English
Based on โAlien contact will challenge every belief system on Earth | Sara Seagerโ from Big Think Watch the original video
The Ultimate Awakening: How Alien Life Will Shatter Our Worldviews
For centuries, humanity has gazed at the stars, pondering the ultimate question: Are we alone? While science fiction has long explored the fantastical implications of extraterrestrial contact, the reality of such a discovery, even of microbial life, promises to be far more profound, challenging the very foundations of our belief systems. According to Dr. Sara Seager, an astrophysicist and planetary scientist at MIT, the robust evidence for life beyond Earth would ignite a โcosmic awakeningโ on par with, if not exceeding, the Copernican revolution.
The scientific ramifications alone would be staggering. If we were to find definitive proof that life emerged independently on another planetโeven just onceโit would provide compelling evidence that our galaxy, perhaps even the entire universe, is teeming with life. โIf life emerged two separate times, like two independent times from each other,โ Seager explains, โit will give us information that thereโs a high likelihood that our galaxy, our universe, is teeming with life.โ This isnโt just wishful thinking; the scientific groundwork is already being laid. Weโve discovered complex organic molecules on Mars, identified liquid environments in Venusโs atmosphere, and theorize about vast subsurface oceans on Mars and many moons of Jupiter and Saturn. These are all critical ingredients for life as we know it, suggesting that the universe might be far more hospitable than we once imagined. And if microbial life is widespread, the possibility of intelligent life inevitably looms largeโa prospect, Seager notes, โweโre gonna have to grapple with.โ
But the true impact of such a discovery wouldnโt be confined to scientific journals. It would resonate through every facet of human society, touching religion, philosophy, culture, and our collective self-perception. Seager vividly illustrates this point with a personal anecdote from a keynote speech she delivered in Thunder Bay, Ontario, for the Royal Astronomical Society of Canada. She was billeted by a devout Catholic woman, whom Seager affectionately calls โSpunky Widow.โ This woman had tried to convey to her Catholic friends the astronomical reality of countless planets beyond our solar system, but they struggled to accept it. โThe people didnโt really wanna believe it โcause it was like a harsh reality for their own system of thinking,โ Seager recounts. Her host simply wanted Seager to legitimize the field of exoplanet research for her community. What ensued was an intense, late-night discussion that Seager and the community dubbed โthe awakeningโโa moment of profound realization that our Earth is not unique in harboring planetary neighbors. If merely the existence of other planets could trigger such a shift, imagine the seismic impact of discovering life itself.
This โawakeningโ finds a powerful historical parallel in the Copernican revolution. For centuries, humanity believed Earth was the unmoving center of the universe, a geocentric model underpinned by religious doctrine. When Nicolaus Copernicus proposed in the 16th century that Earth, in fact, revolved around the Sun, it was considered heresy. It challenged not just scientific understanding but humanityโs very special place in the cosmos. Did the world immediately embrace this new heliocentric truth with excitement? โIn fact, no. No. There was not a definitive moment,โ Seager emphasizes. โIt took a very long time for it to happen.โ Hundreds of years passed before the heliocentric model became widely accepted, a slow and arduous process of re-evaluation and adaptation. Seager predicts a similar trajectory for the acceptance of extraterrestrial life: โIt took a really long time to have like legit definitive proof. And so over hundreds of years, which is likely to be the same for our definitive signs of life, itโll just be something as part of our culture.โ
Beyond the profound philosophical implications, the search for life beyond Earth also underscores the inherent value of โpure scienceโโresearch driven by curiosity rather than immediate practical application. Critics sometimes question why resources should be dedicated to exploring distant stars or searching for alien microbes when so many terrestrial problems demand attention. Yet, history repeatedly demonstrates that pure science often leads to the most transformative, unexpected discoveries. Seager points to the discovery of insulin as a prime example of fundamental biological research yielding life-saving medical breakthroughs.
In astronomy, a field often perceived as purely academic, the practical benefits are ubiquitous. The techniques developed to process faint signals from the night sky into discernible images have directly contributed to advancements in medical imaging. And then thereโs GPS. โWho doesnโt use GPS on a daily basis?โ Seager asks. Yet, this indispensable technology wasnโt the result of a planned project to help people navigate. โNo, it was just people fooling around with rocketry, wanting to see if they could launch rockets and try to get in north orbit.โ From the seemingly abstract pursuit of launching objects into space, an entire global positioning system emerged, revolutionizing transportation, logistics, and countless other industries.
So, will the search for extraterrestrial life lead to similar practical discoveries? โItโs all really to be determined,โ Seager admits. The giant leaps forward in human knowledge and capability are often โfew and far between,โ and crucially, โwe canโt get there purposely.โ Instead, progress often follows a โrandom walk through science, through discovery.โ
The quest for alien life is more than just a scientific endeavor; itโs a journey into the unknown that promises to redefine our understanding of existence itself. Whether itโs the discovery of a microscopic organism beneath an icy moon or the eventual, perhaps gradual, acceptance of intelligent civilizations elsewhere, the impending โawakeningโ will compel humanity to re-evaluate its place in the cosmos. It will be a challenging, often uncomfortable process, echoing the intellectual battles of centuries past. But like the Copernican revolution and the countless unexpected benefits of pure science, it promises to usher in an era of unprecedented understanding and, perhaps, an entirely new perspective on what it means to be alive.
Based on โ35M Users. $100M ARR. My 10-Year Bet Was Right. | Otter.ai, Sam Liangโ from EO Watch the original video
The Spoken Revolution: How Otter.ai is Unlocking a Century of Lost Knowledge
Imagine a world where the wisdom of Shakespeare or the intricate theories of Charles Darwin were not confined to the written word, but preserved in their own voices, accessible for study and understanding. This profound loss of human intelligence, repeated countless times throughout history, is a driving frustration for Sam Liang, co-founder and CEO of Otter.ai. His company, a decade-long bet on the transformative power of voice AI, is not just transcribing meetings; itโs building a future where no valuable spoken word is ever lost again, and business intelligence flows as effortlessly as conversation itself.
With over 35 million users and exceeding $100 million in annual recurring revenue (ARR), Otter.ai has moved far beyond a simple transcription tool. It has evolved into an AI meeting assistant, and is now architecting a โmeeting-centric enterprise knowledge base with agentic workflows on top of it.โ This journey, however, began with a radical idea that initially made many uncomfortable.
The Genesis of a Grand Vision: From Lost Voices to AI Intelligence
Sam Liangโs vision for Otter.ai is rooted in a deep understanding of technologyโs potential to shape the future. He recalls a recent visit to Harvard University, where professors still resist tools like Otter.ai, fearing they might hinder learning. โThatโs old thinking,โ Liang asserts, highlighting that our educational systems, largely conceived over a century ago, must embrace AI. The sheer volume of โvoice knowledgeโ lost over millennia โ from the speeches of ancient leaders to the everyday insights of countless individuals โ represents an โtremendous loss of human knowledge and human intelligence.โ This insight fueled his conviction that โvoice AI will be really huge in the future.โ
Liangโs background prepared him for such ambitious undertakings. He earned his PhD at Stanford University, where his advisor, David Cheriton, instilled in him the importance of โthinking bigโ and identifying ideas that could โgenerate a big impact, what will change the world.โ Cheriton, famously, wrote a $100,000 check to Larry Page and Sergey Brin before Google was even a tangible product, recognizing their talent and vision. Liang himself contributed to groundbreaking innovation as the lead of the Google Maps location platform from 2006 to 2010, before leaving to found a successful mobile startup focused on persistent sensing and personalized mobile services.
It was during the grueling process of building his first startup, constantly juggling investor meetings, internal team discussions, and customer calls, that Liang experienced firsthand the challenge of retaining and sharing crucial meeting content. โReally hard for me to remember all the meeting content. Itโs also hard to share that knowledge with all the team members,โ he explains. He knew there had to be a better way.
A Bet Against Convention: The Evolution of Otter.ai
In 2016, Liang and his team embarked on a mission to โrecord everythingโ and enable seamless sharing. This idea was, at the time, deeply unsettling for most. โBeing recorded is uncomfortable,โ Liang admits, โand also share meeting notes with other people. Itโs uncommon because traditionally people take notes on the paper notebook. Itโs a personal thing.โ
This cultural barrier was perhaps the biggest hurdle. Yet, Liang anticipated a shift. โWe anticipate that the mindset will change, the culture will change. So we build the product that enable that change.โ He understood that while you can convince some, you canโt convince everyone, and thatโs okay. Like any new product, adoption follows a curve. Early adopters, who quickly grasped the value of Otter.ai, became more effective and productive, naturally demonstrating the benefits to their colleagues and helping to drive wider acceptance.
The Power of Deep Technology: Building for the Future, Not Just Today
What truly sets Otter.ai apart, and what Liang believes is essential for any company aiming โto really go big,โ is โdeep technology roots.โ In 2016, building a simple meeting note-taker using existing APIs was feasible for โany college student.โ But Otter.ai chose a far more challenging path: developing its own proprietary speech recognition technology.
โAt that time if we were waiting for someone else to create the API, we would be many years late,โ Liang states. This decision, fraught with risks and requiring significantly fewer resources than giants like Google or Microsoft, was a strategic bet on long-term differentiation. Building their own technology allowed Otter.ai to control costs, offer more robust free services, and ultimately create a โnew revolution in the future.โ
Liang emphasizes that โwhat differentiation can you create? Thatโs the biggest problem for a new startup.โ Relying on third-party APIs means paying significant fees, limiting innovation and the ability to offer competitive pricing or free tiers. Moreover, truly โunsolved problemsโ in AI, such as accurately modeling complex human conversations with multiple speakers and their interactions, demand deep, in-house expertise. โTo solve that problem, you cannot just rely on third party APIs. You have to build your own deep AI tag,โ he asserts. If a solution is โtoo easy for you to build, itโs very easy for 100 other people to build as well.โ
The Future is Spoken: A Paradigm Shift in Business Intelligence
Liang sees a profound shift in human-computer interaction on the horizon, driven by voice AI. He points to historical precedents: before the internet, email felt ubiquitous; then Slack emerged, reducing email reliance. Now, with increasingly mature voice technology, Liang believes โvoice will become the primary interface for enterprise intelligence.โ
His prediction is bold: โYou probably donโt need to write so much in a few years. People will rarely write anything. They will rarely use keyboard to write anything. They can just talk because talk is easier than writing. They can just talk and our AI will write everything for you.โ This isnโt a distant fantasy; itโs already happening, with many using AI to draft documents, emails, and social media posts. This trend, Liang insists, โwill only accelerate.โ The implication is clear: the keyboard, once revolutionary, will largely be superseded by the natural ease of speech, with AI acting as the scribe and organizer of our spoken thoughts.
Building a Generational Company: Persistence and Vision
Despite Otter.aiโs impressive growth, Liang maintains a long-term perspective. โAt least 95% or even higher. I would say 99% of the world hasnโt adopted a tool like Otter yet.โ This vast untapped market fuels his ambition. โWe have to look at the next 10 years not just today. Thatโs how this generational companies are built.โ
Building such a company is no easy feat. Liang, an avid marathon runner with 11 under his belt and another on the horizon, offers a telling comparison: โPeople say building a startup like running a marathon. Actually building a startup is way harder than running a marathon.โ His running habit helps him stay healthy, manage stress, and cultivate the mental fortitude required to โpush through all the challenges.โ The startup journey is defined by expected difficulties, and success hinges on persistence. โMost people give up pretty fast,โ he notes. The key is to โpersist and and continue pursuing your goal.โ
Sam Liangโs 10-year bet on voice AI is paying off handsomely, but for him, itโs just the beginning. Otter.ai isnโt merely a productivity tool; itโs a foundational technology aiming to revolutionize how we capture, process, and leverage human intelligence, ensuring that the spoken word, once ephemeral, becomes an enduring and accessible asset for all.
Based on โFirecrawl AI clearly explained (and how to make $$)โ from Greg Isenberg Watch the original video
Unlocking the Internetโs Treasure Chest: How Firecrawl Gives AI Eyes and Hands (and How You Can Build a Fortune)
In the rapidly evolving landscape of artificial intelligence, a critical limitation has persisted: AI, for all its brilliance, has been largely blind. It can process information, generate text, and even write code, but it struggles to truly see and interact with the dynamic, ever-changing world of the internet. It canโt navigate websites, extract specific data points, or understand the visual context of a webpage. This fundamental โblindnessโ has been a significant barrier to building truly autonomous and intelligent AI applications.
Enter Firecrawl AI, a revolutionary tool poised to become the โeyes and handsโ for artificial intelligence. Itโs not just another web scraper; itโs a foundational piece of infrastructure that could redefine whatโs possible with AI, empowering a new generation of entrepreneurs to build highly valuable businesses.
The Evolution of AI: From Chatbots to Autonomous Agents
To understand Firecrawlโs significance, we need to trace AIโs recent journey. The โfirst eraโ of modern AI, spearheaded by tools like ChatGPT in late 2022, was the chatbot era. These AIs could answer questions, generate creative text, and engage in conversations, but their capabilities were largely confined to the data they were trained on or manually fed.
Next came the co-pilot era, exemplified by tools like GitHub Co-pilot. Here, AI acted as an accelerator, helping humans write code faster or complete tasks more efficiently. The human, however, remained firmly in the driverโs seat, guiding the AIโs every action.
Weโve now entered the AI agent era, often referred to as the โcomputer use era.โ This is where AI begins to do the work for you. Think of tools like Claude Code, which can browse, research, and even build. These agents are designed to act autonomously, but they still face a crucial hurdle: they need clean, reliable data from the web to perform effectively. While advanced models like GPT-4 and Claudeโs computer use API can โseeโ and control computers, they still need a robust mechanism to intelligently gather and process web information.
This is precisely where Firecrawl steps in.
Firecrawl: The Bridge Between AI and the Web
At its core, Firecrawl is a powerful API that transforms messy, unstructured web content into clean, usable data for any AI model. Imagine feeding a website URL into Firecrawl and instantly receiving a perfectly formatted markdown document, a structured JSON object, or even screenshots โ all ready for your AI to consume.
This is a stark contrast to the โold wayโ of web scraping, which was a massive headache:
- Custom scripts per site: Every website required a bespoke scraper, prone to breaking with minor layout changes.
- Proxy and anti-bot management: Navigating CAPTCHAs and IP blocks was a constant battle.
- Manual HTML parsing: Extracting meaningful data from raw HTML was tedious and error-prone.
Firecrawl cuts through this complexity with a single API call. It handles proxies, bypasses anti-bot detection, and uses AI to intelligently extract structured data from virtually any site (reportedly 98-99% of them), even adapting to layout changes. This means developers and entrepreneurs can focus on building intelligent applications, not on the plumbing of data acquisition.
Six Superpowers for Your AI
Firecrawl isnโt just a simple scraper; it equips your AI with a suite of โsuperpowersโ:
- Scrape: Extract clean markdown from a single webpage.
- Crawl: Automatically navigate and gather data from an entire website, like all articles on CNN.com.
- Map URLs: Instantly generate a comprehensive list of all URLs on a domain, providing valuable metadata.
- Search: Leverage Googleโs search capabilities and integrate full content in one call.
- Agent-Driven Data: Describe the data you need (e.g., โ50 highest-rated Cuban restaurants in South Floridaโ), and Firecrawlโs agent will find and deliver it.
- Browser Control (Browser Sandbox): Your AI can control a real browser โ filling out forms, clicking buttons, handling logins, navigating pagination, and even allowing you to watch its actions live.
This level of control and data extraction, often achievable with just three lines of code, is what makes Firecrawl so transformative.
The AWS Moment for Web Data
The impact of Firecrawl can be likened to Amazon Web Services (AWS) in 2006. Before AWS, building a web application meant buying expensive servers, managing racks and cables, and constantly dealing with hardware failures. AWS revolutionized this by offering cloud infrastructure via simple API calls, allowing startups to focus on product innovation rather than server management. This shift enabled the creation of multi-billion and even trillion-dollar companies.
Firecrawl is doing the same for web data. Instead of building custom scrapers, managing proxies, and dealing with anti-bot measures, developers can now make one API call and get clean, structured web data. This liberates builders to concentrate on creating incredibly valuable software products that leverage this โnew oilโ โ clean, structured data.
Your AI Agent Stack: Firecrawl as the โEyes and Handsโ
For any builder looking to create advanced AI applications, a robust โagent stackโ is essential. This stack typically comprises five layers:
- Agent Harness: A platform that orchestrates various AI agents (e.g., Claude Code, Cursor, CodeX).
- Search Layer: Tools for broad information retrieval (e.g., Perplexity, Exa).
- Web Data Layer: This is where Firecrawl shines, providing the crucial scraping, browsing, and extraction capabilities. Itโs how your agents see the internet.
- Ops Brain: A knowledge base for storing context, notes, and institutional memory (e.g., Notion, Obsidian).
- Outbound & Audience Stack: Tools for outreach and distribution (e.g., Instantly, Apollo).
Firecrawl is the critical component that gives your AI agents the ability to perceive and interact with the vast information on the internet, turning raw web pages into actionable intelligence.
From Data to Dollars: Niche Startup Ideas with Firecrawl
The real power of Firecrawl lies in its potential to fuel a new wave of highly profitable, niche-focused businesses. While incumbents like SEMRush or Indeed offer broad, generic tools, Firecrawl allows you to create hyper-specific solutions that perfectly serve a particular customer segment.
Here are six compelling startup ideas, demonstrating how Firecrawl can transform complex problems into lucrative opportunities:
-
Niche Price Monitoring:
- Incumbents: General-purpose price trackers (e.g., Precinct, Visual Ping) costing hundreds of dollars monthly.
- Firecrawl Opportunity: Build a โSneaker Resale Price Tracker.โ Firecrawl monitors platforms like StockX, Goat, and eBay daily. AI alerts users to price drops or rising trends on specific sneaker models.
- Monetization: Charge $50/month for alerts or sell detailed reports for $500. Pick any niche you understand better than others โ collectibles, rare books, specific electronics.
-
Hyper-Niche SEO Gap Finder:
- Incumbents: SEO giants like Ahrefs and SEMRush, offering complex, general-purpose dashboards for hundreds of dollars.
- Firecrawl Opportunity: Create โSEO Audits for Dentists Only.โ Firecrawl reads competitor dentist sites and Google My Business (GMB) listings. AI generates a one-click report showing, for example, โYou rank for 12 keywords, competitors rank for 47.โ
- Monetization: Sell these targeted reports for $200-$500 each, or offer ongoing monitoring.
-
Vertical Job Aggregation:
- Incumbents: Massive horizontal platforms like Indeed or ZipRecruiter, with millions of generic listings.
- Firecrawl Opportunity: Develop โRemote AI/ML Jobs Only.โ Firecrawl monitors 500 top company career pages daily. AI filters and ranks jobs by a โfit score.โ
- Monetization: Charge $29/month for premium alerts, delivering the 50 most relevant jobs instead of 300 million generic ones.
