Published: May 27, 2025

Introduction:

The term "AI Search" is everywhere these days. With new experiences like Google's SGE (Search Generative Experience) making waves, it's clear that the way we find information is on the cusp of a major shift. But amidst all the buzz, many are left wondering: What exactly is AI Search? How different is it from the Google we've known for years? And, perhaps most importantly, why do we even need AI Search? To answer this, we need to rewind and understand how traditional search engines tamed the internet, the challenges they faced, the role SEO played in this ecosystem, and finally, how AI emerged as the pivotal force in this ongoing evolution.

Chapter 1: The Global Digital Library – How Traditional Search Engines Made Sense of the Web

Imagine the internet as an infinitely vast library, constantly expanding every second, filled with billions of books but no central catalog. The mission of traditional search engines was to become the super-librarians of this digital realm, helping us locate the information we needed. They accomplished this monumental task primarily through three core steps:

Exploring the World (Crawling):
- Search engines dispatch armies of "crawlers" (or spiders), much like scouts uncovering the map in Age of Empires. These crawlers don't randomly scan IP addresses, nor do they rely on APIs from individual websites. They start with a "seed URL list" – think of them as globally recognized landmarks – and visit these addresses using basic HTTP requests.
- Next, they parse the returned HTML source code, much like reading the content of a book. They extract textual information and, crucially, identify all hyperlinks (links).
- These newly discovered links are the pathways to uncharted territories. Crawlers add them to their to-do list, continuously repeating this "visit -> parse -> extract links" cycle, gradually mapping out the vast web.
Building the Catalog (Indexing):
- The colossal amount of information brought back by crawlers needs to be organized to be useful. Search engines create a massive "inverted index." This isn't like a traditional book catalog; it's more like an index centered around "words." Under the term "quantum computing," for instance, it lists all web pages containing that term and their perceived importance.
- This enormous index is distributed across thousands of servers worldwide, ensuring the efficiency of subsequent queries.
Recommending the Books (Ranking & Query Processing):
- When a user types in keywords, the search engine interprets their intent and queries this massive index.
- Then, through complex algorithms (considering factors like relevance, authority, freshness, user experience, and more), it sorts through millions of potentially relevant results, ultimately presenting us with the familiar "ten blue links."

Chapter 2: The Art of Being Seen – The Role and Impact of SEO

Since search engines are the primary gateway to information, website owners naturally want their "books" to be recommended. This gave rise to SEO (Search Engine Optimization). The goals of SEO are:

Technical Aspect: To make a website crawler-friendly, ensuring it can be smoothly crawled and indexed (e.g., optimizing site speed, structure, submitting sitemaps).
Content Aspect (On-Page): To create high-quality content that is highly relevant to user queries, and to optimize titles and text so algorithms understand its value.
Authority Aspect (Off-Page): To demonstrate credibility and importance by acquiring high-quality links from other websites.

SEO is essentially an ongoing dialogue and adaptation between websites and search engines, aimed at making valuable information more discoverable.

Chapter 3: The Limits of Tradition – Why Change Was Needed

Despite their incredible achievements, traditional search engines faced growing challenges:

Information Overload: A simple list of links often required users to click through multiple pages to piece together a complete answer.
The Intent Gap: Keywords sometimes struggled to accurately convey complex query intentions.
Quality Challenges: The existence of SEO also brought the risk of spammy content and ranking manipulation.
Evolving User Expectations: People increasingly craved direct answers, not just "signposts."

These challenges indicated that the traditional "catalog-style" search model was becoming less adequate for increasingly complex information needs.

Chapter 4: AI Enters the Scene – The Librarian Becomes a Smart Consultant

This is where AI Search steps in. It's not about tearing down the old system but rather an intelligent upgrade built upon the existing foundation. AI's intervention is most significantly felt in:

Deeper Understanding: Leveraging Large Language Models (LLMs), AI can grasp the nuances of natural language, complex sentence structures, context, and the user's true intent.
Powerful Synthesis: AI doesn't just find relevant pages; it can "read" and "understand" their content. It can extract key pieces of information from multiple sources and synthesize, summarize, compare, or even infer to form a coherent, direct answer.
More Direct Presentation: It can generate generative summaries (like Google's SGE), answer questions directly, offer multi-faceted perspectives, and even engage in follow-up conversations, shifting from "providing options" to "solving problems."

