不知道，这是更难的问题 显式标注知识截止日期 训练模型在时效敏感问题上主动表达不确定性 系统提示里注入当前日期，让模型对比自身知识边界 时间戳溯源，每一条知识附带获取时间，超时自动触发重新验证 越旧的信息置信度越低，触发 fallback 到搜索 对高频变动领域（金融、法律、医疗）设置强制校验节点 一个模型生成答案，另一个专门质疑时效性 #格物/ai 利他性记录：有意义才有持续驱动 2026 年 04 月 30 日 - 14:07:21 利他性的前提，是这件事情对他很有意义 这个平台很 有记录的价值 可以 … ### [Maintaining Self-Worth in the Age of AI](https://cubxxw.com/growth/posts/weight-of-self-in-ai-age/) > Measuring self-worth by utility and being needed is fragile and crumbles when AI can replicate those functions better than humans. Pascal wrote in the 17th century: “All of humanity’s problems stem from one thing: man’s inability to sit quietly in a room alone.” Three hundred years later, I thought of this sentence late at night in Lhasa, and added one more: They’re not sure who exactly is the one sitting in that room. Introduction: The 3 AM Emptiness For a while, I woke up almost every day at 3 … ### [March 2026 Thought Notes](https://cubxxw.com/growth/posts/2026-03-thought-notes/) > In healthy relationships, replace sacrificial one-way giving with co-creation: shared experiences and mutual growth build solid connections stronger than performative gestures. March 2026 Thought Notes 468 entries this month | Recorded: March 1 — March 31, 2026 Monthly Theme: Angkor Wat · AI Agent · Self-Narrative Core Topics: Gewu/AI (~53), Gewu/Cambodia (~38), Gewu/Angkor Wat (~23), Gewu/Hinduism (~15), Guanwo (~14), Gewu/Vietnam (~11) Daily Notes Archive March 1, 2026 Sunday (29 entries) Northeast China’s Direct Culture: How Labor Independence Shapes … ### [Lhasa: Slow and Heavy](https://cubxxw.com/growth/posts/2026-03-27-lhasa-slow-and-heavy/) > The English version of this article is not yet published; full translation is in progress This article is available in Chinese. An English translation is in progress. From February to March 2026, I stayed in Lhasa, Tibet — a city that forces you to slow down before you can begin to understand it. The altitude arrives first, then the silence, and finally the weight of everything the city carries. This piece explores what it means to be a stranger in a place where time moves differently, … ### [Wandering & Growing: 2025-2026 Annual Review](https://cubxxw.com/growth/posts/2025-annual-review/) > Travel initially provided transformative energy but eventually became an avoidance mechanism when internal processing grew harder to sustain through constant movement. On the edge of a cliff, all logic fails. Only intuition can connect with the world. I wrote this sentence one night in Lhasa. Finding it now, I feel it can serve as the entry point to this review. Over these past fourteen months, I’ve been living on the edge of cliffs — almost literally. Encountering ice slopes on ACT mountain trails with drops beside me; every step at the altitude of Genie … ### [AI and Self-Identity: Who Am I in the AI Age](https://cubxxw.com/growth/posts/ai-and-self-identity/) > Human agency rooted in embodied presence and felt experience, not thinking or logic, cannot be replicated by AI. Preface: This article was written on a night in Lhasa. At an altitude of over 3,000 meters, the air is thin and the stars are very close. One evening I stood in the courtyard, alone, and watched the moon in the sky for a long time. That moon doesn’t lose a bit of its light because of AI’s existence. The moonlight is still truly beautiful — What do humans still possess in the AI era? ### [February 2026 Thought Notes](https://cubxxw.com/growth/posts/2026-02-thought-notes/) > Wildlife protection requires controlling human desire rather than expanding resources through internal education or external systemic incentives. February 2026 Thought Notes | 2026 年 2 月思考笔记 Total: 365 notes this month | February 1-28, 2026 Monthly Themes: Travel, Ecology, Urban Observation Core Topics: Travel Experiences (~80 notes), AI Products (~45 notes), Ecological Protection (~25 notes), Urban Governance (~20 notes), Buddhism & Philosophy (~30 notes), Self-Discovery (~40 notes) Selected Notes of the Month Desire & Resources: … ### [January 2026 Thought Notes](https://cubxxw.com/growth/posts/2026-01-thought-notes/) > AI's true profitability comes from deterministic, ROI-measurable work delivered to businesses, not flashier consumer content generation. January 2026 Thought Notes | 2026 年 1 月思考笔记 Total: 388 notes this month | January 1-31, 2026 Monthly Themes: AI Deterministic Value, Finance Patterns, Self-Exploration Core Topics: AI Products & Agent Design (55 notes), Finance & Investment (19 notes), Self-Discovery (11 notes), Biology (10 notes), Social Observation (7 notes), Buddhism (6 notes) Selected Notes of the Month A-Share … ## Open Source Projects ### [Mem0 Explained: AI Memory Layer Architecture & How It Works](https://cubxxw.com/projects/mem0/) > Mem0 is a memory layer for AI agents: it extracts durable facts and preferences from interaction history, stores them, and retrieves the right context when the agent needs it. This project is an ongoing journey: learning AI open source projects through hands-on practice, building real systems with AI tools, and documenting the process. Notion List 1. What Mem0 Is Mem0, pronounced “mem-zero”, is an open source memory layer for AI assistants and agents. Its job is to turn interaction history into reusable memory: user preferences, stable facts, entity … ### [AI Recommendation Systems: How They Work](https://cubxxw.com/projects/ai-recommend/) > Article explores AI-powered recommendation systems and personalized content discovery engine architecture This project is an ongoing process — learning AI open source projects one day at a time, building real-world skills by combining hands-on practice with AI tooling, and documenting the journey. notion List Basic Information: Project Name: GitHub URL: Primary Tech Stack: Related Articles Stages of Growth in Open Source A Complete Guide to Open Source Contribution (For First-Timers) My Practice … ### [NotebookLM: Google's AI Research Tool](https://cubxxw.com/projects/notebooklm/) > Article body is incomplete; only stub headers and related links present with no substantive NotebookLM content This project is an ongoing journey — learning AI open-source projects one step at a time, building real skills through hands-on practice, and using AI tools to sharpen the ability to tackle complex problems. Everything gets documented along the way. Notion List Basic Information: Project Name: GitHub URL: Primary Tech Stack: Related Articles A Stage-by-Stage Growth Guide for Open Source A … ### [TDD for AI: Test-Driven Development Guide](https://cubxxw.com/projects/tdd/) > Applying Test-Driven Development principles to AI and LLM application development practices This project is an ongoing journey — learning AI open-source projects one step at a time, building real skills through hands-on practice, and using AI tools to sharpen the ability to tackle complex problems. Everything gets documented along the way. Notion List Basic Information: Project Name: GitHub URL: Primary Tech Stack: Related Articles A Stage-by-Stage Growth Guide for Open Source A … ### [Microsoft MarkItDown Tutorial: Convert PDF to Markdown](https://cubxxw.com/projects/markitdown/) > MarkItDown converts diverse document formats into structured Markdown, specifically designed for LLM preprocessing with emphasis on content preservation over visual fidelity. This project is an ongoing journey — learning AI open source projects with steady, daily progress. Through hands-on work with real projects and AI tooling, the goal is to develop the ability to solve complex problems and document the process. Notion List 1. Introduction 1.1. MarkItDown and Markdown — Clarifying the Relationship First, it is important to clarify that “MarkItDown” is … ### [LangGraph Architecture: StateGraph, Nodes and Edges](https://cubxxw.com/projects/langgraph/) > LangGraph is a low-level orchestration framework for long-running, stateful agents, where state, nodes, and edges form an explicit execution graph. This project is an ongoing journey — learning AI open source projects with steady, daily progress. Through hands-on work with real projects and AI tooling, the goal is to develop the ability to solve complex problems and document the process. Notion List Basic Information: Project Name: LangGraph GitHub URL: https://github.com/langchain-ai/langgraph Main Tech Stack: Python, JavaScript/TypeScript, … ### [LangChain: Open Source LLM Framework](https://cubxxw.com/projects/langchain/) > LangChain evolved from a monolithic LLM framework into a modular ecosystem with specialized tools—LCEL for composition, LangGraph for complex agents, LangSmith for observability, and platform infrastructure. This project is an ongoing effort to study AI open source projects one step at a time, building real-world skills by combining hands-on practice with AI tooling to tackle complex problems. Everything is documented here. Notion List I. Executive Summary LangChain has emerged as one of the leading frameworks for building applications powered by large language models (LLMs). This report provides an … ### [LLM/AI API Gateway Market Analysis & Startup Stack Recommendations](https://cubxxw.com/projects/ai-gateway/) > AI gateways centralize LLM access, enabling token-based billing, multi-provider routing, cost control, security guardrails, and observability critical for production AI applications. This project is an ongoing process — learning AI open source projects one day at a time, building real-world skills by combining hands-on practice with AI tooling, and documenting the journey. notion List LLM/AI API Gateway Market Analysis & Startup Stack Recommendations 1. Executive Summary As large language model (LLM) and AI applications proliferate, managing their interactions with … ### [Independent Developer in the AI Era: Open Source Deep Dive](https://cubxxw.com/projects/independent-developer/) > Independent developers navigate the AI era through systematic hands-on engagement with open source projects and real-world applications. This project is an ongoing effort — learning AI open source projects one step at a time, building real-world practice through hands-on projects, combining AI tools to strengthen the ability to solve complex problems. Everything is documented along the way. Notion List Basic Information: Project Name: GitHub URL: Primary Tech Stack: Related Articles A Staged Growth Guide for Open Source A Complete … ### [GPT Researcher Explained: Open Source Deep Research Agent](https://cubxxw.com/projects/gpt-researcher/) > GPT Researcher is an open source deep research agent that plans a question, gathers web and local sources, validates evidence, and generates cited reports. This project is an ongoing effort — learning AI open source projects one step at a time, building real-world practice through hands-on projects, combining AI tools to strengthen the ability to solve complex problems. Everything is documented along the way. Notion List Project Overview Basic Information Project Name: GPT Researcher GitHub URL: https://github.com/assafelovic/gpt-researcher Primary … ### [Jina AI: Multimodal Search & Embeddings](https://cubxxw.com/projects/jina/) > No substantive article body exists; file contains only template structure and placeholder sections. Project Overview Basic Information Project Name: GitHub URL: Primary Tech Stack: License: Star Count: Last Updated: Project Introduction [Describe the project’s main features and use cases here] Technical Architecture System Design [Describe the overall architectural design of the project here] Core Modules [List and explain the project’s core modules here] Source Code Analysis Key … ## Feeds - Sitemap: https://cubxxw.com/sitemap.xml - RSS: https://cubxxw.com/index.xml - JSON Feed: https://cubxxw.com/feed.json _Rendered from / on https://cubxxw.com/ (language: en)._