-
Specialized AI Research Reports:
- Incumbents: General research tools (e.g., Consensus, Tavali) where users do the prompting.
- Firecrawl Opportunity: Build โNiche Crypto Token Due Diligence Reports.โ Firecrawl reads whitepapers, Twitter, and other sources. AI auto-generates a risk score and summary.
- Monetization: Sell these highly valuable reports to VCs, private equity, or crypto funds for $1,000-$5,000/month. A $5,000 report is a small price to avoid a bad $500,000 investment.
-
Agent in a Box (for Specific Industries):
- Incumbents: Horizontal enterprise agent platforms (e.g., Harvey AI) with long sales cycles and high customization costs.
- Firecrawl Opportunity: Create a โReal Estate Comp Report Agent.โ Firecrawl pulls listings, tax records, and permit data. The agent generates comprehensive comparative market analyses in 30 seconds.
- Monetization: Sell to realtors for $300/month. This is a direct, tangible value proposition for a specific professional.
-
Review Intelligence for Niche Sellers:
- Incumbents: Broad social and review monitoring tools (e.g., Brand24, AppFollow) with generic sentiment analysis.
- Firecrawl Opportunity: Build an โAmazon FBA Seller Review Tracker.โ Firecrawl monitors competitor product reviews daily. AI spots trends (e.g., โComplaints about battery life up 40%โ).
- Monetization: Sell to Amazon sellers for $99/month. This helps them identify product gaps and market opportunities before competitors.
These examples highlight a critical insight: vertical software is a massive business. People prefer specific tools that do one thing perfectly for their niche, rather than generic, complex platforms. By leveraging Firecrawl, you can create these highly focused, high-margin products.
The Firecrawl Framework for Building Wealth
Ready to start building? Hereโs a simple, five-step framework:
- Pick a Niche: Identify an industry where people actively pay for specific data or insights.
- Build the Scraper: Use Firecrawlโs API (perhaps with a simple Python script or an AI assistant like Claude Code) to get the data.
- Package It: Present the data in a valuable format โ a CSV, a dashboard, Slack alerts, or even an API.
- Sell the Output: Donโt just sell the tool; sell the insights and data it provides. Charge $500-$5,000 per month per client.
- Automate It: Schedule your scrapers and processes to run autonomously, allowing you to compound clients and scale while you sleep.
This flywheel of niche focus, efficient data acquisition, and automated delivery is just beginning.
The Future is Now: AI as Employees
Perhaps the most surprising glimpse into the future offered by Firecrawl comes from its own hiring practices. A year ago, Firecrawl posted a job opening for an โExample Creatorโ โ with the caveat: โPlease only apply if youโre an AI agent.โ They sought an AI agent capable of autonomously researching trends, creating, testing, and refining example applications to showcase Firecrawlโs potential.
This seemingly whimsical job posting reveals a profound shift: the possibility of AI agents becoming specialized โemployees.โ Imagine:
- Content Creator Agent: Writes blog posts, watches metrics, and improves autonomously for $5,000/month.
- Customer Support Agent: Handles tickets, escalates when necessary, all for $5,000/month.
- Junior Developer Agent: Triages GitHub issues, writes documentation and code, also for $5,000/month.
This suggests an opportunity to build AI agents using tools like Firecrawl that companies would want to โhire.โ These agents, powered by Firecrawlโs ability to see and interact with the web, could perform valuable tasks for a fraction of the cost of human employees, signaling a new frontier for entrepreneurship.
Your Opportunity Awaits
Firecrawl AI represents a pivotal moment in the AI revolution. By providing AI with the โeyes and handsโ to navigate and understand the internet, it unlocks unprecedented opportunities for innovation. Whether youโre an experienced developer or an aspiring entrepreneur, understanding the web data layer and leveraging tools like Firecrawl is a 12-month head start in building the next generation of valuable software.
The incumbents are charging hundreds for generic tools; your opportunity is to charge less for a hyper-focused tool that does one thing perfectly for one specific customer. The gold rush is here, and clean, structured data is the new gold. The question is: what will you build with it?
Based on โStop doing admin work manually. Let AI handle it.โ from How I AI Watch the original video
Ditch the Guesswork: How AI is Revolutionizing Cloud Administration
In the intricate world of cloud computing, managing permissions and configurations can feel like navigating a labyrinth blindfolded. For IT professionals, the sheer complexity of assigning the right roles, activating necessary privileges, and remembering obscure prerequisites often leads to frustrating dead ends and wasted time. But what if an intelligent assistant could not only guide you but proactively prevent common pitfalls, making administrative tasks smoother and more efficient?
As one expert from the โHow I AIโ channel reveals, this future is not only possible but already here, transforming the way administrators interact with platforms like Microsoft Azure. The core message is clear: stop doing admin work manually. Let AI handle it.
The Role Assignment Riddle
Imagine needing to grant a team member access to a specific Azure service, like Azure Document Intelligence, but being unsure of the exact permissions required. Manually sifting through documentation, trial-and-error, or escalating to senior engineers can be a significant time sink. This is a common pain point, as the speaker candidly admits: โThere are times when I have no idea what role somebody needs to do something. Iโll be like, give this person whatever role they need to use Azure Document Intelligence and like you figure it out.โ
Traditionally, this โfigure it outโ approach often meant a frustrating cycle of assigning roles, testing, and troubleshooting when permissions inevitably fell short or were over-privileged. However, AI offers a sophisticated alternative. Instead of leaving the AI to its โown devices,โ the speaker connects it directly to comprehensive knowledge basesโspecifically, the Microsoft documentation, which he refers to as an โMCP server.โ
โI can connect it to the Microsoft documentation MCP server and then itโll go look it up and that makes it work much better,โ he explains. This means the AI isnโt guessing; itโs leveraging an authoritative source of truth. When tasked with enabling someone to use Azure Document Intelligence, the AI can intelligently query the documentation, understand the serviceโs requirements, and recommend the precise roles needed. This eliminates guesswork, reduces errors, and ensures that users have just the right level of access โ a critical aspect of cloud security and compliance.
Beyond Lookup: Proactive Problem Prevention
The utility of AI extends far beyond simply looking up information. One of the most compelling insights shared is the AIโs ability to enforce rules and act as a proactive reminder system, preventing common administrative blunders.
A frequent hurdle in Azure administration, for instance, is the necessity to activate โowner accessโ before making significant changes, particularly when assigning roles on a resource group. This is often a security measure, like Azure Privileged Identity Management (PIM), where high-privilege roles are โjust-in-timeโ activated to minimize exposure. Forgetting this step can lead to failed operations and wasted effort.
โThis is one of the common problems that I have is that I have not activated my owner access, which is like a hurdle I have to go through,โ the speaker notes. Rather than enduring repeated failures, heโs configured his AI assistant, which he calls โWarp,โ to intervene.
Warp isnโt just a passive tool; itโs an active participant in the workflow. โI make Warp remind me and Warp will be like, โSo, hey, did you activate your owner access before I start doing this?โ Cuz otherwise, itโs going to fail.โ This proactive intervention is a game-changer. It transforms a potential source of frustration and delay into a seamless, guided process. By embedding these โrulesโ and contextual knowledge into the AI, administrators can offload the mental burden of remembering every prerequisite, allowing them to focus on higher-value tasks.
The Broader Impact: Smarter, Faster, More Reliable Admin
The examples of intelligent role assignment and proactive reminders highlight a fundamental shift in how we approach IT administration. By equipping AI with access to vast documentation and custom-defined rules, we empower it to:
- Reduce Errors: AI minimizes human error by ensuring correct permissions are applied and necessary steps are followed.
- Boost Efficiency: Tasks that once required manual research and troubleshooting are now automated or guided, saving significant time.
- Enhance Security: By ensuring โleast privilegeโ access and adherence to security protocols (like PIM activation), AI strengthens the overall security posture.
- Lower Cognitive Load: Administrators are freed from remembering intricate details and prerequisites, allowing them to concentrate on strategic planning and problem-solving.
While the speaker specifically mentions its utility for โcoding as well,โ he finds it โsuper useful for these kinds of things that I use it,โ referring to the often-overlooked yet critical administrative tasks. The implication is clear: whether itโs configuring cloud resources, managing user access, or even streamlining development workflows, AI is poised to become an indispensable partner.
In an era where cloud environments are growing exponentially in complexity, relying solely on manual processes is no longer sustainable. Tools like Warp, powered by intelligent access to documentation and user-defined rules, represent a significant leap forward. They promise a future where administrative work is less about guesswork and more about intelligent, automated execution, allowing IT professionals to truly โAIโ their way to better productivity and more reliable operations. The message is unequivocal: itโs time to stop doing admin work manually. Let AI handle it.
Based on โโEverything After This Will Be Harderโ: General Stanley McChrystal on Iranโ from New York Times Podcasts Watch the original video
The End of Easy Victories: General McChrystal on Why Everything After This Will Be Harder
In an era of complex global challenges and shifting geopolitical landscapes, the wisdom of seasoned military leaders offers invaluable perspective. Recently, David French, a New York Times columnist and veteran, sat down with General Stanley McChrystalโa figure described by former Secretary of Defense Robert Gates as โperhaps the finest warrior and leader of men in combat I have ever met.โ Their conversation delved deep into the ongoing conflict with Iran, the illusions of modern warfare, and the fundamental questions facing American leadership and society. What emerged was a stark warning: the easy victories are over, and the path ahead promises only greater difficulty.
The Long Shadow of History: Understanding Iranโs Grievances
French, drawing on his own experience as a JAG officer in eastern Diala province, Iraq, during 2007-2008, highlighted the visceral veteranโs perspective on Iran. His unit lost men to โexplosively formed penetratorsโ (EFPs) planted by Iranian-backed militias, fostering a deep-seated animosity. General McChrystal affirmed this sentiment, recounting the American experience from the 1979 seizure of the Tehran embassy and the โdeath to Americaโ chants that followed, just years after Vietnam. For many Americans, Iran transformed overnight from a comfortable ally under the Shah to a recalcitrant, inexplicable enemy.
But McChrystal, ever the historian, stressed that this was only โpart of the story.โ He insisted on a longer view, pushing the narrative back to 1953 when the U.S. and British intelligence services orchestrated the overthrow of Iranโs constitutionally elected Prime Minister Mohammad Mossadegh, reinstating the Shah. The Shahโs regime, particularly through its brutal secret police, Savak, oppressed the Iranian people tremendously. โWhen the Iranian revolution erupts in 1978, we may be surprised. The Iranian people are not surprised,โ McChrystal observed. This historical context provides a crucial lens for understanding why chants of โdeath to Americaโ resonated so deeply within Iran.
The eight-year Iran-Iraq War, a brutal bloodletting twice as long as World War I, further seared itself into the Iranian national psyche. This experience, McChrystal explained, continues to shape the attitudes of Iranโs โbaby boomersโ and lends significant support to the clerical regime. The inclusion of Iran in George W. Bushโs โAxis of Evilโ in 2002 only solidified a continuous chain of grievances. โIf we donโt understand that journey to this point, we donโt understand the attitudes that are going to drive decisions people make,โ McChrystal emphasized, underscoring the deep-seated commitment that fuels the regime, far beyond any quick, external solution.
The Illusion of Quick Fixes: Americaโs Three Seductions
French raised the question of whether a rapid โdecapitation strikeโ could alter the Iranian regime, drawing parallels to a swift raid against Venezuelaโs Maduro. McChrystal quickly dismissed such notions, identifying what he calls the โthree great seductionsโ that consistently mislead American administrations and military leaders:
- Covert Action: The allure of a secret operation that will create a โgreat effectโ without anyone knowing who did it. McChrystalโs experience dictates otherwise: โIt never stays covert and it rarely works.โ Yet, its promise of an โeasy approach to a naughty problemโ remains seductive.
- The Surgical Special Operations Raid: Epitomized by the Maduro raid, these operations demonstrate โextraordinary competenceโ but rarely change the facts on the ground. โNot much changed,โ McChrystal noted, highlighting the disconnect between tactical brilliance and strategic impact.
- Air Power: From World War IIโs โdooh theoriesโ (the bomber will always get through) to Vietnamโs โreostatโ escalation strategy, the belief persists that bombing key targets will produce desired outcomes. However, as Vietnam painfully demonstrated, when an enemy is โasymmetrically committed to the outcomeโโlike the gravely wounded Shia fighters in Iraq who would try to bite their medicsโthere is โno pointโ at which they are willing to quit. Shock and awe in Iraq in 2003 was followed by a decade of fighting. McChrystalโs blunt assessment: โThe outcomes are in the minds of the people and unless youโre going to kill all the people, you may not affect that outcome.โ Iran, he concluded, has an โextraordinary capacity to be bombed.โ
When French countered with the argument that modern air power is โdifferentโ due to advanced drones and deep penetration capabilities, McChrystal, drawing on his investment experience, retorted, โI love that line: โThis time itโs different.โโ While acknowledging increased capability, he remains unconvinced it will be decisive. In Afghanistan, tribal members were โdisdainfulโ of bombing, demanding face-to-face combat to establish moral equivalence. People fight from โpassions,โ not geopolitical calculations. The idea that killing an enemy leaderโs family would make them negotiate, he said, would have โ100% oppositeโ effect on him.
The Inevitable Grind: Hormuz and the Human Cost
The strategic importance of the Strait of Hormuz is undeniable. French questioned the feasibility of forcing it open. McChrystalโs response was sobering: โIt would be hard to keep it open.โ He likened it to the Iraq War: easy to bomb, easy to take Baghdad, easy to remove a government. But changing the reality on the groundโwho controls things, how they workโis an entirely different challenge. โIf you like this war, enjoy this first part because this is the best part,โ McChrystal warned, โbecause everything after this will be harder because it will be more equal.โ
Iran doesnโt need to engage in direct combat with U.S. warships to exert influence. By deploying mines or autonomous vessels, or simply targeting civilian tankers โonce a week,โ they can create enough uncertainty to render the Strait uninsurable. The financial risk would become unacceptable, effectively closing the waterway without direct military confrontation.
Beyond economic and budgetary risks, French highlighted the often-overlooked human cost. While casualties have been few to date, McChrystal reminded that โevery casualty has a family and carries a loss.โ If the war drags on, especially with ground forces, casualties and frustration will inevitably rise. This leads to a critical societal concern: the civilian-military divide.
A Divided Nation, A Politicized Force?
The U.S. military, a volunteer force, is highly respected, but this respect can mask a dangerous divide. French worried that this creates a โsoldier castโ increasingly separate from the rest of society, potentially leading to โtoo great a willingness to use force.โ McChrystal agreed, calling it โnot healthy.โ While professional soldiers are often apolitical, thereโs an inherent โincentive for conflictโโit allows them to practice their craft, earn promotions, and increase defense budgets.
More concerning, McChrystal noted, is the potential for politization. In an environment where generals are โfired simply because they donโt fit in politically,โ the military risks aligning with a specific political leaning. โWhen I was in the service, you never knew what your peers felt politically. You never talked about it,โ he recalled. This neutrality is now โunder pressure,โ raising the specter of the military seeing itself as โthe guardians of the republic,โ a dangerous proposition for a democracy.
Grand Strategy and the โMowing the Lawnโ Fallacy
On the broader canvas of American grand strategy, McChrystal critiqued the โAmerica Firstโ approach. While economically understandable (tariffs, confrontation with China), it fundamentally misunderstands global security. True security, he argued, rests on โcredibility in the world,โ โalliances,โ โrelationships you can trust on,โ and the โrule of law writ large.โ President Trump, he contended, โweakened institutions,โ โchallenged norms,โ and โeliminated relationshipsโ under the misguided belief that it would advantage the U.S. as the โstrongest dog.โ This, McChrystal believes, has proven untrue.
He saw recent โadventurismโโfrom threatening Canada and Greenland to the Maduro raidโas emboldened by perceived cheap successes. The raid, McChrystal reiterated, was a classic example of a president being โseducedโ by the idea of doing something โon the cheap if youโre clever enough.โ The influence of Israel, particularly after October 7th, also played a significant role, with Prime Minister Netanyahuโs drive to โexpand Israelโs securityโ and โdo away with the boogeyman which was Iranโ shaping U.S. actions.
French brought up the pre-October 7th โmowing the lawnโ strategy, where periodic conflicts with groups like Hamas and Hezbollah were meant to โknock them back.โ October 7th, he argued, should have โblown up that idea,โ revealing Hamas as far from cowed and capable of horrific acts. This creates a โserious strategic dilemmaโ: an enormous capacity to damage enemies, but no real capacity to eliminate them. McChrystal concurred, warning that the โresentment you create through what you do now at some point comes back to you.โ Wars, he reminded, are rarely โneat, clean, or produce the kind of outcome we actually want. They produce this messy thing that might be better than before the war, but itโs not a lot better.โ
True Leadership: Beyond Bravado and Big Biceps
The conversation shifted to leadership, particularly the current atmosphere of bravado emanating from the top of the defense establishment. French noted that while some soldiers appreciate a fit, โhands-dirtyโ leader, bravado is often not appreciated in the โshow donโt tellโ military culture. McChrystal expressed his โdisappointment,โ recalling that the elite forces he commanded were never โbraggadociousโ or talked about โloving killing people.โ
He warned that such rhetoric can influence young, 18-year-old service members, fostering a dangerous sense of superiority. McChrystal also challenged the emphasis on physical prowess, noting that the vast majority of the modern force comprises โintelligence, communications, logistics, all the enablersโ rather than โbig biceps.โ He championed diversity and meritocracy, arguing that a healthy force embraces โolder men and women, young people, all this different because they had proven they were contributory to the fight.โ Acceptance, he stressed, should come from being โsmart,โ โcommitted,โ and a โgood colleague,โ not just physical strength.
McChrystal even took issue with the term โwarrior,โ preferring โsoldier.โ The difference between an army and a mob, he explained, is โdiscipline and leadership and the uniform code of military justice.โ This controlled, disciplined approach, rooted in values and culture, is essential when entrusting young people with the power to take life. โBig brains are more important than big biceps,โ he concluded, citing Ukraineโs innovative use of drone warfare as a prime example of an army constantly surprising the world with its resilience and ingenuity.
Bridging Divides: The Case for Mandatory National Service
To conclude, French challenged McChrystal on his advocacy for mandatory national service, a concept French, a believer in voluntary service, found difficult to reconcile with his libertarian leanings. McChrystal, who had once been talked โoff the ledgeโ from mandatory service himself, now firmly believes it is necessary.
