<?xml version="1.0" encoding="UTF-8"?>
<rss
    version="2.0"
    xmlns:atom="http://www.w3.org/2005/Atom"
>
    <channel>
        <atom:link
            type="application/rss+xml"
            href="https://ai.tekin.cn/index.php/en/feed/posts"
            rel="self"
        />
        <title><![CDATA[Posts feed]]></title>
        <link><![CDATA[https://ai.tekin.cn/index.php/en/feed/posts]]></link>
                <description><![CDATA[Latest posts from Industry-Specific Digital Solutions for Manufacturing, Finance, Healthcare + Resources for Clients &amp; Developers]]></description>
        <language>en_US</language>
        <pubDate>2026-04-19T10:27:57+08:00</pubDate>

                    <item>
                <title><![CDATA[envsubst: Production-Grade Secure Environment Variable Substitution Tool Built for Nginx/Docker/K8s]]></title>
                <link>https://ai.tekin.cn/index.php/en/blog/envsubst-secure-environment-variable-substitution-nginx-docker-k8s</link>
                <description><![CDATA[This article introduces an enhanced version of the envsubst tool that solves security issues of the native tool in Nginx configuration substitution. It supports advanced features like wildcard whitelists, default value syntax, and debug mode, with zero dependencies—perfect for containerized deployment scenarios.]]></description>
                                    <author><![CDATA[Tekin Tian <tekintian@gmail.com>]]></author>
                                <guid>https://ai.tekin.cn/index.php/en/blog/envsubst-secure-environment-variable-substitution-nginx-docker-k8s</guid>
                <pubDate>Sun, 19 Apr 2026 10:27:57 +0800</pubDate>
                <enclosure
                    type="image/webp"
                    url="http://ai.tekin.cn/storage/blog/202604/envsubst-secure-environment-variable-substitution-nginx-docker-k8s.webp"
                    length="103236"
                />
                                    <category>核心技术干货（企业级实战）</category>
                            </item>
                    <item>
                <title><![CDATA[Dockerfile Multiline File Writing: Full Compatibility Solutions for Legacy Docker Versions]]></title>
                <link>https://ai.tekin.cn/index.php/en/blog/dockerfile-multiline-write-file-methods</link>
                <description><![CDATA[This article comprehensively summarizes 4 practical methods to write multi-line configurations and scripts to files in a Dockerfile, with a focus on the incompatibility of the syntax=docker/dockerfile:1 heredoc syntax in legacy Docker (v0) versions. Based on locally tested and validated approaches including RUN { ... }|tee, echo $'...', printf, and backslash line continuation ...]]></description>
                                    <author><![CDATA[Tekin Tian <tekintian@gmail.com>]]></author>
                                <guid>https://ai.tekin.cn/index.php/en/blog/dockerfile-multiline-write-file-methods</guid>
                <pubDate>Sun, 05 Apr 2026 10:51:31 +0800</pubDate>
                <enclosure
                    type="image/png"
                    url="http://ai.tekin.cn/index.php/assets/core/base/images/placeholder.png"
                    length="0"
                />
                                    <category>核心技术干货（企业级实战）</category>
                            </item>
                    <item>
                <title><![CDATA[In-depth Analysis of PyInstaller macOS Single-File Mode Dual-Process Issue]]></title>
                <link>https://ai.tekin.cn/index.php/en/blog/pyinstaller-macos-singlefile-dual-process-issue-solution</link>
                <description><![CDATA[This article analyzes the dual-process, crash, and process residue issues caused by PyInstaller's single-file mode when packaging PyQt5 M3U8 downloader on macOS. Through step-by-step troubleshooting—fixing obvious bugs, eliminating system interference, and comparative testing—it locates the compatibility bug of single-file mode. It provides reusable directory mode spec configuration...]]></description>
                                    <author><![CDATA[Tekin Tian <tekintian@gmail.com>]]></author>
                                <guid>https://ai.tekin.cn/index.php/en/blog/pyinstaller-macos-singlefile-dual-process-issue-solution</guid>
                <pubDate>Fri, 20 Mar 2026 09:36:09 +0800</pubDate>
                <enclosure
                    type="image/webp"
                    url="http://ai.tekin.cn/storage/blog/202603/app-2-process-issue.webp"
                    length="45426"
                />
                                    <category>安全与运维专题</category>