Crawlers are still crawling, and indexes are still being built, but the "processing" in the middle and the final "presentation" have been qualitatively transformed by AI.

Chapter 5: Why AI Search? – The Driving Forces and Future Outlook

We need AI Search because:

It aligns more intuitively with how humans seek knowledge: We often prefer a direct answer to a list of references to sift through.
It can handle the exponential growth of information: AI's ability to synthesize is an effective way to process vast amounts of data.
The technology has matured: Breakthroughs in LLMs have made it possible.

However, AI Search also introduces new challenges: the accuracy of AI-generated content (hallucinations), the potential impact on traffic to original websites, the risk of algorithmic bias, and the transformative pressure on the SEO industry. The existence of niche search engines like DuckDuckGo (privacy-focused) and Kagi (paid, ad-free, customizable) also demonstrates the market's demand for differentiation and specific value propositions. How they evolve in the AI era will be fascinating to watch.

Conclusion:

AI Search isn't an isolated invention; it's a necessary evolution driven by search engines' efforts to overcome their own limitations, meet user expectations, and leverage the latest technology. It marks a shift from an era of "searching for information" to a new age of "interacting with information and gaining insights." As we embrace the convenience AI Search offers, we must also maintain a critical perspective, understand its underlying mechanisms, and continually assess the profound impact its development will have on the information ecosystem, content creators, and indeed, each one of us. This journey of discovery has only just begun.

Why AI Search? 揭開搜尋引擎的演進脈絡與未來圖景

發布日期： 2025年5月27日

前言：

近年來，「AI Search」的浪潮席捲而來，以 Google 的 SGE (Search Generative Experience) 為代表的新型態搜尋模式，預示著我們獲取資訊的方式將迎來巨變。然而，在這股熱潮之下，許多人不禁要問：AI Search 究竟是什麼？它與我們熟悉的 Google 搜尋有何不同？更重要的是，我們為什麼需要 AI Search？ 要回答這個問題，我們必須回溯源頭，深入理解傳統搜尋引擎如何運作，它們面臨何種挑戰，以及 SEO 如何在其中扮演角色，最終才能看清 AI 如何成為這場演進的關鍵驅動力。

第一章：全球數位圖書館 — 傳統搜尋引擎如何馴服網路

想像網際網路是一座無邊無際、每秒都在擴建的巨大圖書館，藏書億萬，卻沒有總目錄。傳統搜尋引擎的任務，就是成為這座圖書館的超級管理員，幫助我們找到所需的資訊。它們主要透過三個核心步驟完成這項壯舉：

探索世界 (Crawling - 網路爬取)：
- 搜尋引擎派出無數的「爬蟲大軍」，如同《世紀帝國》中探索地圖的斥侯。它們並非漫無目的地掃描 IP 位址，更非依賴各網站的 API。它們從一個「種子 URL 列表」（像是全球知名地標）出發，透過基礎的 HTTP 請求訪問這些網址。
- 接著，它們解析 (Parse) 回傳的 HTML 原始碼，像閱讀書本內容一樣，提取文字資訊，並最關鍵地，找出所有超連結 (Links)。
- 這些新發現的連結，就是通往未知領域的新路標。爬蟲將它們加入待辦清單，不斷循環這個「訪問 -> 解析 -> 提取連結」的過程，逐步繪製出龐大的網路地圖。
建立目錄 (Indexing - 索引建立)：
- 爬蟲帶回的海量資訊，需要被整理才能使用。搜尋引擎會建立一個巨大的「倒排索引 (Inverted Index)」。這不像傳統的書目卡，而是像一本以「詞彙」為核心的索引：在「量子計算」這個詞條下，列出所有包含它的網頁及其重要性。
- 這個龐大的索引被分散儲存在全球數千台伺服器上，確保了後續查詢的效率。
推薦書籍 (Ranking - 查詢處理與排名)：
- 當使用者輸入關鍵字時，搜尋引擎會理解其意圖，並查詢這個巨大的索引。
- 接著，透過複雜的演算法（考量相關性、權威性、新鮮度、使用者體驗等多重因素），對數百萬個可能相關的結果進行排序，最終呈現出我們熟悉的「十條藍色連結」。