He argued that people mature at different rates; a 17 or 18-year-old often makes different choices than a 36-year-old. Waiting for everyone to โarrive at the right answerโ is insufficient. Mandatory national service, offering a range of options beyond military conscription (like Teach for America or Peace Corps), would be a โgreat levelerโ in American society. It would create a shared experience, a common bond (โwhere did you serve?โ), bridging divides across different backgrounds and regions. โAll of us could use a period in our lives when weโre doing something thatโs inconvenient or maybe unpleasant or or that sort of thing,โ McChrystal asserted. โWe come out better for it.โ
General McChrystalโs insights paint a sobering picture of contemporary warfare and leadership. From the weight of history shaping Iranian resolve to the false promises of quick military solutions, and from the dangers of a politicized military to the need for a more unified society, his message is clear: the challenges ahead are profound, requiring not just strength, but deep understanding, humility, and a collective commitment to service. The era of easy victories, if it ever truly existed, is unequivocally over.
ํ๊ตญ์ด
โAlien contact will challenge every belief system on Earth | Sara Seagerโ โ Big Think ๊ธฐ๋ฐ ๊ธฐ์ฌ ์๋ณธ ์์ ๋ณด๊ธฐ
์ฐ์ฃผ ์๋ช ์ฒด ๋ฐ๊ฒฌ: ์ธ๋ฅ์ ์ ๋ ์ฒด๊ณ๋ฅผ ๋คํ๋ค ๋๊ฐ์ฑ ์๊ณ
์ฐ๋ฆฌ๊ฐ ์ฐ์ฃผ์์ ์ธ๊ณ ์๋ช ์ฒด๋ฅผ ๋ฐ๊ฒฌํ๋ค๋ฉด, ๊ทธ ์๊ฐ์ ๋จ์ํ ๊ณผํ์ ์ฑ๊ณผ๋ฅผ ๋์ด ์ธ๋ฅ์ ๋ชจ๋ ์ ๋ ์ฒด๊ณ๋ฅผ ๋คํ๋ค๊ณ ์กด์ฌ๋ก ์ ์ง๋ฌธ์ ๋์ง๋ ๋์ ํ์ ์ด ๋ ๊ฒ์ ๋๋ค. MIT์ ์ ๋ช ํ ํ์ฑ ๊ณผํ์ ์ฌ๋ผ ์๊ฑฐ(Sara Seager) ๊ต์๋ ๋น ์ฑํฌ(Big Think)์์ ์ธํฐ๋ทฐ์์ ์ธ๊ณ ์๋ช ์ฒด ๋ฐ๊ฒฌ์ด ๊ฐ์ ธ์ฌ ๊ณผํ์ , ์ฒ ํ์ , ์ฌํ์ ํ๊ธ๋ ฅ์ ๋ํด ์ฌ๋ ๊น์ ํต์ฐฐ์ ์ ์ํ์ต๋๋ค. ๊ทธ๋ ๋ ์ด ๋ฐ๊ฒฌ์ด ์ธ๋ฅ์๊ฒ โ๋๊ฐ์ฑ(awakening)โ์ ์๊ฐ์ด ๋ ๊ฒ์ด๋ผ๊ณ ์์ธกํ๋ฉฐ, ์ฐ๋ฆฌ๊ฐ ๊ณผ์ฐ ์ด ์ฅ๋ํ ์ฐ์ฃผ์ ์ง์ค์ ๋ฐ์๋ค์ผ ์ค๋น๊ฐ ๋์ด ์๋์ง ๋ฌป์ต๋๋ค.
์๋ช ์ฒด๋ ์ฐ์ฃผ์ ๋์ณ๋ ๊น? ๊ณผํ์ ์ฆ๊ฑฐ๋ค
์ฌ๋ผ ์๊ฑฐ ๊ต์๋ ๋ง์ฝ ์๋ช ์ด ๋ ๋ฒ, ์ฆ ๋ ๋ฆฝ์ ์ผ๋ก ๋ ๋ฒ ๋ฐ์ํ๋ค๋ ๊ฐ๋ ฅํ ์ฆ๊ฑฐ๋ฅผ ํ๋ณดํ๋ค๋ฉด, ๊ทธ ํ๊ธ๋ ฅ์ ์์ฒญ๋ ๊ฒ์ด๋ผ๊ณ ๊ฐ์กฐํฉ๋๋ค. ์ด๋ ์ฐ๋ฆฌ ์ํ์ ์ฐ์ฃผ์ ์๋ช ์ฒด๊ฐ ๋์ณ๋ ๊ฐ๋ฅ์ฑ์ด ๋งค์ฐ ๋๋ค๋ ์ค์ํ ์ ๋ณด๋ฅผ ์ป๊ฒ ๋ ๊ฒ์ด๊ธฐ ๋๋ฌธ์ ๋๋ค. ํ์ฌ ์ฐ๋ฆฌ๋ ์ด๋ฌํ ๊ฐ๋ฅ์ฑ์ ๋ท๋ฐ์นจํ๋ ์ฌ๋ฌ ๋จ์๋ค์ ํฌ์ฐฉํ๊ณ ์์ต๋๋ค.
- ํ์ฑ: ๋ณต์กํ ์ ๊ธฐ ๋ถ์๋ค์ด ๋ฐ๊ฒฌ๋์์ผ๋ฉฐ, ์งํ์๋ ์ก์ฒด ์ํ์ ๋ฌผ์ด ์กด์ฌํ ๊ฒ์ผ๋ก ์ถ์ ๋ฉ๋๋ค. ์ก์ฒด ์ํ์ ๋ฌผ์ ์๋ช ์ฒด ์กด์ฌ์ ํ์ ์กฐ๊ฑด ์ค ํ๋์ ๋๋ค.
- ๊ธ์ฑ: ๋๊ธฐ๊ถ์ ์๋ช ์ฒด๊ฐ ํ์๋ก ํ๋ ์ก์ฒด ํ๊ฒฝ์ด ์กด์ฌํจ์ ์๊ณ ์์ต๋๋ค.
- ๋ชฉ์ฑ๊ณผ ํ ์ฑ์ ์์ฑ๋ค: ์ ๋กํ(Europa), ์์ ๋ผ๋์ค(Enceladus) ๋ฑ ๋ง์ ์์ฑ์์ ๊ฑฐ๋ํ ์งํ ๋ฐ๋ค๊ฐ ์กด์ฌํจ์ด ํ์ธ๋์์ผ๋ฉฐ, ์ด๋ ์ธ๊ณ ์๋ช ์ฒด ํ์ฌ์ ๊ฐ์ฅ ์ ๋ ฅํ ํ๋ณด์ง๋ก ๊ผฝํ๋๋ค.
์ด๋ฌํ ๋ฐ๊ฒฌ๋ค์ ์ฐ์ฃผ ์ด๋๊ฐ์ ์ง์ ์๋ช ์ฒด๊ฐ ์กด์ฌํ ์ ์์์ ๊ฐ๋ ฅํ ์์ฌํ๋ฉฐ, ์๊ฑฐ ๊ต์๋ ์ธ๋ฅ๊ฐ ์ด ๊ฐ๋ฅ์ฑ๊ณผ ์ง์งํ๊ฒ ์จ๋ฆํด์ผ ํ ๊ฒ์ด๋ผ๊ณ ๋งํฉ๋๋ค.
์ ๋ ์ฒด๊ณ์ ๋์ : ์ฝํ๋ฅด๋์ฟ ์ค์ ๋์ ํ
ํ์ง๋ง ์ธ๊ณ ์๋ช ์ฒด์ ๋ฐ๊ฒฌ์ ๋จ์ํ ๊ณผํ์ ํธ๊ธฐ์ฌ์ ๋์ด ์ธ๋ฅ์ ๊ทผ๋ณธ์ ์ธ ์ ๋ ์ฒด๊ณ์ ๊ฑฐ๋ํ ๋์ ์ ์ ๊ธฐํ ๊ฒ์ ๋๋ค. ์๊ฑฐ ๊ต์๋ ์์ ์ ๊ฒฝํ๋ด์ ํตํด ์ด๋ฌํ ์ธ๋ฅ์ ์ ํญ๊ณผ ๋ณํ์ ๊ณผ์ ์ ์ค๋ช ํฉ๋๋ค.
๋ช ๋ ์ ์บ๋๋ค ์จํ๋ฆฌ์ค์ฃผ์ ์ฌ๋๋ฒ ์ด์์ ์๋ฆฝ์ฒ๋ฌธํํ ๊ธฐ์กฐ ์ฐ์ค์ ํ์ ๋์ ์ผ์ ๋๋ค. ์๊ฑฐ ๊ต์๋ ๋ ์คํ ๊ฐํจ๋ฆญ ์ ์์๋ โ์คํํค ์๋์ฐ(Spunky Widow)โ๋ผ๋ ๋ถ์ ์ง์ ๋จธ๋ฌผ๊ฒ ๋์๊ณ , ๊ทธ๋ ๋ ์๊ฑฐ ๊ต์๋ฅผ ์์ ์ ๊ฐํจ๋ฆญ ์น๊ตฌ๋ค์๊ฒ ์๊ฐํด์ฃผ๊ณ ์ถ์ด ํ์ต๋๋ค. ์คํํค ์๋์ฐ๋ ์น๊ตฌ๋ค์๊ฒ ์ฐ์ฃผ์ ๋ง์ ํ์ฑ์ด ์๋ค๊ณ ์ด์ผ๊ธฐํ์ง๋ง, ์น๊ตฌ๋ค์ ์ ๋ป ๋ฏฟ์ผ๋ ค ํ์ง ์์์ต๋๋ค. ์ด๋ ๊ทธ๋ค์ ๊ธฐ์กด ์ฌ๊ณ ์ฒด๊ณ์ ์ผ์ข ์ ๊ฐํนํ ํ์ค์ด์๊ธฐ ๋๋ฌธ์ ๋๋ค.
์๊ฑฐ ๊ต์๋ ์น๊ตฌ๋ค์๊ฒ ์ฐ์ฃผ์ ํ์ฑ์ด ์กด์ฌํ๋ค๋ ๊ฒ์ด ์ ๋นํ ๊ณผํ ๋ถ์ผ์ด๋ฉฐ, ์ค์ ๋ก ๋ง์ ์ฆ๊ฑฐ๊ฐ ์์์ ์ค๋ช ํด๋ฌ๋ผ๋ ์์ฒญ์ ๋ฐ์์ต๋๋ค. ์์ง ์๋ช ์ฒด๋ ์ง๊ตฌ์ ๊ฐ์ ํ์ฑ์ ๋ฐ๊ฒฌ๋์ง ์์์ง๋ง์. ์ด ์ค๋ช ์ ๊ทธ ๊ณต๋์ฒด์ ํฐ ์ํฅ์ ๋ฏธ์ณค์ต๋๋ค. ๊ทธ๋ค์ ๋ฐค๋ฆ๊ฒ๊น์ง ์ด์ผ๊ธฐ๋ฅผ ๋๋์๊ณ , ์ด๋ฅผ โ๋๊ฐ์ฑ(awakening)โ์ด๋ผ๊ณ ๋ถ๋ ์ต๋๋ค. ์ฐ์ฃผ์ ๋ค๋ฅธ ํ์ฑ๋ค์ด ์กด์ฌํ๋ค๋ ์ฌ์ค์ ๊นจ๋ฌ์ ๊ฒ์ด์ฃ . ์๊ฑฐ ๊ต์๋ ์๋ช ์ฒด์ ํ์ ์ ๋ฐ๊ฒฌํ๊ฒ ๋๋ฉด, ์ฐ๋ฆฌ์ ์ฌ๊ณ ๋ฐฉ์์ ์ฌ์ ๋ฆฝํด์ผ ํ๋ ํจ์ฌ ๋ ํฐ โ๋๊ฐ์ฑโ์ ๋ง์ดํ๊ฒ ๋ ๊ฒ์ด๋ผ๊ณ ์์ธกํฉ๋๋ค.
์๊ฑฐ ๊ต์๋ ์ด๋ฌํ ๋ณํ๋ฅผ ์ญ์ฌ ์ โ์ฝํ๋ฅด๋์ฟ ์ค์ ๋์ ํโ์ ๋น์ ํฉ๋๋ค. ์๋ฐฑ ๋ ์ , ์ง๊ตฌ๋ ์ฐ์ฃผ์ ์ค์ฌ์ด ์๋๋ฉฐ ํ์์ด ์ค์ฌ์ด๊ณ ์ง๊ตฌ๋ฅผ ๋น๋กฏํ ๋ชจ๋ ํ์ฑ์ด ํ์ ์ฃผ์๋ฅผ ๋๋ค๋ ์ฝํ๋ฅด๋์ฟ ์ค(Nicolaus Copernicus)์ ์ง๋์ค์ ๋น์ ์ข ๊ต์ ๊ด์ ์์ ์ด๋จ์ผ๋ก ์ฌ๊ฒจ์ก์ต๋๋ค. ๊ณผ์ฐ ๋ชจ๋ ์ฌ๋์ด โ์, ๋๋๋ค! ์ง๊ตฌ๊ฐ ํ์ ์ฃผ์๋ฅผ ๋๋ค๋ ์ ๋ง ์ ๋๋ ์ผ์ด์ผ!โ๋ผ๊ณ ์ธ์น๋ฉฐ ์ฆ๊ฐ์ ์ผ๋ก ๋์ํ์๊น์? ์๋๋๋ค. ๋ช ํํ ์ ํ์ ์ ์์์ผ๋ฉฐ, ์ด ๊ฐ๋ ์ด ๋ฐ์๋ค์ฌ์ง๋ ๋ฐ๋ ๋งค์ฐ ์ค๋ ์๊ฐ์ด ๊ฑธ๋ ธ์ต๋๋ค. ๊ทธ๋ฆฌ๊ณ ๋ง์นจ๋ด ๋ฐ์๋ค์ฌ์ก์ ๋๋ ์ฑ๋ํ ์ถํ๋ณด๋ค๋ ์ ์ง์ ์ผ๋ก ์ฐ๋ฆฌ ๋ฌธํ์ ์ผ๋ถ๊ฐ ๋์์ต๋๋ค.
๋ง์ฐฌ๊ฐ์ง๋ก ์ธ๊ณ ์๋ช ์ฒด์ ๋ช ํํ ํ์ ์ ๋ฐ๊ฒฌํ๋ ์ผ๋ ์๋ฐฑ ๋ ์ ๊ฑธ์ณ ์ ์ง์ ์ผ๋ก ์ฐ๋ฆฌ์ ๋ฌธํ์ ์ค๋ฉฐ๋ค๊ฒ ๋ ๊ฒ์ด๋ผ๊ณ ์๊ฑฐ ๊ต์๋ ์์ธกํฉ๋๋ค. ์ด๋ ์ฆ๊ฐ์ ์ธ ์ถฉ๊ฒฉ๋ณด๋ค๋ ์ค๋ ์๊ฐ์ ๊ฑธ์ณ ์ธ๋ฅ์ ์ธ๊ณ๊ด์ ์ฌํธํ๋ ๊ณผ์ ์ด ๋ ๊ฒ์ ๋๋ค.
์์ ๊ณผํ์ ๊ฐ์น: ์์ธก ๋ถ๊ฐ๋ฅํ ๋ฐ๊ฒฌ๋ค
์ด๋ค ์ฌ๋๋ค์ ์์ ๊ณผํ์ ์ ๊ทธ๋ ๊ฒ ๋ง์ ๋ ธ๋ ฅ์ ๊ธฐ์ธ์ฌ์ผ ํ๋์ง ์๋ฌธ์ ์ ๊ธฐํ๊ธฐ๋ ํฉ๋๋ค. ๋น์ฅ ์ค์ฉ์ ์ธ ๊ฒฐ๊ณผ๋ฌผ์ด ๋ณด์ด์ง ์์ ๋ ๋์ฑ ๊ทธ๋ ์ต๋๋ค. ํ์ง๋ง ์๊ฑฐ ๊ต์๋ ์์ ๊ณผํ์ด ์ฐ๋ฆฌ๋ฅผ ๋๋ผ์ด ๋ฐ๊ฒฌ์ผ๋ก ์ด๋์ด์์์ ๊ฐ์กฐํฉ๋๋ค.
- ์ธ์๋ฆฐ์ ๋ฐ๊ฒฌ: ์ง๋ณ ์น๋ฃ์ ํ๋ช ์ ์ธ ๋ณํ๋ฅผ ๊ฐ์ ธ์จ ์ธ์๋ฆฐ(Insulin)์ ์์ ์๋ฌผํ ์ฐ๊ตฌ์ ๊ฒฐ๊ณผ์์ต๋๋ค.
- ์๋ฃ ์์ ๊ธฐ์ : ์ฒ๋ฌธํ ๋ถ์ผ์์๋ ๋ฐคํ๋์ ๋ฐ์ดํฐ๋ฅผ ์ด๋ฏธ์ง๋ก ๋ณํํ๋ ๊ณผ์ ์์ ๋ฐ์ ํ ๊ธฐ์ ์ด ์ค๋๋ MRI๋ CT์ ๊ฐ์ ์๋ฃ ์์ ๊ธฐ์ ๋ฐ์ ์ ํฌ๊ฒ ๊ธฐ์ฌํ์ต๋๋ค.
- GPS: ์๊ฑฐ ๊ต์๊ฐ ๊ฐ์ฅ ์ข์ํ๋ ์์๋ GPS(Global Positioning System)์ ๋๋ค. ๋งค์ผ GPS๋ฅผ ์ฌ์ฉํ์ง ์๋ ์ฌ๋์ด ์์๊น์? ํ์ง๋ง ๋๊ตฐ๊ฐ๊ฐ โ์ฐ๋ฆฌ๊ฐ ์ด๋์ ์๊ณ ๋ค๋ฅธ ์ฌ๋๋ค์ ์ด๋์ ์๋์ง ์ ์ ์๋ ๋ฐฉ๋ฒ์ ๋ง๋ค์!โ๋ผ๊ณ ์ฒ์๋ถํฐ ๊ณํํ์๊น์? ์๋๋๋ค. ๋ก์ผ์ ๋ฐ์ฌํ์ฌ ๋ถ์ชฝ ๊ถค๋์ ์ง์ ํ ์ ์๋์ง ์คํํ๋ฉฐ ๋จ์ํ ๋ก์ผ ๊ธฐ์ ์ ๊ฐ์ง๊ณ โ์ฅ๋์น๋(fooling around)โ ์ฌ๋๋ค์ด ๋ง๋ค์ด๋ธ ๊ฒฐ๊ณผ๋ฌผ์ ๋๋ค.