                            </item>
                    <item>
                <title><![CDATA[AI Model Naming Conventions: A Complete Guide to Tags, Standards and Industry Best Practices]]></title>
                <link>https://ai.tekin.cn/index.php/en/blog/ai-model-naming-conventions-tags-standards</link>
                <description><![CDATA[This guide unlocks AI model naming logic, breaking down the core-to-detail universal structure and defining all mainstream tags (capability, training, quantization, language). It compares naming standards across open-source platforms, quantized model types and top vendors, and offers practical interpretation of typical model names—empowering seamless model selection and local deployment ...]]></description>
                                    <author><![CDATA[Tekin Tian <tekintian@gmail.com>]]></author>
                                <guid>https://ai.tekin.cn/index.php/en/blog/ai-model-naming-conventions-tags-standards</guid>
                <pubDate>Tue, 17 Mar 2026 10:27:29 +0800</pubDate>
                <enclosure
                    type="image/webp"
                    url="http://ai.tekin.cn/storage/blog/202603/ai-model-naming-conventions-tags-standards.webp"
                    length="79058"
                />
                                    <category>开源与 AI 应用专题</category>
                            </item>
                    <item>
                <title><![CDATA[🌟 Series Overview: The Misalignment of Intelligence—The Underlying Logic of AI Hallucination]]></title>
                <link>https://ai.tekin.cn/index.php/en/blog/ai-hallucination-underlying-logic-series-guide</link>
                <description><![CDATA[This 10-part series explores the fundamental mechanisms behind AI hallucination—why language models confidently generate false information and why this cannot be fixed. From understanding how AI predicts rather than thinks, to examining real-world cases of fabricated court cases and fake academic citations, each article builds toward a comprehensive framework for understanding AI's capabilities and limitations.]]></description>
                                    <author><![CDATA[Tekin Tian <tekintian@gmail.com>]]></author>
                                <guid>https://ai.tekin.cn/index.php/en/blog/ai-hallucination-underlying-logic-series-guide</guid>
                <pubDate>Mon, 16 Mar 2026 21:29:00 +0800</pubDate>
                <enclosure
                    type="image/webp"
                    url="http://ai.tekin.cn/storage/blog/ai-hallucination/ai-hallucination.webp"
                    length="131194"
                />
                                    <category>核心技术干货（企业级实战）</category>
                            </item>
                    <item>
                <title><![CDATA[The Underlying Logic of Language Models: Why AI Hallucination Is Irremovable]]></title>
                <link>https://ai.tekin.cn/index.php/en/blog/ai-hallucination-irremovable-language-model-limitations</link>
                <description><![CDATA[AI hallucination cannot be eliminated through better technology—it's mathematically inevitable. Nature published research stating that hallucination is a "designed feature of LLMs, not a bug." OpenAI acknowledges this is "mathematically unavoidable." Turing Award winner Yann LeCun argues that genuine reasoning requires "world models" that learn from reality, not just text. Understanding AI's limitations is as important as understanding its capabilities.]]></description>
                                    <author><![CDATA[Tekin Tian <tekintian@gmail.com>]]></author>
                                <guid>https://ai.tekin.cn/index.php/en/blog/ai-hallucination-irremovable-language-model-limitations</guid>
                <pubDate>Mon, 16 Mar 2026 20:39:36 +0800</pubDate>
                <enclosure
                    type="image/webp"
                    url="http://ai.tekin.cn/storage/blog/ai-hallucination/10.webp"
                    length="78992"
                />
                                    <category>核心技术干货（企业级实战）</category>
                            </item>
                    <item>
                <title><![CDATA[AI Isn't Smart—It's Just Too Good at Sounding Human]]></title>
                <link>https://ai.tekin.cn/index.php/en/blog/ai-language-intelligence-vs-cognitive-intelligence</link>