第二章：被看見的藝術 — SEO 的角色與影響

既然搜尋引擎是資訊的主要入口，網站營運者自然希望自己的「書」能被優先推薦。這就催生了 SEO (Search Engine Optimization)。SEO 的目標是：

技術層面： 讓網站對爬蟲友好，確保能被順利爬取和索引（例如：優化網站速度、結構、提交 Sitemap）。
內容層面 (On-Page)： 創作高品質、與使用者查詢高度相關的內容，並優化標題和內文，讓演算法理解其價值。
權威層面 (Off-Page)： 透過獲取其他網站的高品質連結，證明自身的信譽與重要性。

SEO 本質上是網站與搜尋引擎之間的一場持續對話與適應，旨在讓有價值的資訊更容易被需要的人找到。

第三章：傳統搜尋的困境 — 為何需要改變？

儘管傳統搜尋引擎創造了奇蹟，但它也面臨著日益嚴峻的挑戰：

資訊過載： 簡單的連結列表，往往需要使用者點擊多個頁面才能拼湊出完整答案。
意圖鴻溝： 關鍵字有時難以精確表達複雜的查詢意圖。
品質挑戰： SEO 的存在也伴隨著垃圾內容和排名操縱的風險。
使用者期待： 人們越來越渴望直接的答案，而非僅僅是「路標」。

這些挑戰表明，傳統的「目錄式」搜尋模式，已逐漸難以滿足日益複雜的資訊需求。

第四章：AI 登場 — 圖書館管理員變身智慧顧問

這就是 AI Search 登場的契機。它並非推倒重來，而是建立在現有基礎之上的一次智慧升級。AI 的介入，主要體現在：

更深層的理解： 借助大型語言模型 (LLM)，AI 能理解自然語言的細微之處，掌握使用者的真實意圖。
更強大的整合： AI 能「閱讀」並「整合」多個網頁的資訊，合成一個全面、直接的答案。
更直接的呈現： 它能生成摘要性文字，直接回答問題，甚至進行多輪對話，從「提供選項」轉變為「解決問題」。

爬蟲依然在爬取，索引依然在建立，但中間的「處理」與最終的「呈現」環節，因為 AI 的加入而發生了質變。

第五章：Why AI Search? — 驅動力與未來展望

我們需要 AI Search，因為：

它更符合人類獲取知識的直覺： 我們希望直接獲得答案，而非一堆待讀的參考文獻。
它能應對資訊的爆炸式增長： AI 的整合能力是處理海量資訊的有效途徑。
技術已經成熟： LLM 的突破為其提供了可能。

然而，AI Search 也帶來新的挑戰：AI 生成內容的準確性（幻覺）、對原始網站流量的衝擊、演算法偏見的風險，以及 SEO 行業的轉型壓力。小眾搜尋引擎如 DuckDuckGo（注重隱私）和 Kagi（付費無廣告、可客製化）的存在，也證明了市場對差異化和特定價值主張的需求，它們在 AI 時代將如何演進，值得關注。

結論：

AI Search 不是一個獨立的發明，而是搜尋引擎為了克服自身局限、滿足使用者期待，並利用最新技術而進行的必然演進。它標誌著我們從「搜尋資訊」的時代，邁向「與資訊互動、獲取洞察」的新紀元。當我們擁抱 AI Search 帶來的便利時，也必須保持批判性思維，理解其背後的運作機制，並持續關注其發展對資訊生態、內容創作者乃至我們每一個人所帶來的深遠影響。這趟探索之旅，才剛剛開始。