์ธ๊ณ ์๋ช ์ฒด ํ์ฌ๊ฐ ์ด๋ค ์ค์ฉ์ ์ธ ๋ฐ๊ฒฌ์ผ๋ก ์ด์ด์ง์ง๋ ์์ง ๋ฏธ์ง์์ ๋๋ค. ํ์ง๋ง ์๊ฑฐ ๊ต์๋ ์ธ๋ฅ์ ๊ฑฐ๋ํ ๋์ฝ์ ๋๋ฌผ๊ฒ ์ผ์ด๋์ง๋ง, ์ฐ๋ฆฌ๋ ์๋์ ์ผ๋ก ๊ทธ๊ณณ์ ๋๋ฌํ ์ ์๋ค๊ณ ๋งํฉ๋๋ค. ๊ณผํ์ด๋ผ๋ โ๋ฌด์์์ ์ธ ์ฐ์ฑ (random walk)โ์ ํตํด, ์ฆ ์์ธก ๋ถ๊ฐ๋ฅํ ๋ฐ๊ฒฌ์ ๊ณผ์ ์ ํตํด ๋๋ฌํด์ผ ํฉ๋๋ค. ์์ ๊ณผํ์ ๋น์ฅ์ ์ด์ต์ ๋์ด ์ธ๋ฅ์ ์ง์๊ณผ ๊ธฐ์ ์ ํ์ฅํ๊ณ , ์๊ธฐ์น ์์ ๊ณณ์์ ์๋ํ ํ์ ์ ๊ฐ์ ธ์ค๋ ๊ทผ๋ณธ์ ์ธ ๋๋ ฅ์ธ ๊ฒ์ ๋๋ค.
์ธ๋ฅ๋ ์ค๋น๋์๋๊ฐ?
์ฌ๋ผ ์๊ฑฐ ๊ต์์ ํต์ฐฐ์ ์ธ๊ณ ์๋ช ์ฒด ํ์ฌ๊ฐ ๋จ์ํ ๊ณผํ์ ํ๊ตฌ๋ฅผ ๋์ด ์ธ๋ฅ์ ์กด์ฌ๋ก ์ ์ง๋ฌธ์ ๋์ง๋ ์ฌ์ ์์ ๋ณด์ฌ์ค๋๋ค. ์ฐ๋ฆฌ๊ฐ ์ฐ์ฃผ์์ ์๋ช ์ฒด๋ฅผ ๋ฐ๊ฒฌํ๋ ์๊ฐ์ ์ธ๋ฅ์ ๊ณผํ์ ์งํ์ ๋ํ๊ณ , ์ฐ๋ฆฌ์ ์ ๋ ์ฒด๊ณ๋ฅผ ์ฌ์ ๋ฆฝํ๋ฉฐ, ๊ถ๊ทน์ ์ผ๋ก๋ ์ฐ๋ฆฌ ์์ ์ ์ดํดํ๋ ๋ฐ ๊ฒฐ์ ์ ์ธ ์ ํ์ ์ด ๋ ๊ฒ์ ๋๋ค.
๊ทธ๊ฒ์ด ์๋ฐฑ ๋ ์ด ๊ฑธ๋ฆฌ๋ ์ ์ง์ ์ธ โ๋๊ฐ์ฑโ์ด๋ , ๊ฐ์์ค๋ฌ์ด ๊นจ๋ฌ์์ด๋ , ์ธ๋ฅ๋ ์ด ์ฅ๋ํ ์ฐ์ฃผ์ ์ง๋ฌธ์ ๋ํ ๋ต์ ์ฐพ์๊ฐ๋ ์ฌ์ ์์์ ๋์์์ด ์งํํ ๊ฒ์ ๋๋ค. ์ธ๊ณ ์๋ช ์ฒด ๋ฐ๊ฒฌ์ ๋จ์ง โ๋ฌด์์ ๋ฐ๊ฒฌํ๋๊ฐโ์ ๋ฌธ์ ๋ฅผ ๋์ด, โ์ฐ๋ฆฌ๊ฐ ๋๊ตฌ์ด๋ฉฐ ์ฐ์ฃผ์์ ์ด๋ค ์กด์ฌ์ธ๊ฐโ๋ผ๋ ๊ทผ๋ณธ์ ์ธ ์ง๋ฌธ์ ๋ํ ๋ต์ ์ฐพ์๊ฐ๋ ์ธ๋ฅ์ ์๋ํ ์ฌ์ ์ ์์์ด ๋ ๊ฒ์ ๋๋ค.
โFirecrawl AI clearly explained (and how to make $$)โ โ Greg Isenberg ๊ธฐ๋ฐ ๊ธฐ์ฌ ์๋ณธ ์์ ๋ณด๊ธฐ
AI์ โ๋๊ณผ ์โ Firecrawl: ์น ๋ฐ์ดํฐ๊ฐ AI ์๋์ ์๋ก์ด ๊ธ๋งฅ์ด ๋๋ ์ด์
์ธ๊ณต์ง๋ฅ(AI)์ ๋๋ผ์ธ ์ ๋๋ก ๋๋ํ์ง๋ง, ์น๋ช ์ ์ธ ์ฝ์ ์ ๊ฐ์ง๊ณ ์์ต๋๋ค. ๋ฐ๋ก โ์๊ฐโ์ด ์๋ค๋ ๊ฒ์ ๋๋ค. AI๋ ์ค์ค๋ก ์ธํฐ๋ท์ ํ์ํ๊ณ ์น์ฌ์ดํธ์์ ๋ฐ์ดํฐ๋ฅผ ์์งํ ์ ์์ต๋๋ค. ์ด๋ ๋ง์น ์ฒ์ฌ์ ์ธ ๋๋๋ฅผ ๊ฐ์ก์ง๋ง ๋์ด ๋ฉ๊ณ ์์ด ์๋ ๊ฒ๊ณผ ๊ฐ์ต๋๋ค. ํ์ง๋ง ์ด์ ์ด๋ฌํ ํ๊ณ๋ฅผ ๊ทน๋ณตํ๊ณ AI์ โ๋๊ณผ ์โ์ ๋ถ์ฌํ๋ ํ์ ์ ์ธ ๋๊ตฌ๊ฐ ๋ฑ์ฅํ์ต๋๋ค. ๋ฐ๋ก Firecrawl AI์ ๋๋ค.
์ด ๊ธ์์๋ Greg Isenberg ์ฑ๋์ Greg Isenberg๊ฐ Firecrawl AI๋ฅผ ๋ช ํํ๊ฒ ์ค๋ช ํ๊ณ , ์ด๋ฅผ ํ์ฉํด ์ค์ ์์ต์ ์ฐฝ์ถํ๋ ๋น์ฆ๋์ค๋ฅผ ๊ตฌ์ถํ๋ ๋ฐฉ๋ฒ์ ๋ํ ํต์ฐฐ์ ๊น์ด ์๊ฒ ๋ค๋ฃน๋๋ค. Firecrawl์ AI๊ฐ ์น์ โ๋ณด๊ณ โ โํ๋โํ๋ฉฐ ๋ฐ์ดํฐ๋ฅผ ์์งํ ์ ์๋๋ก ๋๋ ๋๊ตฌ๋ก, ์ ํ ๊ฐ๋ฐ ๋ฐฉ์๊ณผ ๋ฐ์ดํฐ ์์ง, ๊ทธ๋ฆฌ๊ณ AI๋ก ๊ฐ๋ฅํ ๊ฒ๋ค์ ๋ํ ์ฐ๋ฆฌ์ ์๊ฐ์ ๊ทผ๋ณธ์ ์ผ๋ก ๋ณํ์ํฌ ๊ฒ์ ๋๋ค.
1. AI, ๋ ์ด์ ๋ ๋จผ ์ฅ๋์ด ์๋๋ค: Firecrawl์ ๋ฑ์ฅ
ํ์ฌ AI ๋ชจ๋ธ, ํนํ ๋๊ท๋ชจ ์ธ์ด ๋ชจ๋ธ(LLM)์ ๋ฐฉ๋ํ ํ์ต ๋ฐ์ดํฐ๋ฅผ ๊ธฐ๋ฐ์ผ๋ก ๋๋ผ์ด ๋ฅ๋ ฅ์ ๋ณด์ฌ์ค๋๋ค. ํ์ง๋ง ์ด๋ค์ ์ต์ ์น ๋ฐ์ดํฐ๋ฅผ ์ค์๊ฐ์ผ๋ก ์ง์ ๊ฒ์ํ๊ณ ํ์ฉํ๋ ๋ฐ๋ ํ๊ณ๊ฐ ์์ต๋๋ค. ์ฆ, AI๋ ๋๋ํ์ง๋ง โ๋งน์ธ(blind)โ ์ํ์ ๋๋ค. ์น์ฌ์ดํธ์ ์ ์ํ๊ฑฐ๋ ๋ฐ์ดํฐ๋ฅผ ์ง์ ์ถ์ถํ ์ ์์ฃ . Firecrawl์ ๋ฐ๋ก ์ด ๋ฌธ์ ๋ฅผ ํด๊ฒฐํฉ๋๋ค.
Firecrawl์ AI์ โ๋๊ณผ ์โ์ ๋ถ์ฌํ์ฌ ์ธํฐ๋ท์ ๋ณผ ์ ์๊ฒ ํ๊ณ ๋ฐ์ดํฐ๋ฅผ ๊ฐ์ ธ์ฌ ์ ์๊ฒ ํฉ๋๋ค. ์ผ๋จ Firecrawl์ด ์๋ํ๋ ๋ชจ์ต์ ๋ณด๋ฉด, ์ ํ์ ๋ง๋ค๊ณ ๋ฐ์ดํฐ๋ฅผ ์์งํ๋ฉฐ AI๋ก ๋ฌด์์ด ๊ฐ๋ฅํ์ง์ ๋ํ ์๊ฐ์ด ์์ ํ ๋ฐ๋ ๊ฒ์ ๋๋ค. Greg Isenberg๋ Firecrawl์ ํ์ฉํ์ฌ ํธ๋ ๋์ ์ต๊ณ ์ ์คํํธ์ ์์ด๋์ด๋ฅผ ์ ๊ณตํ๋ ideabser.com์ ๊ตฌ์ถํ์ผ๋ฉฐ, ์ด ๊ฒฝํ์ ํตํด Firecrawl์ ์ ์ฌ๋ ฅ์ ์ง์ ํ์ธํ์ต๋๋ค.
2. AI ์๋์ ์งํ: ์ฑ๋ด์์ AI ์์ด์ ํธ๊น์ง
AI์ ๋ฐ์ ์ ํฌ๊ฒ ์ธ ์๋๋ก ๋๋ ์ ์์ต๋๋ค.
- ์ฑ๋ด ์๋ (2022๋ ): ChatGPT์ ๋ฑ์ฅ์ ์ง๋ฌธ์ ๋ตํ๋ AI์ ์ด๊ธฐ ๊ฐ๋ฅ์ฑ์ ๋ณด์ฌ์ฃผ์์ง๋ง, ๊ทธ ๊ธฐ๋ฅ์ ๋ค์ ์ ํ์ ์ด์์ต๋๋ค.
- ์ฝํ์ผ๋ฟ(Co-pilot) ์๋: Cursor๋ GitHub Co-pilot๊ณผ ๊ฐ์ ๋๊ตฌ๋ค์ด ๋ฑ์ฅํ๋ฉฐ ๊ฐ๋ฐ ๊ณผ์ ์ ๊ฐ์ํํ์ง๋ง, ์ฌ์ ํ ์ธ๊ฐ์ ์ฃผ๋์ ์ธ ์กฐ์์ด ํ์ํ์ต๋๋ค.
- AI ์์ด์ ํธ(Agent) ์๋ (ํ์ฌ): ์ด์ AI๊ฐ ์ค์ค๋ก ์์ ์ ์ํํ๋ ์๋๋ก ์ ์ด๋ค์์ต๋๋ค. Claude Code์ ๊ฐ์ AI๋ ์น์ ํ์ํ๊ณ , ์ฐ๊ตฌํ๊ณ , ์ฌ์ง์ด ์ฝ๋๋ฅผ ๊ตฌ์ถํ๊ธฐ๋ ํฉ๋๋ค. ํ์ง๋ง ์ด ๋ชจ๋ ๊ณผ์ ์์ AI๋ ์ฌ์ ํ โ๋ฐ์ดํฐโ๋ฅผ ํ์๋ก ํฉ๋๋ค.
์ด๋ฌํ ๋ณํ๋ โ์ปดํจํฐ ์ฌ์ฉ(Computer Use) ์๋โ๋ก ์ด์ด์ง๋๋ค. AI ์์ด์ ํธ๊ฐ ์ปดํจํฐ๋ฅผ ๋ณด๊ณ ์ ์ดํ ์ ์๊ฒ ๋ ๊ฒ์ด์ฃ . ๊ณผ๊ฑฐ์๋ ์ธ๊ฐ์ด ๋ง์ฐ์ค์ ํค๋ณด๋๋ฅผ ์ฌ์ฉํด ํด๋ฆญํ๊ณ ์์ ์ ์ํํ์ง๋ง, ์ด์ ๋ Perplexity, OpenAI์ Operator, Claude์ ์ปดํจํฐ ์ฌ์ฉ API์ ๊ฐ์ ๋๊ตฌ๋ค์ด AI๊ฐ ์น์ ํ์ํ๊ณ ์คํฌ๋ฆฐ์ท์ ์ฐ๊ณ ํด๋ฆญํ๋ฉฐ ์ฌ์ง์ด ๋ฐ์คํฌํฑ ์ ์ฒด๋ฅผ ์ ์ดํ ์ ์๊ฒ ํฉ๋๋ค. ์ด ๋ชจ๋ AI ์์ด์ ํธ๋ค์ด ๊ณตํต์ ์ผ๋ก ํ์๋ก ํ๋ ๊ฒ์ด ๋ฐ๋ก **โ๊นจ๋ํ ์น ๋ฐ์ดํฐ(Clean Web Data)โ**์ด๋ฉฐ, Firecrawl์ ์ด ๋ฐ์ดํฐ๋ฅผ ์ ๊ณตํ๋ ํต์ฌ์ ์ธ ์ญํ ์ ํฉ๋๋ค.
3. ์น ์คํฌ๋ํ์ ํจ๋ฌ๋ค์ ์ ํ: Firecrawl์ 6๊ฐ์ง ์ํผํ์
๊ณผ๊ฑฐ์ ์น ์คํฌ๋ํ์ ๊ณ ํต์ค๋ฌ์ด ์์ ์ด์์ต๋๋ค. ์น์ฌ์ดํธ๋ง๋ค ๋ง์ถคํ ์คํฌ๋ํผ๋ฅผ ์ฝ๋ฉํด์ผ ํ๊ณ , ํ๋ก์(proxy)์ ๋ธ๋ผ์ฐ์ ๋ฅผ ๊ด๋ฆฌํ๋ฉฐ, ๋ด ๊ฐ์ง ์์คํ ์ ๋์ํด์ผ ํ์ต๋๋ค. ๋ํ ์ง์ ๋ถํ HTML์ ์๋์ผ๋ก ํ์ฑํด์ผ ํ๊ณ , ์น์ฌ์ดํธ๊ฐ ๋ณ๊ฒฝ๋ ๋๋ง๋ค ์คํฌ๋ฆฝํธ๊ฐ ๊นจ์ง๋ ์ผ์ด ๋น์ผ๋น์ฌํ์ต๋๋ค. ์ด๋ ์์ฒญ๋ ๋ํต๊ฑฐ๋ฆฌ์์ฃ .
๊ทธ๋ฌ๋ Firecrawl์ ์ด๋ฌํ ๋ฌธ์ ๋ฅผ ๋จ ํ๋์ API ํธ์ถ๋ก ํด๊ฒฐํฉ๋๋ค. ๋ช ์ด ๋ง์ ๊นจ๋ํ ๋ฐ์ดํฐ๋ฅผ ๋ฐํํ๋ฉฐ, ๊ฑฐ์ ๋ชจ๋ ์น์ฌ์ดํธ์์ ์๋ํ๊ณ , AI๊ฐ ๋ ์ด์์ ๋ณ๊ฒฝ๊น์ง ์ฒ๋ฆฌํฉ๋๋ค. Firecrawl์ ๊ฐ๋ฐ์๋ค์ด AI ์ ํ์ ๊ตฌ์ถํ ๋ ํ์ํ โ์์ด์ ํธ ์คํ(Agent Stack)โ์ ํต์ฌ ๊ณ์ธต, ์ฆ **์น ๋ฐ์ดํฐ ๊ณ์ธต(Web Data Layer)**์ ๋ด๋นํฉ๋๋ค.
Firecrawl์ 6๊ฐ์ง ํต์ฌ ๊ธฐ๋ฅ(์ํผํ์)์ ๋ค์๊ณผ ๊ฐ์ต๋๋ค.
- ์คํฌ๋ํ(Scrape): ๋จ์ผ ์น ํ์ด์ง๋ฅผ ๊นจ๋ํ ๋งํฌ๋ค์ด(Markdown) ํ์์ผ๋ก ์ถ์ถํฉ๋๋ค. ์๋ฅผ ๋ค์ด, ํน์ ๋ธ๋ก๊ทธ ๊ฒ์๋ฌผ ํ๋๋ฅผ ๊ฐ์ ธ์ฌ ์ ์์ต๋๋ค.
- ํฌ๋กค๋ง(Crawl): ์ ์ฒด ์น์ฌ์ดํธ๋ฅผ ์๋์ผ๋ก ํฌ๋กค๋งํ์ฌ ๋ชจ๋ ๊ธฐ์ฌ๋ ํ์ด์ง์ ๋ฐ์ดํฐ๋ฅผ ์์งํฉ๋๋ค. cnn.com๊ณผ ๊ฐ์ ๋๋ฉ์ธ์ ๋ชจ๋ ๊ธฐ์ฌ๋ฅผ ๊ฐ์ ธ์ค๋ ์์ ๋๋ค.
- URL ๋งคํ(Map URLs): ๋๋ฉ์ธ ๋ด์ ๋ชจ๋ URL์ ์ฆ์ ๋งคํํ์ฌ ํด๋น URL์ ํฌํจ๋ ๋ ์ง, ์ ๋ชฉ ๋ฑ ํ๋ถํ ๋ฉํ๋ฐ์ดํฐ๋ฅผ ์ ๊ณตํฉ๋๋ค.
- ๊ฒ์(Search): ๊ตฌ๊ธ ๊ฒ์ ๊ฒฐ๊ณผ์ ํจ๊ป ํด๋น ํ์ด์ง์ ์ ์ฒด ์ฝํ ์ธ ๋ฅผ ๋จ์ผ API ํธ์ถ๋ก ๊ฐ์ ธ์ต๋๋ค.
- ์์ด์ ํธ(Agent): ์ํ๋ ๋ฐ์ดํฐ๋ฅผ ์ค๋ช ํ๋ฉด AI๊ฐ ์น์ ํ์ํ์ฌ ์ฐพ์์ค๋๋ค. ์๋ฅผ ๋ค์ด, โ์ฌ์ฐ์ค ํ๋ก๋ฆฌ๋ค์์ ํ์ ๋์ ์ฟ ๋ฐ ์๋น 50๊ณณ์ ์ฐพ์์คโ๋ผ๊ณ ์ง์ํ๋ฉด ๊ฐ์ฅ ๋ช ํํ ๋ฐ์ดํฐ๋ฅผ ์ ๊ณตํฉ๋๋ค.