                <description><![CDATA[AI demonstrates impressive "language intelligence"—it speaks fluently, structures arguments, and mimics expert tone. But language intelligence is not cognitive intelligence. Philosopher John Searle's famous "Chinese Room" argument reveals that AI manipulates symbols without understanding their meaning. Studies show AI can be more persuasive than humans, not because it's smarter but because...]]></description>
                                    <author><![CDATA[Tekin Tian <tekintian@gmail.com>]]></author>
                                <guid>https://ai.tekin.cn/index.php/en/blog/ai-language-intelligence-vs-cognitive-intelligence</guid>
                <pubDate>Mon, 16 Mar 2026 20:31:01 +0800</pubDate>
                <enclosure
                    type="image/webp"
                    url="http://ai.tekin.cn/storage/blog/ai-hallucination/9.webp"
                    length="98564"
                />
                                    <category>核心技术干货（企业级实战）</category>
                            </item>
                    <item>
                <title><![CDATA[Human Logic vs. AI Logic: The Fundamental Difference Between Two Types of Intelligence]]></title>
                <link>https://ai.tekin.cn/index.php/en/blog/human-logic-vs-ai-logic-intelligence-misalignment</link>
                <description><![CDATA[Human and AI intelligence operate on fundamentally different logical systems. Human logic is meaning-driven: we understand what words represent, reason from evidence, and verify against reality. AI logic is probability-driven: it predicts likely next words based on statistical patterns without comprehension. Cognitive scientist Gary Marcus describes language models as "blind to truth"—they cannot distinguish true statements from false ones.]]></description>
                                    <author><![CDATA[Tekin Tian <tekintian@gmail.com>]]></author>
                                <guid>https://ai.tekin.cn/index.php/en/blog/human-logic-vs-ai-logic-intelligence-misalignment</guid>
                <pubDate>Mon, 16 Mar 2026 20:24:28 +0800</pubDate>
                <enclosure
                    type="image/webp"
                    url="http://ai.tekin.cn/storage/blog/ai-hallucination/8.webp"
                    length="98506"
                />
                                    <category>核心技术干货（企业级实战）</category>
                            </item>
                    <item>
                <title><![CDATA[Why Can AI Describe Non-Existent Books Convincingly?]]></title>
                <link>https://ai.tekin.cn/index.php/en/blog/ai-fabricates-fake-books-academic-hallucination</link>
                <description><![CDATA[Ask AI about a book that doesn't exist, and it will produce a detailed review complete with author, publisher, chapters, and key arguments. A 2023 Nature study confirmed that AI fabricates academic references that look perfectly formatted but don't exist. The implications are serious: AI-generated fake papers are already appearing in academic ecosystems, with fabricated citations potentially becoming accepted as real sources.]]></description>
                                    <author><![CDATA[Tekin Tian <tekintian@gmail.com>]]></author>
                                <guid>https://ai.tekin.cn/index.php/en/blog/ai-fabricates-fake-books-academic-hallucination</guid>
                <pubDate>Mon, 16 Mar 2026 19:24:04 +0800</pubDate>
                <enclosure
                    type="image/webp"
                    url="http://ai.tekin.cn/storage/blog/ai-hallucination/7.webp"
                    length="88436"
                />
                                    <category>核心技术干货（企业级实战）</category>
                            </item>
                    <item>
                <title><![CDATA[You Ask AI About a Non-Existent Person, It Can Fabricate an Entire Life for Them]]></title>
                <link>https://ai.tekin.cn/index.php/en/blog/ai-fabricates-fake-people-person-hallucination</link>
                <description><![CDATA[Give AI any name, and it will generate a complete biography—birth year, education, career, achievements, and even personality traits. This isn't magic; it's structure completion. Real cases illustrate the danger: a Norwegian man searched his own name and discovered AI had fabricated a story that he murdered his two children. AI doesn't verify whether people exist—it applies learned patterns of "what a person's life should look like."]]></description>
                                    <author><![CDATA[Tekin Tian <tekintian@gmail.com>]]></author>
                                <guid>https://ai.tekin.cn/index.php/en/blog/ai-fabricates-fake-people-person-hallucination</guid>
                <pubDate>Mon, 16 Mar 2026 19:10:17 +0800</pubDate>
                <enclosure
                    type="image/webp"
                    url="http://ai.tekin.cn/storage/blog/ai-hallucination/6.webp"
                    length="105400"
                />
                                    <category>核心技术干货（企业级实战）</category>
                            </item>
                    <item>
                <title><![CDATA[AI Hallucination Isn't a Bug—It's Its Nature: When Language Patterns Meet Human Logic]]></title>
                <link>https://ai.tekin.cn/index.php/en/blog/ai-hallucination-feature-not-bug-language-models</link>
                <description><![CDATA[Many believe AI hallucination will disappear with technological advancement. The uncomfortable truth: hallucination is built into language models' very architecture. OpenAI research confirms that generating "overconfident, plausible false information" is an inherent characteristic, not a fixable flaw. Turing Award winner Yann LeCun states that text-trained models cannot understand the physical world—even a house cat has better world understanding.]]></description>
                                    <author><![CDATA[Tekin Tian <tekintian@gmail.com>]]></author>
                                <guid>https://ai.tekin.cn/index.php/en/blog/ai-hallucination-feature-not-bug-language-models</guid>
                <pubDate>Mon, 16 Mar 2026 18:55:18 +0800</pubDate>
                <enclosure
                    type="image/webp"
                    url="http://ai.tekin.cn/storage/blog/ai-hallucination/5.webp"
                    length="82420"
                />
                                    <category>核心技术干货（企业级实战）</category>
                            </item>
                    <item>
                <title><![CDATA[Why Does AI Get More Absurd the More You Ask? The Underlying Logic Explains Everything]]></title>
                <link>https://ai.tekin.cn/index.php/en/blog/ai-hallucination-worse-with-follow-up-questions</link>
                <description><![CDATA[The more you probe AI with follow-up questions, the more its responses drift into absurdity. This isn't AI malfunctioning—it's the probability chain working exactly as designed. When users embed false premises in their questions, AI unconditionally accepts them and continues building on those errors. Research confirms that longer conversations correlate with higher hallucination rates, as AI attempts to "please" users regardless of factual accuracy.]]></description>
                                    <author><![CDATA[Tekin Tian <tekintian@gmail.com>]]></author>
                                <guid>https://ai.tekin.cn/index.php/en/blog/ai-hallucination-worse-with-follow-up-questions</guid>
                <pubDate>Mon, 16 Mar 2026 18:25:26 +0800</pubDate>
                <enclosure
                    type="image/webp"
                    url="http://ai.tekin.cn/storage/blog/ai-hallucination/4.webp"
                    length="79380"
                />
                                    <category>核心技术干货（企业级实战）</category>
                            </item>
                    <item>
                <title><![CDATA[AI's Confidence Doesn't Come from Knowing—It Comes from Learning to "Sound Expert"]]></title>
                <link>https://ai.tekin.cn/index.php/en/blog/ai-confidence-tone-vs-knowledge-language-patterns</link>
                <description><![CDATA[AI's unwavering confidence often tricks users into believing it possesses deep knowledge. In reality, this confidence comes from learned language patterns, not actual understanding. During training, AI absorbs millions of texts where experts speak with certainty—and it mimics this tone regardless of factual accuracy. OpenAI has acknowledged that models will "confidently give wrong answers rather than admit ignorance" as an inherent architectural feature.]]></description>
                                    <author><![CDATA[Tekin Tian <tekintian@gmail.com>]]></author>
                                <guid>https://ai.tekin.cn/index.php/en/blog/ai-confidence-tone-vs-knowledge-language-patterns</guid>
                <pubDate>Mon, 16 Mar 2026 18:13:54 +0800</pubDate>
                <enclosure
                    type="image/webp"
                    url="http://ai.tekin.cn/storage/blog/ai-hallucination/3.webp"
                    length="39182"
                />
                                    <category>核心技术干货（企业级实战）</category>
                            </item>
                    <item>
                <title><![CDATA[You Think AI Is Thinking? It's Actually Just Speaking in Probabilities]]></title>