- ๋ธ๋ผ์ฐ์ (Browser): AI๊ฐ ์ค์ ๋ธ๋ผ์ฐ์ ๋ฅผ ์ ์ดํ์ฌ ์์์ ์ฑ์ฐ๊ฑฐ๋, ๋ฒํผ์ ํด๋ฆญํ๊ณ , ๋ก๊ทธ์ธ ๋ฐ ์ธ์ฆ์ ์ฒ๋ฆฌํ๋ฉฐ, ํ์ด์ง๋ค์ด์ (pagination)์ ํ์ํ ์ ์์ต๋๋ค. AI๊ฐ ์น์ ํ์ํ๋ ๊ณผ์ ์ ์ค์๊ฐ์ผ๋ก ์ง์ผ๋ณผ ์๋ ์์ต๋๋ค.
์ด ๋ชจ๋ ๊ฐ๋ ฅํ ๊ธฐ๋ฅ์ ๋จ ์ธ ์ค์ ์ฝ๋๋ก ๊ตฌํ ๊ฐ๋ฅํฉ๋๋ค.
4. ์น ๋ฐ์ดํฐ์ AWS ์๊ฐ: Firecrawl์ด ์ฌ๋ ์๋ก์ด ๋น์ฆ๋์ค ๊ธฐํ
Firecrawl์ ๋ฑ์ฅ์ ๋ง์น 2006๋ AWS(Amazon Web Services)์ ํ์๊ณผ ๋น๊ฒฌ๋ ๋งํ ์ค๋ํ ๋ณํ์ ๋๋ค. 2006๋ ์ด์ ์๋ ์น ์ฑ์ ๊ตฌ์ถํ๋ ค๋ฉด ์์ฒ ๋ฌ๋ฌ๋ฅผ ๋ค์ฌ ์๋ฒ๋ฅผ ๊ตฌ๋งคํ๊ณ , ๋๊ณผ ์ผ์ด๋ธ์ ๊ด๋ฆฌํ๋ฉฐ, ์ฆ์ ๊ณ ์ฅ์ ์ง์ ์ฒ๋ฆฌํด์ผ ํ์ต๋๋ค. ํ์ง๋ง AWS๋ โ๋จ ํ ๋ฒ์ API ํธ์ถ๋ก ํด๋ผ์ฐ๋ ์๋ฒ๋ฅผ ์ฌ์ฉํ์ธ์โ๋ผ๊ณ ๋งํ๋ฉฐ ์ด๋ฌํ ๋ชจ๋ ๋ฒ๊ฑฐ๋ก์์ ์์ด์ต๋๋ค. ๋๋ถ์ ๊ฐ๋ฐ์๋ค์ ์๋ฒ ๊ด๋ฆฌ ๋์ ์ ํ ๊ฐ๋ฐ์๋ง ์ง์คํ ์ ์์๊ณ , ์ด๋ ์์กฐ ๋ฌ๋ฌ ๊ท๋ชจ์ ๊ธฐ์ ๋ค์ด ํ์ํ๋ ๊ธฐ๋ฐ์ด ๋์์ต๋๋ค.
๋ง์ฐฌ๊ฐ์ง๋ก 2026๋ , AI๊ฐ ์น ๋ฐ์ดํฐ๋ฅผ ์ฌ์ฉํ๊ฒ ํ๋ ค๋ฉด ์คํฌ๋ํผ๋ฅผ ๋ง๋ค๊ณ , ํ๋ก์๋ฅผ ๊ด๋ฆฌํ๊ณ , ๋ธ๋ผ์ฐ์ ๋ฅผ ๊ด๋ฆฌํ๋ฉฐ, ๋ณด์ ๋ฌธ์ ์ ์จ๋ฆํด์ผ ํ ๊ฒ์ ๋๋ค. ๊ทธ๋ฌ๋ Firecrawl์ โ๋จ ํ ๋ฒ์ API ํธ์ถ๋ก ๋ชจ๋ ๊ฒ์ ํด๊ฒฐํด ๋๋ฆฝ๋๋คโ๋ผ๊ณ ๋งํฉ๋๋ค. ์ด๋ ์์ฒญ๋ ๋ณํ์ด๋ฉฐ, Firecrawl ์์ ๊ตฌ์ถ๋๋ ํ์ฌ๋ค์ AWS ์์ ์ธ์์ง ํ์ฌ๋ค์ฒ๋ผ ๋ง๋ํ ๊ฐ์น๋ฅผ ์ฐฝ์ถํ ์ ์์ต๋๋ค. โ๊นจ๋ํ๊ณ ๊ตฌ์กฐํ๋ ๋ฐ์ดํฐ๋ ์๋ก์ด ์ค์ผ(new oil)์ ๋๋ค.โ ์ด ๋ฐ์ดํฐ๋ฅผ LLM๊ณผ ์ํํธ์จ์ด์ ํจ๊ณผ์ ์ผ๋ก ํตํฉํ๋ ๋ฐฉ๋ฒ์ ์ดํดํ๋ ์ฌ๋๋ค์ด ํฅํ 12๊ฐ์ ๋์ ๊ฐ์ฅ ๊ฐ์น ์๋ ์คํํธ์ ์ ๋ง๋ค ์ ์์ ๊ฒ์ด๋ฉฐ, ์ด๋ค์ ์ต์ 12๊ฐ์์ ์ ๋๋ฅผ ์ฐจ์งํ ๊ฒ์ ๋๋ค.
5. Firecrawl๋ก ๋ง๋๋ ์ฑ๊ณต ๋น์ฆ๋์ค ์์ด๋์ด: ํ์ ์์ฅ์ ๊ณต๋ตํ๋ผ
Firecrawl์ ํ์ฉํ๋ฉด ์์ต์์ ์์ญ์ต ๋ฌ๋ฌ ๊ท๋ชจ์ ๊ฑฐ๋ ๊ธฐ์ ์ ๋ชฉํ๋ก ํ๋ ๋์ , ์ฐ๊ฐ 100๋ง ๋ฌ๋ฌ์์ 5์ฒ๋ง ๋ฌ๋ฌ ๊ท๋ชจ์ ์์ ์ ์ธ ํ์ ๋น์ฆ๋์ค๋ฅผ ๊ตฌ์ถํ ์ ์์ต๋๋ค. ํต์ฌ์ **โ๋(Brain, LLM)โ, โ์ ๊ฒฝ๊ณ(Nervous System, ํ๋กํ ์ฝ)โ, ๊ทธ๋ฆฌ๊ณ โ๋๊ณผ ์(Eyes & Hands, Firecrawl)โ**์ ์กฐํฉ์ ๋๋ค. Firecrawl์ ๋ธ๋ผ์ฐ์ ์๋๋ฐ์ค ๊ธฐ๋ฅ์ AI๊ฐ ๋ณด์๋ ํ๊ฒฝ์์ ์น์ฌ์ดํธ์ ์์์ ์ฑ์ฐ๊ณ ๋ฒํผ์ ํด๋ฆญํ๋ฉฐ ๋ก๊ทธ์ธ ์ํ๋ฅผ ์ ์งํ๋ ๋ฑ ๋ณต์กํ ์ํธ์์ฉ์ ๊ฐ๋ฅํ๊ฒ ํฉ๋๋ค. ์ด๋ฌํ ๊ฐ๋ ฅํ โ๋๊ณผ ์โ์ ๊ฐ์ง๊ณ ์ด๋ค ๋น์ฆ๋์ค ์์ด๋์ด๋ฅผ ๊ตฌ์ํ ์ ์์๊น์?
๋ค์์ Firecrawl์ ํ์ฉํ ๊ตฌ์ฒด์ ์ธ ์ฌ์ ์์ด๋์ด๋ค์ ๋๋ค.
- ๊ฐ๊ฒฉ ๋ชจ๋ํฐ๋ง ์ํํธ์จ์ด:
- ๊ธฐ์กด ๋ฌธ์ : Precinct๋ Visual Ping ๊ฐ์ ๊ธฐ์กด ๋๊ตฌ๋ค์ ์ 200~1000๋ฌ๋ฌ๋ฅผ ๋ถ๊ณผํ๋ฉฐ ์ผ๋ฐ์ ์ธ ์ด์ปค๋จธ์ค ์ ํ์ ์ถ์ ํฉ๋๋ค.
- Firecrawl ์๋ฃจ์ : ์ฃผ๋ง ๋ง์ ์ค๋์ปค์ฆ ์ฌํ๋งค ๊ฐ๊ฒฉ ๋ชจ๋ํฐ๋ง ์ํํธ์จ์ด๋ฅผ ๊ตฌ์ถํฉ๋๋ค. StockX, Goat, eBay์ ์ค๋์ปค์ฆ ๊ฐ๊ฒฉ ๋ณ๋์ ์๋์ผ๋ก ๊ฐ์งํ์ฌ ์๋ฆผ์ ์ ๊ณตํ๊ณ , ์ 50๋ฌ๋ฌ ๋๋ 500๋ฌ๋ฌ์ ํ๋งคํฉ๋๋ค. ํน์ ์์งํ, ์นด๋, ์์ ํ ๋ฑ ์ํ๋ ํ์ ์์ฅ์ ์ ์ฉํ ์ ์์ต๋๋ค.
- SEO ๊ฐญ ํ์ธ๋ (SEO Gapfinder):
- ๊ธฐ์กด ๋ฌธ์ : Ahrefs๋ SEMrush ๊ฐ์ ๊ฑฐ๋ SEO ๋๊ตฌ๋ค์ ์ ์๋ฐฑ ๋ฌ๋ฌ๋ฅผ ๋ถ๊ณผํ๋ฉฐ SEO ์ ๋ฌธ๊ฐ์๊ฒ ๋ณต์กํ ๋์๋ณด๋๋ฅผ ์ ๊ณตํฉ๋๋ค.
- Firecrawl ์๋ฃจ์ : Firecrawl์ ์ด์ฉํด โ์น๊ณผ ์์ฌ ์ ์ฉ SEO ๊ฐ์ฌโ ๋๊ตฌ๋ฅผ ๋ง๋ญ๋๋ค. ๊ฒฝ์์ฌ ์น์ฌ์ดํธ์ GMB(Google My Business) ๋ชฉ๋ก์ ๋ถ์ํ์ฌ, โ๋น์ ์ 12๊ฐ ํค์๋์์, ๊ฒฝ์์ฌ๋ 47๊ฐ ํค์๋์์ ์์๊ฐ ๋์ต๋๋คโ์ ๊ฐ์ ์ํด๋ฆญ ๋ณด๊ณ ์๋ฅผ ์์ฑํ๊ณ , ๋ณด๊ณ ์๋น 500๋ฌ๋ฌ ๋๋ ์ 200๋ฌ๋ฌ์ ํ๋งคํฉ๋๋ค.
- ํ์ ์ฑ์ฉ ํ๋ซํผ:
- ๊ธฐ์กด ๋ฌธ์ : Indeed, Zillow ๊ฐ์ ๋๊ท๋ชจ ํ๋ซํผ์ ์์ญ์ต ๋ฌ๋ฌ์ ํฌ์๋ฅผ ๋ฐ์ ๋ชจ๋ ์ฌ๋์ ์ํ ์ผ๋ฐ์ ์ธ ๊ฒ์ ๊ธฐ๋ฅ์ ์ ๊ณตํ๋ฉฐ ๊ด๊ณ ๊ธฐ๋ฐ ๋ชจ๋ธ์ ์ฌ์ฉํฉ๋๋ค.
- Firecrawl ์๋ฃจ์ : โ์๊ฒฉ AI/ML ์ผ์๋ฆฌโ๋ง์ ์ํ ํ๋ซํผ์ ๋ง๋ญ๋๋ค. Firecrawl์ด 500๊ฐ ๊ธฐ์ ์ ์ฑ์ฉ ํ์ด์ง๋ฅผ ๋งค์ผ ๋ชจ๋ํฐ๋งํ๊ณ , AI๊ฐ ์ ํฉ๋ ์ ์(fit score)์ ๋ฐ๋ผ ํํฐ๋ง ๋ฐ ์์๋ฅผ ๋งค๊น๋๋ค. ์ 29๋ฌ๋ฌ์ ํ๋ฆฌ๋ฏธ์ ์๋ฆผ ์๋น์ค๋ฅผ ์ ๊ณตํฉ๋๋ค. ์ฌ๋๋ค์ 3์ต ๊ฐ์ ์ผ์๋ฆฌ ๋ชฉ๋ก๋ณด๋ค ์์ ์๊ฒ ์ค์ํ 50๊ฐ์ ์ผ์๋ฆฌ๋ฅผ ์ํฉ๋๋ค.
- AI ์ฐ๊ตฌ ๋ณด๊ณ ์:
- ๊ธฐ์กด ๋ฌธ์ : Consensus๋ Tavali ๊ฐ์ ๊ธฐ์กด ์ฐ๊ตฌ ๋ณด๊ณ ์ ์๋น์ค๋ ์ผ๋ฐ์ ์ธ ํ์ ์ฐ๊ตฌ๋ ๊ด๋ฒ์ํ ๋ถ์ผ๋ฅผ ๋ค๋ฃจ๋ฉฐ, ์ฌ์ฉ์๊ฐ ์ง์ ํ๋กฌํํธ๋ฅผ ์ ๋ ฅํด์ผ ํ๊ณ ์ ๋ฌธ์ฑ์ด ๋ถ์กฑํฉ๋๋ค.
- Firecrawl ์๋ฃจ์ : โํ์ ์ํธํํ ํ ํฐ ์ค์ฌ ๋ณด๊ณ ์โ๋ฅผ ์ ๊ณตํฉ๋๋ค. Firecrawl์ด ๊ด๋ จ ๋ ผ๋ฌธ, ํธ์ํฐ ๋ฐ ๊ธฐํ ์๋ฃ๋ฅผ ์ฝ๊ณ , AI๊ฐ ์๋์ผ๋ก ์ํ ์ ์์ ์์ฝ์ ์์ฑํฉ๋๋ค. ์ด๋ฅผ ๋ฒค์ฒ์บํผํ, ์ฌ๋ชจํ๋ ๋ฑ์ ์ 1000~5000๋ฌ๋ฌ์ ํ๋งคํฉ๋๋ค. VC๋ 50๋ง ๋ฌ๋ฌ์ ์๋ชป๋ ํฌ์๋ก๋ถํฐ ์์ ์ ๊ตฌํด์ค ๋ณด๊ณ ์์ ๊ธฐ๊บผ์ด 5000๋ฌ๋ฌ๋ฅผ ์ง๋ถํ ๊ฒ์ ๋๋ค.
- ์์ด์ ํธ ์ธ ์ด ๋ฐ์ค(Agent-in-a-Box):
- ๊ธฐ์กด ๋ฌธ์ : Harvey AI์ ๊ฐ์ ํ๋ซํผ์ ์ํฐํ๋ผ์ด์ฆ ๋์์ด๋ฉฐ, ์ปค์คํฐ๋ง์ด์ง์ ๋ช ๋ฌ์ด ๊ฑธ๋ฆฝ๋๋ค.
- Firecrawl ์๋ฃจ์ : โ๋ถ๋์ฐ ๋น๊ต ๋ณด๊ณ ์ ์์ด์ ํธโ๋ฅผ ๋ง๋ญ๋๋ค. Firecrawl์ด ๋งค๋ฌผ ๋ชฉ๋ก, ์ธ๊ธ ๊ธฐ๋ก, ํ๊ฐ์ฆ ๋ฑ์ ๊ฐ์ ธ์ค๊ณ , AI ์์ด์ ํธ๊ฐ 30์ด ๋ง์ ๋น๊ต ๋ณด๊ณ ์๋ฅผ ์์ฑํฉ๋๋ค. ์ด๋ฅผ ์ค๊ฐ์ธ์๊ฒ ์ 300๋ฌ๋ฌ์ ํ๋งคํฉ๋๋ค.
- ๋ฆฌ๋ทฐ ์ธํ
๋ฆฌ์ ์ค:
- ๊ธฐ์กด ๋ฌธ์ : Brand24๋ AppFollow ๊ฐ์ ์๋น์ค๋ ์์ ๋ฏธ๋์ด์ ๋ฆฌ๋ทฐ๋ฅผ ๊ด๋ฒ์ํ๊ฒ ๋ชจ๋ํฐ๋งํ๋ฉฐ, ๋ง์ผํ ํ์ ์ํ ์ผ๋ฐ์ ์ธ ๊ฐ์ฑ ๋ถ์ ๋์๋ณด๋๋ฅผ ์ ๊ณตํฉ๋๋ค.
- Firecrawl ์๋ฃจ์ : โ์๋ง์กด FBA ํ๋งค์ ๋ฆฌ๋ทฐ ํธ๋์ปคโ๋ฅผ ๊ตฌ์ถํฉ๋๋ค. Firecrawl์ด ๊ฒฝ์์ฌ์ ๋ฆฌ๋ทฐ๋ฅผ ๋งค์ผ ๋ชจ๋ํฐ๋งํ๊ณ , AI๊ฐ ํธ๋ ๋๋ฅผ ๋ถ์ํฉ๋๋ค(์: โ๋ฐฐํฐ๋ฆฌ ์๋ช ์ ๋ํ ๋ถ๋ง์ด 40% ์ฆ๊ฐํ์ต๋๋คโ). ์ด๋ฅผ ์๋ง์กด ํ๋งค์์๊ฒ ์ 99๋ฌ๋ฌ์ ํ๋งคํฉ๋๋ค. ์๋ง์กด ํ๋งค์๋ค์ ๊ฒฝ์์ฌ๋ณด๋ค ๋จผ์ ์ ํ์ ๋ฌธ์ ์ ์ ํ์ ํ๊ธฐ ์ํด ๊ธฐ๊บผ์ด ๋์ ์ง๋ถํ ๊ฒ์ ๋๋ค.
์ด๋ฌํ ์์ด๋์ด๋ค์ ๋ชจ๋ ํ์ ์์ฅ์ ๊ณต๋ตํ๊ณ , Firecrawl์ ํตํด ๊นจ๋ํ๊ณ ๊ตฌ์กฐํ๋ ๋ฐ์ดํฐ๋ฅผ ํจ์จ์ ์ผ๋ก ์์งํ๋ฉฐ, AI๋ฅผ ํ์ฉํ์ฌ ๊ฐ์น ์๋ ๊ฒฐ๊ณผ๋ฌผ์ ์ ๊ณตํ๋ ๊ฒ์ด ํต์ฌ์ ๋๋ค. Constellation Software์ ๊ฐ์ ๊ธฐ์ ์ด ์๋ฐฑ ๊ฐ์ ์์งํ ์ํํธ์จ์ด ํ์ฌ๋ค์ ์์ ํ๋ฉฐ 750์ต ๋ฌ๋ฌ ๊ท๋ชจ์ ๊ธฐ์ ์ด ๋ ๊ฒ์ฒ๋ผ, ํน์ ๊ณ ๊ฐ์ ์ํ ์๋ฒฝํ ๋๊ตฌ๋ฅผ ๋ง๋ค๋ฉด ํฐ ๊ธฐํ๊ฐ ์์ต๋๋ค.