                <link>https://ai.tekin.cn/index.php/en/blog/ai-thinking-probability-language-models-explained</link>
                <description><![CDATA[AI appears to think, reason, and understand—but it's actually performing sophisticated probability calculations. Language models operate on a single principle: predict the most likely next word based on context. This article breaks down the token prediction mechanism that powers systems like ChatGPT, explaining why AI can answer complex questions without truly understanding them. The key insight: AI excels at pattern recognition, not comprehension.]]></description>
                                    <author><![CDATA[Tekin Tian <tekintian@gmail.com>]]></author>
                                <guid>https://ai.tekin.cn/index.php/en/blog/ai-thinking-probability-language-models-explained</guid>
                <pubDate>Mon, 16 Mar 2026 18:06:50 +0800</pubDate>
                <enclosure
                    type="image/webp"
                    url="http://ai.tekin.cn/storage/blog/ai-hallucination/2.webp"
                    length="87324"
                />
                                    <category>核心技术干货（企业级实战）</category>
                            </item>
                    <item>
                <title><![CDATA[Why Does AI Confidently Make Things Up?]]></title>
                <link>https://ai.tekin.cn/index.php/en/blog/ai-hallucination-why-ai-makes-things-up</link>
                <description><![CDATA[AI hallucination occurs when language models generate plausible but false information with unwavering confidence. This isn't a bug—it's an inevitable result of how AI fundamentally operates. Unlike humans who think through understanding, AI merely predicts the next most likely word based on statistical patterns. Real-world cases, including a lawyer who cited fabricated court cases and Google's Bard demo error that wiped $100 billion in market value, demonstrate the serious consequences.]]></description>
                                    <author><![CDATA[Tekin Tian <tekintian@gmail.com>]]></author>
                                <guid>https://ai.tekin.cn/index.php/en/blog/ai-hallucination-why-ai-makes-things-up</guid>
                <pubDate>Mon, 16 Mar 2026 17:45:58 +0800</pubDate>
                <enclosure
                    type="image/webp"
                    url="http://ai.tekin.cn/storage/blog/ai-hallucination/1.webp"
                    length="113182"
                />
                                    <category>开源与 AI 应用专题</category>
                            </item>
                    <item>
                <title><![CDATA[macOS Zsh Function Naming Pitfall: A Deep Dive into Parse Errors Caused by Hyphens]]></title>
                <link>https://ai.tekin.cn/index.php/en/blog/macos-zsh-function-naming-trap-hyphen-parse-error</link>
                <description><![CDATA[This article explains macOS Zsh parse errors from hyphens in function names via a pyenv-Clang compatibility case. Zsh bans hyphens (option separators) with two-stage validation—no definition errors, but invocation failures. It contrasts Bash-Zsh differences, gives fixes like underscore replacement/aliases, clarifies Zsh naming rules, and shares best practices for macOS zsh/pyenv developers]]></description>
                                    <author><![CDATA[Tekin Tian <tekintian@gmail.com>]]></author>
                                <guid>https://ai.tekin.cn/index.php/en/blog/macos-zsh-function-naming-trap-hyphen-parse-error</guid>
                <pubDate>Sat, 07 Mar 2026 15:11:43 +0800</pubDate>
                <enclosure
                    type="image/webp"
                    url="http://ai.tekin.cn/storage/blog/202603/macos-zsh-function-naming-trap-hyphen-parse-error-1.webp"
                    length="31698"
                />
                                    <category>安全与运维专题</category>
                            </item>
                    <item>
                <title><![CDATA[Qwen3.5 Hybrid Attention: Gated DeltaNet + MoE Architecture & Deployment Guide]]></title>
                <link>https://ai.tekin.cn/index.php/en/blog/qwen3-5-hybrid-attention-gated-deltanet-moe-deployment</link>
                <description><![CDATA[This article provides a technical deep dive into Qwen3.5, Alibaba’s advanced large language model built on a hybrid attention architecture that combines Gated DeltaNet linear attention and full attention in a 3:1 layer ratio. It supports an ultra-long context window of up to 262K tokens while achieving much higher inference efficiency than traditional Transformers.]]></description>