6. Firecrawl ๋น์ฆ๋์ค ๊ตฌ์ถ ํ๋ ์์ํฌ
Firecrawl์ ํ์ฉํ์ฌ ์์ต์ฑ ์๋ ๋น์ฆ๋์ค๋ฅผ ๊ตฌ์ถํ๊ธฐ ์ํ 5๋จ๊ณ ํ๋ ์์ํฌ๋ ๋ค์๊ณผ ๊ฐ์ต๋๋ค.
- ํ์ ์์ฅ ์ ์ (Pick a niche): ์ด ์ฐ์ ์ ์ฌ๋๋ค์ด ์ด๋ค ๋ฐ์ดํฐ์ ๊ธฐ๊บผ์ด ๋์ ์ง๋ถํ ๊ฒ์ธ๊ฐ?
- ์คํฌ๋ํผ ๊ตฌ์ถ (Build the scraper): Firecrawl ์์ด์ ํธ, ๊ฐ๋จํ Python ์คํฌ๋ฆฝํธ, ๋๋ Claude Code๋ฅผ ํ์ฉํ์ฌ ์คํฌ๋ํผ๋ฅผ ๋ง๋ญ๋๋ค.
- ๋ฐ์ดํฐ ํจํค์ง (Package it): ๋ฐ์ดํฐ๋ฅผ CSV ํ์ผ, ๋์๋ณด๋, ์ฌ๋(Slack) ์๋ฆผ, ๋๋ API ํํ๋ก ์ ๊ณตํฉ๋๋ค.
- ๊ฒฐ๊ณผ๋ฌผ ํ๋งค (Sell the output): ๋จ์ํ ๋๊ตฌ๋ฅผ ํ๋งคํ๋ ๊ฒ์ด ์๋๋ผ, ๋ฐ์ดํฐ๋ฅผ ํ๋งคํฉ๋๋ค. ๊ณ ๊ฐ๋น ์ 500๋ฌ๋ฌ์์ 5000๋ฌ๋ฌ๋ฅผ ์ฒญ๊ตฌํ ์ ์์ต๋๋ค.
- ์๋ํ (Automate it): ์คํฌ๋ํผ๋ฅผ ์์ฝํ๊ณ ์๋ํํ์ฌ ์ ์๋ ๋์์๋ ์๋ํ๊ฒ ํฉ๋๋ค.
์ด๋ฌํ ํ๋ผ์ดํ (Flywheel) ๋ชจ๋ธ์ ์ด์ ๋ง ์์๋๊ณ ์์ผ๋ฉฐ, ๋ถ๋์ฐ ๊ฐ๊ฒฉ ๋ฐ์ดํฐ, SaaS ๊ฒฝ์์ฌ ๋ชจ๋ํฐ๋ง, ์ฑ์ฉ ์ ๋ณด ์ง๊ณ, ํนํ ๋ฐ ๋ฒ๋ฅ ๋ฌธ์, ์ธํ๋ฃจ์ธ์ ์ฐ๋ฝ์ฒ ๋ฐ์ดํฐ๋ฒ ์ด์ค, ์ ๋ถ ์ฐ๋ฝ์ฒ ์๋ฆผ, ์ด์ปค๋จธ์ค ๊ฐ๊ฒฉ ์ถ์ , ํ์ ์ฐ๊ตฌ ๋ฐ์ดํฐ ์ธํธ ๋ฑ ๋ค์ํ ๋ถ์ผ์์ ํ์ ์์ฅ์ ๋์ฑ ์ธ๋ถํํ์ฌ ์ ์ฉํ ์ ์์ต๋๋ค.
7. ๋ฏธ๋๋ฅผ ์ฟ๋ณด๋ค: AI ์์ด์ ํธ ์ฑ์ฉ ์๋
Firecrawl์ ์ค์ ๋ก โFirecrawl ์์ ํฌ๋ฆฌ์์ดํฐโ๋ผ๋ ์ง์ฑ ์ AI ์์ด์ ํธ๋ฅผ ์ฑ์ฉํ๊ฒ ๋ค๊ณ ๊ณต๊ณ ๋ฅผ ๋์ต๋๋ค. ์ด๋ค์ AI ์์ด์ ํธ๊ฐ ์์จ์ ์ผ๋ก ํธ๋ ๋ ๊ธฐ์ ๋ฐ ๋ชจ๋ธ์ ์ฐ๊ตฌํ๊ณ , ์ด ์ ๋ณด๋ฅผ ์ฌ์ฉํ์ฌ ๊ณ ํ์ง ์์ ์ ํ๋ฆฌ์ผ์ด์ ์ ๋ง๋ค๊ณ ํ ์คํธํ๋ฉฐ ๊ฐ์ ํ ์ ์๊ธฐ๋ฅผ ์ํ์ต๋๋ค.
์ด๋ ๋ฏธ๋์ AI ์์ด์ ํธ๊ฐ ์ฝํ ์ธ ํฌ๋ฆฌ์์ดํฐ, ๊ณ ๊ฐ ์ง์ ์์ด์ ํธ, ์ฃผ๋์ด ๊ฐ๋ฐ์ ์์ด์ ํธ ๋ฑ์ผ๋ก ๊ณ ์ฉ๋์ด ์ 5000๋ฌ๋ฌ ์๋น์ ์ ๋ฌด๋ฅผ ์ํํ๊ฒ ๋ ๊ฒ์ด๋ผ๋ ๊ฐ๋ฅ์ฑ์ ์์ฌํฉ๋๋ค. Firecrawl๊ณผ ๊ฐ์ ๋๊ตฌ๋ฅผ ์ฌ์ฉํ์ฌ ๊ธฐ์ ์ด ์ฑ์ฉํ๊ณ ์ถ์ด ํ ๋งํ AI ์์ด์ ํธ์ ์ ํ์ ๊ตฌ์ถํ๋ ๊ฒ์ด ๋ค์ ํฐ ๊ธฐํ๊ฐ ๋ ์ ์์ต๋๋ค.
Firecrawl์ AI์ ์น ๋ฐ์ดํฐ๋ฅผ ์ ๊ณตํ๋ ํต์ฌ์ ์ธ ์ธํ๋ผ๋ก์, ์น ๋ฐ์ดํฐ ๊ณ์ธต์ ์์ฒญ๋ ๋น์ฆ๋์ค ๊ธฐํ๋ฅผ ์ด์ด์ฃผ๊ณ ์์ต๋๋ค. AI๊ฐ ๋ ์ด์ ๋ ๋จผ ์ํ๋ก ์์ง ์๊ณ โ๋๊ณผ ์โ์ ๊ฐ์ง๊ณ ์ธํฐ๋ท์ ํ์ํ๋ฉฐ ๊ท์คํ ๋ฐ์ดํฐ๋ฅผ ์์งํ ์ ์๊ฒ ๋จ์ผ๋ก์จ, ์ฐ๋ฆฌ๋ ์ด์ ์๋ ์์ํ๊ธฐ ์ด๋ ค์ ๋ ํ์ ์ ์ธ ์ ํ๊ณผ ์๋น์ค๋ฅผ ๋ง๋ค ์ ์๊ฒ ๋ ๊ฒ์ ๋๋ค. ์ง๊ธ์ด์ผ๋ง๋ก Firecrawl์ ํ์ฉํ์ฌ AI ์๋์ ์๋ก์ด ๊ธ๋งฅ์ ์บ๋ผ ๋์ ๋๋ค.
โStop doing admin work manually. Let AI handle it.โ โ How I AI ๊ธฐ๋ฐ ๊ธฐ์ฌ ์๋ณธ ์์ ๋ณด๊ธฐ
ํด๋ผ์ฐ๋ ๊ด๋ฆฌ์ ๋ณต์ก์ฑ, AI๊ฐ ํ์ด๋ธ๋ค: ๋ฒ๊ฑฐ๋ก์ด ์์์ ์ ์ด์ ๊ทธ๋ง!
ํ๋ ํด๋ผ์ฐ๋ ํ๊ฒฝ์ ๊ธฐ์ ์ด์์ ํ์ ์์๊ฐ ๋์์ง๋ง, ๊ทธ ๋ณต์ก์ฑ์ IT ๊ด๋ฆฌ์๋ค์๊ฒ ๋์์๋ ๋์ ๊ณผ์ ๋ฅผ ์๊ฒจ์ค๋๋ค. ํนํ, ์๋ง์ ์๋น์ค์ ๋ฆฌ์์ค์ ๋ํ ์ ํํ ๊ถํ ์ค์ ๋ฐ ์ญํ ํ ๋น์ ์ ๋ฌธ ์ง์๊ณผ ์๋นํ ์๊ฐ์ ์๊ตฌํ๋ ๊ณ ๋ ์์ ์ ๋๋ค. ์๋ชป๋ ์ค์ ์ ๋ณด์ ์ทจ์ฝ์ ์ด๋ ์ด์ ๋นํจ์จ๋ก ์ด์ด์ง ์ ์์ด ๋์ฑ ์ ์ค์ ๊ธฐํด์ผ ํฉ๋๋ค. ํ์ง๋ง, ์ด๋ฐ ๋ฒ๊ฑฐ๋ก์ด ์๋ ์์ ์ ์ธ๊ณต์ง๋ฅ(AI)์๊ฒ ๋งก๊ธด๋ค๋ฉด ์ด๋จ๊น์? ์ฌ๊ธฐ, AI๊ฐ ํด๋ผ์ฐ๋ ๊ด๋ฆฌ์ ๋์ ๋ฅผ ์ด๋ป๊ฒ ํด๊ฒฐํ๊ณ , ์ฐ๋ฆฌ๊ฐ ํต์ฌ ์ ๋ฌด์ ์ง์คํ ์ ์๋๋ก ๋๋์ง์ ๋ํ ํฅ๋ฏธ๋ก์ด ์ฌ๋ก๊ฐ ์์ต๋๋ค.
ํด๋ผ์ฐ๋ ๊ถํ ์ค์ , ์ ๊ทธ๋ ๊ฒ ์ด๋ ค์ธ๊น?
ํด๋ผ์ฐ๋ ํ๊ฒฝ, ํนํ ๋ง์ดํฌ๋ก์ํํธ ์ ์ (Microsoft Azure)์ ๊ฐ์ ํ๋ซํผ์์๋ ์ฌ์ฉ์ ๋๋ ์๋น์ค์ ํ์ํ โ์ญํ (Role)โ๊ณผ โ๊ถํ(Permission)โ์ ์ ํํ๊ฒ ๋ถ์ฌํ๋ ๊ฒ์ด ๋งค์ฐ ์ค์ํฉ๋๋ค. ์๋ฅผ ๋ค์ด, ํน์ ์ง์์ด โAzure Document Intelligenceโ ๊ธฐ๋ฅ์ ์ฌ์ฉํด์ผ ํ ๋, ๊ทธ์๊ฒ ์ ํํ ์ด๋ค ์ญํ ์ด ํ์ํ์ง ํ์ ํ๋ ๊ฒ์ ๊ฒฐ์ฝ ์ฌ์ด ์ผ์ด ์๋๋๋ค. ์๋ง์ ๋ฌธ์์ ์ค์ ํญ๋ชฉ์ ์ผ์ผ์ด ์ฐพ์๋ณด๊ณ ์ ์ฉํ๋ ๊ณผ์ ์ ๋นํจ์จ์ ์ผ ๋ฟ๋ง ์๋๋ผ, ์ธ์ ์ค๋ฅ์ ๊ฐ๋ฅ์ฑ๋ ๋์ต๋๋ค.
์๋ฌธ์ ํ์ ๋ํ โ์ด๋ค ์ญํ ์ด ํ์ํ์ง ์ ํ ๊ฐ์ด ์ค์ง ์์ ๋๊ฐ ๋ง์ต๋๋คโ๋ผ๊ณ ๊ณ ์ถฉ์ ํ ๋กํ๋ฉฐ, ๋๋ก๋ โ์ด ์ฌ๋์๊ฒ Azure Document Intelligence๋ฅผ ์ฌ์ฉํ๋๋ฐ ํ์ํ ์ญํ ์ ๋น์ ์ด ์์์ ์ฐพ์์ ๋ถ์ฌํ๋ผโ๋ ์์ผ๋ก ์์ ์ ๋งก๊ธฐ๊ธฐ๋ ํ๋ค๊ณ ์ธ๊ธํฉ๋๋ค. ์ด๋ ํด๋ผ์ฐ๋ ๊ด๋ฆฌ์๋ค์ด ๊ณตํต์ ์ผ๋ก ๊ฒช๋ ์ด๋ ค์์ผ๋ก, ์ ํํ ์ ๋ณด ๋ถ์กฑ๊ณผ ๋ฐฉ๋ํ ๋ฌธ์ ์์์ ํค๋งค๋ ์๊ฐ ๋ญ๋น๊ฐ ๋ฐ๋ณต๋๋ ์ ํ์ ์ธ ์ํฉ์ ๋๋ค.
AI ๋น์, ํด๋ต์ ์ ์ํ๋ค
์ด๋ฌํ ์ํฉ์์ AI๋ ๊ฐ๋ ฅํ ํด๊ฒฐ์ฑ ์ ์ ์ํฉ๋๋ค. ์๋์ผ๋ก ๋ฌธ์๋ฅผ ๋ค์ง๋ ๋์ , AI๋ ๋ง์น ์ ๋ด ๋น์์ฒ๋ผ ํ์ํ ์ ๋ณด๋ฅผ ์ฆ์ ์ฐพ์์ค๋๋ค. ์์ปจ๋, ํ์๊ฐ ์ฌ์ฉํ๋ AI ๋๊ตฌ์ธ โWarp(์ํ)โ๋ ๋ง์ดํฌ๋ก์ํํธ ๊ณต์ ๋ฌธ์ ์๋ฒ(Microsoft documentation MCP server)์ ์ง์ ์ฐ๊ฒฐ๋์ด, ํน์ ์์ ์ ํ์ํ ์ญํ ์ ์๋์ผ๋ก ์กฐํํ๊ณ ์ ์ฉํ ์ ์์ต๋๋ค.
์ฌ์ฉ์๊ฐ โ์ด ์ฌ๋์๊ฒ Azure Document Intelligence๋ฅผ ์ฌ์ฉํ ์ ์๋ ์ญํ ์ ๋ถ์ฌํ๋ผโ๊ณ ์ง์ํ๋ฉด, Warp๋ ๋ฐฉ๋ํ ๊ณต์ ๋ฌธ์๋ฅผ ๋ถ์ํ์ฌ ๊ฐ์ฅ ์ ํฉํ ์ญํ ์ ์ฐพ์ ์ ์ํ๊ฑฐ๋ ์ง์ ํ ๋นํ๋ ๋ฐฉ์์ ๋๋ค. ์ด๋ ๋ง์น ์๋ จ๋ ์ ๋ฌธ๊ฐ๊ฐ ์์์ ์กฐ์ธํด์ฃผ๋ ๊ฒ๊ณผ ๊ฐ์, ์์ ์ ์ ํ์ฑ๊ณผ ์๋๋ฅผ ๋น์ฝ์ ์ผ๋ก ํฅ์์ํต๋๋ค. ๋ ์ด์ ๋ณต์กํ ๋ฌธ์ ์์์ ํค๋งค๊ฑฐ๋ ์๋ชป๋ ์ญํ ์ ๋ถ์ฌํ ๊น ๊ฑฑ์ ํ ํ์๊ฐ ์์ต๋๋ค. AI๊ฐ ์ ํํ๊ณ ์ ์ํ๊ฒ ์ต์ ์ ์๋ฃจ์ ์ ์ ๊ณตํ๋ ๊ฒ์ ๋๋ค.
๋จ์ ์๋ํ๋ฅผ ๋์ด์ โ์ค๋งํธโํ ์ง์
Warp์ ์ง๋ฅ์ ์ธ ๊ธฐ๋ฅ์ ์ฌ๊ธฐ์ ๊ทธ์น์ง ์์ต๋๋ค. ๋จ์ํ ์ ๋ณด ๊ฒ์์ ๋์ด, ์ฌ์ฉ์๊ฐ ์ ์ํ โ๊ท์น(Rules)โ์ ๊ธฐ๋ฐ์ผ๋ก ๋์ฑ ์ ๊ตํ ์์ ์ ์ํํ๊ณ , ์ฌ์ง์ด ์ ์ฌ์ ์ธ ๋ฌธ์ ๊น์ง ๋ฏธ๋ฆฌ ๊ฐ์งํ์ฌ ์๋ ค์ค๋๋ค.
์๋ฅผ ๋ค์ด, ํด๋ผ์ฐ๋ ๋ฆฌ์์ค ๊ทธ๋ฃน(Resource Group)์ ์ญํ ์ ํ ๋นํ ๋, โ์์ ์ ์ก์ธ์ค(Owner Access)โ ๊ถํ์ด ๋จผ์ ํ์ฑํ๋์ด ์์ด์ผ ํ๋ค๋ ํน์ ๊ท์น์ด ์๋ค๊ณ ๊ฐ์ ํด ๋ด ์๋ค. ์ด๋ ํด๋ผ์ฐ๋ ๊ด๋ฆฌ์์ ํํ ๋ฐ์ํ๋ โ๊ฑธ๋ฆผ๋โ ์ค ํ๋์ ๋๋ค. ๋ง์ ๊ด๋ฆฌ์๊ฐ ์ด ๋จ๊ณ๋ฅผ ์์ด๋ฒ๋ ค ์์ ์ด ์คํจํ๋ ๊ฒฝ์ฐ๊ฐ ๋น๋ฒํฉ๋๋ค. Warp๋ ์ด๋ฌํ ๊ท์น์ ํ์ตํ๊ณ ๊ธฐ์ตํ์ฌ, ์์ ์ ์์ํ๊ธฐ ์ ์ ์ฌ์ฉ์์๊ฒ ์ ์ ์ ์ผ๋ก ์๋ฆผ์ ๋ณด๋ ๋๋ค. ์๋ฌธ ํ์๋ Warp๊ฐ ์์ ์๊ฒ **โ์ ๊น๋ง์, ์ด ์์ ์ ์์ํ๊ธฐ ์ ์ ์์ ์ ์ก์ธ์ค๋ฅผ ํ์ฑํํ์ จ๋์? ๊ทธ๋ ์ง ์์ผ๋ฉด ์คํจํ ๊ฒ๋๋คโ**๋ผ๊ณ ๋ฌป๋๋ค๊ณ ์ค๋ช ํฉ๋๋ค.
์ด์ฒ๋ผ AI๋ ๋จ์ํ ์๋ํ๋ฅผ ๋์ด, ์ฌ์ฉ์์ ์ค์๋ฅผ ์๋ฐฉํ๊ณ ์์ ํ๋ฆ์ ์ต์ ํํ๋ โ์ค๋งํธ ๋น์โ ์ญํ ์ ํกํกํ ํด๋ ๋๋ค. ๋ณต์กํ ํด๋ผ์ฐ๋ ํ๊ฒฝ์์ ๋ฐ์ํ ์ ์๋ ์ ์ฌ์ ๋ฌธ์ ๋ฅผ ๋ฏธ๋ฆฌ ํ์ ํ๊ณ ์ ์ ํ ์กฐ์น๋ฅผ ์๋ดํจ์ผ๋ก์จ, ์์ ์คํจ์จ์ ์ค์ด๊ณ ์ ๋ฐ์ ์ธ ์ด์ ํจ์จ์ฑ์ ํฌ๊ฒ ํฅ์์ํต๋๋ค.
๊ด๋ฆฌ ์ ๋ฌด๋ฅผ ๋์ด์ AI์ ํ์ฅ ๊ฐ๋ฅ์ฑ
์ด๋ฌํ AI ๊ธฐ๋ฐ์ ์ ๊ทผ ๋ฐฉ์์ ํด๋ผ์ฐ๋ ๊ด๋ฆฌ๋ฟ๋ง ์๋๋ผ ๋ค์ํ IT ๋ถ์ผ์์ ํ์ฉ๋ ์ ์์ต๋๋ค. ํนํ ์ฝ๋ฉ(Coding) ์์ ์ ํ์ํ ๋ผ์ด๋ธ๋ฌ๋ฆฌ๋ API(Application Programming Interface) ๋ฌธ์ ๊ฒ์, ํน์ ํ๋ ์์ํฌ ์ฌ์ฉ๋ฒ ๋ฑ ๋ณต์กํ๊ณ ๋ฐ๋ณต์ ์ธ ์ ๋ณด ํ์ ๊ณผ์ ์ AI๋ฅผ ์ ์ฉํ๋ฉด ๊ฐ๋ฐ ์์ฐ์ฑ์ ํฌ๊ฒ ๋์ผ ์ ์์ต๋๋ค. AI๋ ๋ฐฉ๋ํ ๋ฐ์ดํฐ๋ฅผ ๋ถ์ํ๊ณ ํจํด์ ์ธ์ํ๋ ๋ฅ๋ ฅ์ ๋ฐํ์ผ๋ก, ์ธ๊ฐ์ด ์๊ฐ๊ณผ ๋ ธ๋ ฅ์ ๋ค์ฌ์ผ ํ๋ ๋ง์ ์์ ์ ํจ์จ์ ์ผ๋ก ๋์ฒดํ๋ฉฐ, ๊ถ๊ทน์ ์ผ๋ก ์ธ๊ฐ์ด ๋ ์ฐฝ์์ ์ด๊ณ ์ ๋ต์ ์ธ ์ ๋ฌด์ ์ง์คํ ์ ์๋ ํ๊ฒฝ์ ์กฐ์ฑํฉ๋๋ค.
๊ฒฐ๋ก : AI์ ํจ๊ปํ๋ ์ค๋งํธํ ๋ฏธ๋
๊ฒฐ๋ก ์ ์ผ๋ก, AI๋ ๋ ์ด์ ๋ฏธ๋ ๊ธฐ์ ์ด ์๋ ํ์ฌ์ ๊ฐ๋ ฅํ ์ ๋ฌด ๋๊ตฌ์ ๋๋ค. ํด๋ผ์ฐ๋ ๊ถํ ์ค์ ๊ณผ ๊ฐ์ ๋ณต์กํ๊ณ ์ค๋ฅ ๋ฐ์ ๊ฐ๋ฅ์ฑ์ด ๋์ ๊ด๋ฆฌ ์ ๋ฌด๋ฅผ AI์ ์์ํจ์ผ๋ก์จ, ๊ธฐ์ ์ ์ด์ ํจ์จ์ฑ์ ๊ทน๋ํํ๊ณ ๋ณด์์ ๊ฐํํ๋ฉฐ, ๊ท์คํ ์ธ์ ์์์ ํต์ฌ ๋น์ฆ๋์ค ๊ฐ์น ์ฐฝ์ถ์ ์ง์ค์ํฌ ์ ์์ต๋๋ค. ๋ฒ๊ฑฐ๋ก์ด ์์์ ์์ ๋ฒ์ด๋, AI์ ํ๋ ฅํ์ฌ ๋์ฑ ์ค๋งํธํ๊ณ ์์ฐ์ ์ธ ๋ฏธ๋๋ฅผ ๋ง๋ค์ด๊ฐ๋ ์๋๊ฐ ๋๋ํ์ต๋๋ค. ์ด์ ์ฐ๋ฆฌ์๊ฒ ํ์ํ ๊ฒ์ AI์ ์ ์ฌ๋ ฅ์ ์ดํดํ๊ณ , ์ด๋ฅผ ์ฐ๋ฆฌ์ ์ ๋ฌด์ ์ ๊ทน์ ์ผ๋ก ํตํฉํ๋ ค๋ ์ ์ฐํ ์ฌ๊ณ ๋ฐฉ์์ผ ๊ฒ์ ๋๋ค.
โโEverything After This Will Be Harderโ: General Stanley McChrystal on Iranโ โ New York Times Podcasts ๊ธฐ๋ฐ ๊ธฐ์ฌ ์๋ณธ ์์ ๋ณด๊ธฐ
๋งฅํฌ๋ฆฌ์คํธ ์ฅ๊ตฐ์ด ๊ฒฝ๊ณ ํ๋ ์ด๋๊ณผ์ ์ ์: โ์ง๊ธ ์ดํ๋ ๋ชจ๋ ๊ฒ์ด ๋ ์ด๋ ค์ธ ๊ฒ์ด๋คโ
์ต๊ทผ ์ค๋ ์ง์ญ์ ๊ธด์ฅ์ด ๊ณ ์กฐ๋๋ฉด์ ์ด๋๊ณผ์ ์ ์ฌ์ ์ถฉ๋ ๊ฐ๋ฅ์ฑ์ด ์ ์ธ๊ณ์ ์ด๋ชฉ์ ์ง์ค์ํค๊ณ ์์ต๋๋ค. ๋ด์ํ์์ค ํ์บ์คํธ โThe Opinionsโ์์๋ ๋ฒ ํ ๋ ์นผ๋ผ๋์คํธ ๋ฐ์ด๋น๋ ํ๋ ์น(David French)๊ฐ ์ํ๊ฐ๋์คํ ์ฃผ๋ ๋ฏธ๊ตฐ ์ฌ๋ น๊ด์ ์ญ์ํ๊ณ ํฉ๋ํน์์์ ์ฌ๋ น๋ถ(Joint Special Operations Command, JSOC)๋ฅผ ์ด๋์๋ ์คํ ๋ฆฌ ๋งฅํฌ๋ฆฌ์คํธ(Stanley McChrystal) ์๋น์ญ ๋์ฅ๊ณผ ์ฌ์ธต ๋ํ๋ฅผ ๋๋ด์ต๋๋ค. ๋ก๋ฒํธ ๊ฒ์ด์ธ (Robert Gates) ์ ๊ตญ๋ฐฉ์ฅ๊ด์ผ๋ก๋ถํฐ โ๋ด๊ฐ ๋ง๋ ์ต๊ณ ์ ์ ์ฌ์ด์ ๋ฆฌ๋โ๋ผ๋ ์ฐฌ์ฌ๋ฅผ ๋ฐ์ ๋งฅํฌ๋ฆฌ์คํธ ์ฅ๊ตฐ์ ์ด๋๊ณผ์ ๊ฐ๋ฑ์ ๊ธฐ์, ๋ฏธ๊ตญ์ด ์ง๋ฉดํ ๋์ , ์ ์ ๋ณธ์ง, ๋ฆฌ๋์ญ, ๊ทธ๋ฆฌ๊ณ ๊ตญ๊ฐ ๋ด์ฌ์ ์ค์์ฑ์ ๋ํ ๊น์ด ์๋ ํต์ฐฐ์ ์ ๊ณตํ์ต๋๋ค.
์ํ ์ญ์ฌ, ๋ฟ๋ฆฌ ๊น์ ๊ฐ๋ฑ์ ์์
ํ๋ ์น ์นผ๋ผ๋์คํธ๋ ์ด๋๊ณผ์ ์ต๊ทผ ๊ฐ๋ฑ์ด ํฐ์ง ์ดํ, ์ฐธ์ ์ฉ์ฌ๋ค์ ์๊ฐ์ด ๋น์ฐธ์ ์ธ์ฌ๋ค์ ์๊ฐ๊ณผ ๋ค๋ฅด๋ค๋ ์ ์ ์ง์ ํ๋ฉฐ ๋ํ๋ฅผ ์์ํ์ต๋๋ค. ํนํ ์ด๋ผํฌ์ ์ฐธ์ ์ฉ์ฌ๋ค์ ์ด๋์ด ์ง์ํ๋ ๋ฏผ๋ณ๋๊ฐ ์ค์นํ ํญ๋ฐ์ฑ ๊ดํต์(Explosively Formed Penetrator, EFP)๋ก ์ธํด ๋๋ฃ๋ค์ ์์ ์ํ ๊ธฐ์ต์ ๊ฐ์ง๊ณ ์์ต๋๋ค. ์ด๋ ์ด๋์ ๋ํ ๊น์ ๊ฐ์ ์ ๋ฐ๊ฐ์ผ๋ก ์ด์ด์ง๋๋ค.
๋งฅํฌ๋ฆฌ์คํธ ์ฅ๊ตฐ์ ๋ฏธ๊ตญ์ ์ด๋ ๊ฒฝํ์ 1979๋ ํ ํค๋ ๋ฏธ๊ตญ ๋์ฌ๊ด ์ธ์ง ์ฌ๊ฑด๊ณผ โ๋ฏธ๊ตญ์ ์ฃฝ์์โ์ด๋ผ๋ ๊ตฌํธ๋ก ๊ฑฐ์ฌ๋ฌ ์ฌ๋ผ๊ฐ ์ค๋ช ํ์ต๋๋ค. ๋ฒ ํธ๋จ์ ์ดํ ์ ์์ ์ผ๋ก ์ทจ์ฝํ๋ ๋ฏธ๊ตญ์ธ๋ค์๊ฒ ๊ณผ๊ฑฐ ๋๋งน๊ตญ์ด์๋ ์ด๋์ ์ ๋์ ์ธ ํ๋๋ ํฐ ์ถฉ๊ฒฉ์ด์์ต๋๋ค. ์ดํ ์ด๋-์ด๋ผํฌ ์ ์์ ์ง์ผ๋ณด๋ฉฐ ๋ง์ ๋ฏธ๊ตญ์ธ๋ค์ โ์ด๋์ ์๋ํด์ฃผ๋ ๋๊ตฐ๊ฐ๊ฐ ์๊ฒผ๋คโ๋ฉฐ ๊ธ์ ์ ์ผ๋ก ๋ฐ๋ผ๋ณด๊ธฐ๋ ํ์ต๋๋ค. 1988๋ ์๋ ๋ฏธ ํด๊ตฐ ๊ตฌ์ถํจ ๋น์ผ์คํธ(USS Vincennes)๊ฐ ์ด๋ ์ฌ๊ฐ๊ธฐ๋ฅผ ์ ํฌ๊ธฐ๋ก ์ค์ธํด ๊ฒฉ์ถ, 290๋ช ์ ๋ฏผ๊ฐ์ธ์ด ์ฌ๋งํ๋ ๋น๊ทน๋ ์์์ต๋๋ค. ์ด ์๊ธฐ ๋ฏธ๊ตญ์ธ๋ค์๊ฒ ์ด๋์ โ์ดํดํ ์ ์๋ ์ด์ ๋ก ์ฐ๋ฆฌ๋ฅผ ๋ฏธ์ํ๋ ์๊ฐํ ์ โ์ผ๋ก ๊ฐ์ธ๋์์ต๋๋ค. 2007๋ ์ด๋ผํฌ์์๋ ์ด๋์ ์ง์์ ๋ฐ๋ ์์ํ ๋ฏผ๋ณ๋์์ ์น์ดํ ์ ํฌ๋ฅผ ๊ฒช์ผ๋ฉฐ, ๋งฅํฌ๋ฆฌ์คํธ ์ฅ๊ตฐ๊ณผ ๊ทธ์ ๋ถ๋์๋ค์๊ฒ ์ด๋์ ๋ช ๋ฐฑํ ์ ์ด์์ต๋๋ค.
ํ์ง๋ง ๋งฅํฌ๋ฆฌ์คํธ ์ฅ๊ตฐ์ ์ด๊ฒ์ด โ์ด์ผ๊ธฐ์ ์ผ๋ถ์ผ ๋ฟโ์ด๋ผ๊ณ ๊ฐ์กฐํฉ๋๋ค. ๊ทธ๋ ๋ฏธ๊ตญ์ ๋จ๊ธฐ์ ์ธ ์๊ฐ์ ๋นํํ๋ฉฐ, ์ด๋๊ณผ์ ๊ฐ๋ฑ์ ๋ฟ๋ฆฌ๊ฐ ํจ์ฌ ๋ ๊น๋ค๊ณ ์ง์ ํ์ต๋๋ค. ์ง์ง ์์์ 1953๋ ๋ฏธ๊ตญ๊ณผ ์๊ตญ ์ ๋ณด๊ธฐ๊ด์ด ์ด๋์ ๋ฏผ์ฃผ์ ์ผ๋ก ์ ์ถ๋ ์ด๋ฆฌ๋ฅผ ์ ๋ณต์ํค๊ณ ํ๋ ๋น ์์กฐ์ ์ค(Shah)๋ฅผ ๋ณต๊ถ์ํจ ๋๋ก ๊ฑฐ์ฌ๋ฌ ์ฌ๋ผ๊ฐ๋๋ค. ์ค ์ ๊ถ์ ๋น๋ฐ๊ฒฝ์ฐฐ ์ฌ๋ฐํฌ(Savvak)๋ฅผ ํตํด ๊ตญ๋ฏผ์ ์ํนํ๊ฒ ์ต์ํ๊ณ , ์ด๋ 1978๋ ์ด๋ ํ๋ช ์ ์ด๋งค์ ๊ฐ ๋์์ต๋๋ค. ์ด๋์ธ๋ค์๊ฒ โ๋ฏธ๊ตญ์ ์ฃฝ์์โ์ด๋ผ๋ ๊ตฌํธ๋ ๋จ์ํ ์ฆ์ค๊ฐ ์๋, ์ค๋ ์ต์์ ๋ํ ๋ถ๋ ธ์ ํ์ถ์ด์๋ ๊ฒ์ ๋๋ค. 8๋ ๊ฐ์ ์ํนํ ์ด๋-์ด๋ผํฌ ์ ์์ ์ด๋์ธ๋ค์๊ฒ ๊น์ ์์ฒ๋ฅผ ๋จ๊ฒผ๊ณ , 2002๋ ์กฐ์ง W. ๋ถ์ ๋ํต๋ น์ด ์ด๋์ โ์ ์ ์ถ(Axis of Evil)โ์ผ๋ก ์ง๋ชฉํ๋ฉด์ ์ด๋ค์ ๋ถ๋ง์ ๋์ฑ ์ปค์ก์ต๋๋ค. ๋งฅํฌ๋ฆฌ์คํธ ์ฅ๊ตฐ์ โํ์ฌ ๋ฒ์ด์ง๋ ์ผ์ ์ดํดํ๋ ค๋ฉด, ์ด ์ง์ ๊น์ง์ ์ฌ์ ์ ์ดํดํด์ผ๋ง ์ฌ๋๋ค์ ์์ฌ๊ฒฐ์ ์ ์ด๋๋ ํ๋๋ฅผ ์ดํดํ ์ ์๋คโ๊ณ ๊ฐ์กฐํ์ต๋๋ค.
์ธ ๊ฐ์ง ์ ํน: ๋ฏธ๊ตฐ์ ์ ๋ต์ ๋๋ ๋ง
๋งฅํฌ๋ฆฌ์คํธ ์ฅ๊ตฐ์ ์ด๋ ์ ๊ถ์ โ์ฐธ์ ์์ (decapitation strike)โ์ด๋ ์ ์ํ ์ ๊ถ ๊ต์ฒด๊ฐ ์ด๋ ต๋ค๊ณ ์ ๋งํ์ต๋๋ค. ์ด๋์ ๋ฒ ๋ค์์๋ผ์ ๋ง๋๋ก(Maduro) ๊ฐ์ ๋จ๋ฏธ์ ๋ ์ฌ์์๋ ๊ทผ๋ณธ์ ์ผ๋ก ๋ค๋ฅธ ์ ์ด๋ผ๋ ๊ฒ์ ๋๋ค. ์ด์ฌ๋ ํ๋ช ์ ๊ถ์ ์ ํ์ ์ด๊ณ ์ข ๋ง๋ก ์ ์ธ ์์ค์ ํ์ ์ ๊ฐ์ง๊ณ ์์ผ๋ฉฐ, ์ด๋ผํฌ์์ ์์ํ ๋ฏผ๋ณ๋์ ์ธ์ธ ๋ ๋ถ์๋นํ ๋ณ์ฌ๋ค์ด ์ฌ์ง์ด ์น๋ฃํด์ฃผ๋ ์๋ฃ์ง์ ๋ฌผ์ด๋ฏ์ผ๋ ค ํ ์ ๋๋ก ๊ทน๋จ์ ์ธ ํ์ ์ ๋ณด์๋ค๋ ์ผํ๋ฅผ ๋ค๋ ค์ฃผ์์ต๋๋ค. ๊ทธ๋ ์ด๋ ๋ด ๋ฐ์ ๋ถ ์ธ๋ ฅ์ ์ค์ฒด๊ฐ ๋ถ๋ถ๋ช ํ๊ณ , ์ ์์ ์ข ์ข ๊ตญ๋ฏผ๋ค์ ์ ๋ถ ์ฃผ๋ณ์ผ๋ก ๊ฒฐ์ง์ํค๋ ํจ๊ณผ๊ฐ ์๋ค๊ณ ์ค๋ช ํ์ต๋๋ค.
๋งฅํฌ๋ฆฌ์คํธ ์ฅ๊ตฐ์ ๋ฏธ๊ตญ ํ์ ๋ถ์ ๊ตฐ์ด ๋น ์ง๊ธฐ ์ฌ์ด โ์ธ ๊ฐ์ง ์ ํนโ์ ๊ฒฝ๊ณ ํ์ต๋๋ค.