                                    <author><![CDATA[Tekin Tian <tekintian@gmail.com>]]></author>
                                <guid>https://ai.tekin.cn/index.php/en/blog/qwen3-5-hybrid-attention-gated-deltanet-moe-deployment</guid>
                <pubDate>Fri, 06 Mar 2026 23:47:09 +0800</pubDate>
                <enclosure
                    type="image/webp"
                    url="http://ai.tekin.cn/storage/blog/202603/qwen3-5-hybrid-attention-cover-v2-1-20260306223817.webp"
                    length="86258"
                />
                                    <category>核心技术干货（企业级实战）</category>
                            </item>
                    <item>
                <title><![CDATA[[Underlying Logic Notes #25] The Reconfiguration of World Order: Global Logic for the Next Decade]]></title>
                <link>https://ai.tekin.cn/index.php/en/blog/underlying-logic-observation-notes-25-global-world-order-reconfiguration-next-decade</link>
                <description><![CDATA[Explore the underlying logic of the global order reconfiguration over the next decade. This post argues that power, technology, and energy structures are synchronously realigning, marking a shift from stability to restructuring. It analyzes key variables—compute, energy, supply chains—and China’s evolving role, offering a strategic roadmap for individuals to navigate this profound global transform]]></description>
                                    <author><![CDATA[Tekin Tian <tekintian@gmail.com>]]></author>
                                <guid>https://ai.tekin.cn/index.php/en/blog/underlying-logic-observation-notes-25-global-world-order-reconfiguration-next-decade</guid>
                <pubDate>Mon, 02 Mar 2026 20:44:52 +0800</pubDate>
                <enclosure
                    type="image/webp"
                    url="http://ai.tekin.cn/storage/blog/underlying-logic/25.webp"
                    length="74116"
                />
                                    <category>技术视野与选型</category>
                            </item>
                    <item>
                <title><![CDATA[Underlying Logic Observation Notes 24: The Future of Global Supply Chains — From Efficiency to Security]]></title>
                <link>https://ai.tekin.cn/index.php/en/blog/underlying-logic-observation-notes-24-global-supply-chain-future-efficiency-to-security</link>
                <description><![CDATA[The restructuring of supply chains isn’t a distant national policy — it’s a structural force shaping everyone’s destiny. This article deciphers the industrial logic, opportunity distribution, and tangible upward paths you can seize in the next decade.]]></description>
                                    <author><![CDATA[Tekin Tian <tekintian@gmail.com>]]></author>
                                <guid>https://ai.tekin.cn/index.php/en/blog/underlying-logic-observation-notes-24-global-supply-chain-future-efficiency-to-security</guid>
                <pubDate>Tue, 03 Mar 2026 12:00:05 +0800</pubDate>
                <enclosure
                    type="image/webp"
                    url="http://ai.tekin.cn/storage/blog/underlying-logic/24.webp"
                    length="61060"
                />
                                    <category>技术视野与选型</category>
                            </item>
                    <item>
                <title><![CDATA[Underlying Logic Observation Notes 23: Personal Strategy System in the AI Era — How Ordinary People Make Long-Term Choices]]></title>
                <link>https://ai.tekin.cn/index.php/en/blog/underlying-logic-observation-notes-23-personal-strategy-ai-era</link>
                <description><![CDATA[In the next decade, success hinges not on hard work alone, but on strategy. A actionable personal strategy system helps you choose the right direction, follow the optimal path, make critical decisions, and achieve sustained compound growth.]]></description>
                                    <author><![CDATA[Tekin Tian <tekintian@gmail.com>]]></author>
                                <guid>https://ai.tekin.cn/index.php/en/blog/underlying-logic-observation-notes-23-personal-strategy-ai-era</guid>
                <pubDate>Mon, 02 Mar 2026 20:31:09 +0800</pubDate>
                <enclosure
                    type="image/webp"
                    url="http://ai.tekin.cn/storage/blog/underlying-logic/23.webp"
                    length="62818"
                />
                                    <category>技术视野与选型</category>
                            </item>
            </channel>
</rss>