- ๋น๋ฐ ์์ (Covert Action): ์๋ก์ด ๋ํต๋ น์ด ๋ค์ด์๋ฉด ์ ๋ณด๊ธฐ๊ด์ โ์๋ฌด๋ ๋ชจ๋ฅด๊ฒ ์์ฒญ๋ ํจ๊ณผ๋ฅผ ๋ผ ์ ์๋คโ๊ณ ์ ํนํ์ง๋ง, ๊ทธ์ ๊ฒฝํ์ ๋น๋ฐ์ ์ ์ง๋์ง ์๊ณ ์ฑ๊ณตํ๋ ๊ฒฝ์ฐ๋ ๋๋ฌผ๋ค๋ ๊ฒ์ ๋๋ค.
- ์ ๋ฐ ํน์์์ (Surgical Special Operations Raid): ๋ง๋๋ก ์ ๊ถ ์ถ์ถ ์๋์์ ๋ณด๋ฏ์ด, ํน์์์ ์ ํ์ํ ์ญ๋์ ๋ณด์ฌ์ค ์ ์์ง๋ง, ํ์ฅ์ ํ์ค์ ์ค์ง์ ์ผ๋ก ๋ฐ๊พธ๋ ๋ฅ๋ ฅ์ ์ ํ์ ์ ๋๋ค.
- ๊ณต๊ตฐ๋ ฅ(Air Power): ์ 2์ฐจ ์ธ๊ณ๋์ ์ โํญ๊ฒฉ๊ธฐ๋ ํญ์ ๋ซ๊ณ ๊ฐ๋คโ๋ ๋ฏฟ์๋ถํฐ ๋ฒ ํธ๋จ์ ์ ์ ์ง์ ํ์ ์ ๋ต, 2003๋ ์ด๋ผํฌ์ ์ โ์ถฉ๊ฒฉ๊ณผ ๊ณตํฌ(shock and awe)โ๊น์ง, ๋ฏธ๊ตญ์ ๊ณต๊ตฐ๋ ฅ์ ๋ํ ๊ณผ๋ํ ๊ธฐ๋๋ฅผ ๊ฐ์ ธ์์ต๋๋ค. ๊ทธ๋ฌ๋ ๋ฒ ํธ๋จ์ ์์ ๋ถ๋ฒ ํธ๋จ์ ๋น๋์นญ์ ์ผ๋ก ์ ์ ๊ฒฐ๊ณผ์ ํ์ ํ๊ณ , ๊ณต๊ตฐ๋ ฅ๋ง์ผ๋ก๋ ๊ทธ๋ค์ ์์ง๋ฅผ ๊บพ์ ์ ์์์ต๋๋ค. ๋งฅํฌ๋ฆฌ์คํธ ์ฅ๊ตฐ์ โ์ด๋์ ํญ๊ฒฉ์ ๊ฒฌ๋๋ผ ๋น๋ฒํ ๋ฅ๋ ฅ์ ๊ฐ์ง ๋๋ผ์ผ ์ ์๋คโ๋ฉฐ, ์ฌ๋๋ค์ ๋ง์์ ๋ฐ๊พธ์ง ์๊ณ ๋ ์ํ๋ ๊ฒฐ๊ณผ๋ฅผ ์ป๊ธฐ ์ด๋ ต๋ค๊ณ ๋งํ์ต๋๋ค.
ํ๋ ์น ์นผ๋ผ๋์คํธ๋ โ์ด๋ฒ์ ๋ค๋ฅด๋คโ๋ฉฐ ํ๋์ ์ ๋ฐ ๋๋ก , ์ ์ฅ ๊ฐ์์ฑ, ์ด๋ ์ ๊ถ์ ๋ํ ๊น์ ์นจํฌ ๋ฅ๋ ฅ์ ๋ค์ด ๊ณต๊ตฐ๋ ฅ์ ์งํ๋ฅผ ์ฃผ์ฅํ์ง๋ง, ๋งฅํฌ๋ฆฌ์คํธ ์ฅ๊ตฐ์ ์ด๋ฅผ โํฌ์์ ์ธ๊ณ์์ ์ข์ํ๋ ๋งโ์ด๋ผ๋ฉฐ ํ์์ ์ธ ๋ฐ์์ ๋ณด์์ต๋๋ค. ๊ทธ๋ ์ํ๊ฐ๋์คํ์์ ํญ๊ฒฉ์ด ๋ถ์กฑ์๋ค๋ก๋ถํฐ โ๋ ์ ๋ด๋ ค์ ๋์ ๋ง์ฃผํ๊ณ ์ง์ ์ธ์ธ ์์ง๊ฐ ์๋ค๋ฉด ๋๋์ ์ผ๋ก ์ฐ๋ฆฌ ์์ค์ด ์๋๋คโ๋ผ๋ ๊ฒฝ๋ฉธ์ ์๋ ๊ฒฝํ์ ์ธ๊ธํ๋ฉฐ, ์ฌ๋๋ค์ด ์ธ์ฐ๋ ์ด์ ๋ ์ง์ ํ์ ๊ณ์ฐ์ด ์๋ โ์ด์ โ ๋๋ฌธ์ด๋ผ๊ณ ๊ฐ์กฐํ์ต๋๋ค. ์ง๋์์ ๊ฐ์กฑ์ ์ดํดํ๋ ๋ฑ์ โ์ฐธ์ ์์ โ์ ํ์ ์์ง๋ฅผ ๋์ด๋ ๊ฒ์ด ์๋๋ผ ์คํ๋ ค ๋ฐ๊ฐ์ ํค์ธ ๋ฟ์ด๋ผ๋ ๊ฒ์ ๋๋ค.
ํธ๋ฅด๋ฌด์ฆ ํดํ์ ํ์ค: โ์ง๊ธ์ด ๊ฐ์ฅ ์ฌ์ด ๋ถ๋ถ์ด๋คโ
๋ฏธ๊ตญ์ธ๋ค์ ๋ฏธ๊ตฐ์ โ์ํผ๋งจโ์ผ๋ก ์ฌ๊ธฐ๋ฉฐ ํธ๋ฅด๋ฌด์ฆ ํดํ๊ณผ ๊ฐ์ ๋์ ๋ ์ฝ๊ฒ ํด๊ฒฐํ ์ ์๋ค๊ณ ๋ฏฟ๋ ๊ฒฝํฅ์ด ์์ต๋๋ค. ๊ทธ๋ฌ๋ ๋งฅํฌ๋ฆฌ์คํธ ์ฅ๊ตฐ์ ์ด๋ฅผ โ์ด๋ ค์ธ ๊ฒโ์ด๋ผ๊ณ ๋จ์ธํ์ต๋๋ค. ์ด๋ผํฌ ํญ๊ฒฉ๊ณผ ๋ฐ๊ทธ๋ค๋ ์ ๋ น์ ์ฌ์ ์ง๋ง, ํ์ค์ ๋ฐ๊พธ๊ณ ๋๊ฐ ํต์ ๊ถ์ ๊ฐ์ง๋์ง ๊ฒฐ์ ํ๋ ๊ฒ์ โ3๋ง ํผํธ ์๊ณต์ด ์๋ 6ํผํธ ๋์ด์์ ์ ๊ณผ ๊ฐ์ ๋์ด์์ ์ธ์ฐ๋ ๊ฒโ๊ณผ ๊ฐ๋ค๊ณ ๋น์ ํ์ต๋๋ค.
๊ทธ๋ โ์ด ์ ์์ด ๋ง์์ ๋ ๋ค๋ฉด ์ง๊ธ ์ด ์ฒซ ๋ถ๋ถ์ ์ฆ๊ธฐ๋ผ. ์๋ํ๋ฉด ์ง๊ธ ์ดํ๋ก๋ ๋ชจ๋ ๊ฒ์ด ๋ ์ด๋ ค์์ง ๊ฒ์ด๊ธฐ ๋๋ฌธ์ด๋ค. ๊ทธ๋๋ ๋ ๋๋ฑํ ์ธ์์ด ๋ ๊ฒ์ด๋คโ๋ผ๊ณ ๊ฒฝ๊ณ ํ์ต๋๋ค. ํธ๋ฅด๋ฌด์ฆ ํดํ์ ๊ฐ๋ฐฉํ๊ณ ์ ์งํ๋ ๊ฒ์ ์ง๋ขฐ, ์์จ ์์ ๋ฐ ์์ค ํญ๊ณต๊ธฐ, ์ฌ์ง์ด ๋ฏผ๊ฐ ์ ์กฐ์ ์ด๋ ํ๋ฌผ์ ์ ํ์ ์ผ๋ก ์ผ๋ ์ ๊ฐ๋ ๊ณต๊ฒฉ ๋ฑ ์ด๋์ด ๋์ํ ์ ์๋ ๋ค์ํ ์ํ์ผ๋ก ์ธํด ๊ทนํ ์ด๋ ค์ธ ๊ฒ์ ๋๋ค. ์ด๋์ด ์ผ์ฃผ์ผ์ ํ ๋ฒ ๋ฏผ๊ฐ ์ ๋ฐ์ ๊ณต๊ฒฉํ๊ธฐ๋ง ํด๋, ๋ณดํ์ฌ๋ค์ ๋ณดํ ๊ฐ์ ์ ๊ฑฐ๋ถํ ๊ฒ์ด๊ณ , ์ด๋ ํดํ ํตํญ์ ์ฌ์ค์ ๋ถ๊ฐ๋ฅํ๊ฒ ๋ง๋ค ๊ฒ์ด๋ผ๋ ์ค๋ช ์ ๋๋ค.
์ด๋ฌํ ๊ฐ๋ฑ์์ ์ค๋ ์ํ์ ๊ฒฝ์ ์ (๊ณ ์ ๊ฐ, ๊ธ๋ก๋ฒ ๊ฒฝ์ ๋ฌธ์ )์ด๊ณ ์์ฐ์ (2์ฒ์ต ๋ฌ๋ฌ ์ด์์ ๊ตญ๋ฐฉ ์์ฐ ์ฆ์ก ์๊ตฌ)์ผ ๋ฟ๋ง ์๋๋ผ, ๊ฐ์ฅ ์ค์ํ โ์ธ๋ช ํผํดโ๋ก ์ด์ด์ง๋ค๊ณ ๋งฅํฌ๋ฆฌ์คํธ ์ฅ๊ตฐ์ ๊ฐ์กฐํ์ต๋๋ค. ์ง๊ธ๊น์ง๋ ์ฌ์์๊ฐ ์ ์์ง๋ง, ์ ์์ด ์ฅ๊ธฐํ๋๊ณ ์ง์๊ตฐ์ด ํฌ์ ๋ ๊ฒฝ์ฐ ์ฌ์์๋ ๊ธ์ฆํ ๊ฒ์ด๋ฉฐ, ์ด๋ ์์๊ตฐ์ผ๋ก ์ด๋ฃจ์ด์ง ์ฌํ์ ๊ตฐ์ธ ์ฌ์ด์ โ๋ถ์ด๊ณผ ๋ฐ๊ฐโ์ ์ด๋ํ ์ ์๋ค๊ณ ๊ฒฝ๊ณ ํ์ต๋๋ค.
๋ฆฌ๋์ญ๊ณผ ๊ตฐ์ ๋ฏธ๋: ๋๋๊ฐ ๋ํ๊ทผ๋ณด๋ค ์ค์ํ๋ค
ํ๋ ์น ์นผ๋ผ๋์คํธ๋ ํ์ฌ ๊ตญ๋ฐฉ์ฅ๊ด์ โ๊ฐ๋ ฅํจโ์ ๊ณผ์ํ๋ ๋ฏํ ํ๋(๋ฒค์นํ๋ ์ค, ํธ์ฌ์ , โ์ฐ๋ฆฌ๋ ์น๋ช ์ ์ด๊ณ ํ๊ดดํ ๊ฒ์ด๋คโ ๊ฐ์ ํธ์ธ์ฅ๋ด)์ ๋ํ ๊ตฐ์ธ๋ค์ ๋ฐ์์ ๋ฌผ์์ต๋๋ค. ๋งฅํฌ๋ฆฌ์คํธ ์ฅ๊ตฐ์ ์ด๋ฌํ ์ต๊ณ ์์ธต์ ๋ฉ์์ง์ โ์ค๋งํ๋คโ๊ณ ๋ตํ์ต๋๋ค. ๊ทธ๋ ์ง์ ์ผ๋ก ์๋ฆฌํธ ๋ถ๋์ ์ผ์์ด์๋ ์ฌ๋๋ค์ ๊ฐ์ด์ ์น๊ฑฐ๋ ์๋ํ์ง ์์ผ๋ฉฐ, โ์ฌ๋ ์ฃฝ์ด๋ ๊ฒ์ ์ข์ํ๋คโ๊ณ ๋งํ์ง ์๋๋ค๊ณ ์ง์ ํ์ต๋๋ค.
๋ฌธ์ ๋ ์ด๋ฌํ ์ธํ์ด 18์ธ์ ์ ์ ๋ณ์ฌ๋ค์๊ฒ ์ํฅ์ ๋ฏธ์ณ โ์ฐ๋ฆฌ๊ฐ ๊ทธ๋ ๊ฒ ์๊ฐํด์ผ ํ๋ค. ์ฐ๋ฆฌ๋ ์ฐ์ํ๋คโ๊ณ ๋ฏฟ๊ฒ ๋ง๋ค ์ ์๋ค๋ ๊ฒ์ ๋๋ค. ๊ทธ๋ โํฐ ์ด๋๋ฐ๊ทผ์ ๊ฐ์ง๊ณ ๋ฌธ์ ๋ฐ๋ก ์ฐจ๊ณ ๋ค์ด๊ฐ ํ์๊ฐ ์๋ ์ฌ๋์ ์๋ ๊ทนํ ์ ๋คโ๋ฉฐ, ๊ตฐ์ ๋๋ถ๋ถ์ ์ ๋ณด, ํต์ , ๋ฌผ๋ฅ ๋ฑ ์ง์ ๋ณ๋ ฅ์ผ๋ก ์ด๋ฃจ์ด์ ธ ์๋ค๊ณ ์ค๋ช ํ์ต๋๋ค. โ๋ชจ๋ ์ฌ๋์ด ๋์ฒ๋ผ ๋ณด์ฌ์ผ ํ๋คโ๋ ์๊ฐ์ ์ฌ์์ ์ด๋ํ ๊ฒ์ด๋ฉฐ, ๊ฒ์ด๋ ํธ๋์ค์ ๋ ๋ฑ ๋๊ตฌ๋ ๋ฅ๋ ฅ์ด ์๋ค๋ฉด ๋ณต๋ฌดํด์ผ ํ๋ค๊ณ ์ฃผ์ฅํ์ต๋๋ค. ๋งฅํฌ๋ฆฌ์คํธ ์ฅ๊ตฐ์ ์์ ์ด ์ ์์ ๋๋ ๋์ง์ ์ด์๋ ๋ํ ๋ฌ ๋ถ๋๊ฐ ์ด๋ผํฌ์ ๋น์์๋ ๋์ด ๋ ๋จ๋ , ์ ์์ด๋ค ๋ฑ ๋ค์ํ ๋ฐฐ๊ฒฝ์ ๊ฐ์ง ์ฌ๋๋ค์ด ๋ชจ์ธ โ๋ฅ๋ ฅ์ฃผ์ ์ง๋จโ์ผ๋ก ๋ฐ์ ํ์ผ๋ฉฐ, ์ด๋ ํจ์ฌ ๊ฑด๊ฐํ ๊ตฐ๋์๋ค๊ณ ํ๊ณ ํ์ต๋๋ค.
๊ทธ๋ ์ฌ์ง์ด โ์ ์ฌ(warrior)โ๋ผ๋ ๋จ์ด์๋ ๋ฌธ์ ๋ฅผ ์ ๊ธฐํ์ต๋๋ค. ์ ํต์ ์ผ๋ก ์ ์ฌ๋ ๋ณ์ฌ์ ๋ถ๋ฆฌ๋์์ผ๋ฉฐ, ๊ตฐ๋์ ํญ๋(mob)์ ์ฐจ์ด๋ โ๊ท์จโ๊ณผ โ๋ฆฌ๋์ญโ, ๊ทธ๋ฆฌ๊ณ โ๊ตฐ์ฌ๋ฒ๊ท(UCMJ)โ์ ์๋ค๊ณ ๊ฐ์กฐํ์ต๋๋ค. ์ ์์ด๋ค์๊ฒ ์๋ช ์ ์์๊ฐ ์ ์๋ ๋ฌด๊ธฐ๋ฅผ ์ฅ์ฌ์ค ๋๋ ๊ฐ์น์ ๋ฌธํ, ๊ทธ๋ฆฌ๊ณ ๊ตฐ์ฌ์ ์ผ๋ก ๊ท์ ๋ ๊ท์จ์ด ํ์์ ์ด๋ผ๋ ๊ฒ์ ๋๋ค. ๋งฅํฌ๋ฆฌ์คํธ ์ฅ๊ตฐ์ โํฐ ๋๊ฐ ํฐ ์ด๋๋ฐ๊ทผ๋ณด๋ค ์ค์ํ๋คโ๊ณ ์ญ์คํ๋ฉฐ, ๋๋ก ์ ์์์์ ํ์ ๊ณผ ๋๋ผ์ด ํ๋ณต๋ ฅ์ ๋ณด์ฌ์ฃผ๋ ์ฐํฌ๋ผ์ด๋๊ตฐ์ ๊ทธ ์ฆ๊ฑฐ๋ก ์ ์ํ์ต๋๋ค.
์๋ฌด์ ๊ตญ๊ฐ ๋ด์ฌ: ๋ถ์ด๋ ์ฌํ๋ฅผ ์๋ ๋ค๋ฆฌ
๋ํ์ ๋ง์ง๋ง์ ๊ตญ๊ฐ ๋ด์ฌ์ ์ค์์ฑ์ผ๋ก ์ด์ด์ก์ต๋๋ค. ํ๋ ์น ์นผ๋ผ๋์คํธ๋ ์์๋ด์ฌ๋ฅผ ์ง์งํ์ง๋ง, ์๋ฌด์ ์ธ ๊ตญ๊ฐ ๋ด์ฌ์ ๋ํด์๋ ์์ ์ง์์ฃผ์์ ๊ด์ ์์ ํ์์ ์ด๋ผ๊ณ ๋ฐํ์ต๋๋ค. ๋ฐ๋ฉด ๋งฅํฌ๋ฆฌ์คํธ ์ฅ๊ตฐ์ ๊ณผ๊ฑฐ์๋ ์์๋ด์ฌ๋ฅผ ์ง์งํ์ง๋ง, ์ด์ ๋ โ์๋ฌด์ (mandatory)โ ๊ตญ๊ฐ ๋ด์ฌ๋ฅผ ์ฃผ์ฅํ๊ณ ์์ต๋๋ค.
๊ทธ๋ 36์ธ์ ๋ค๋ฆ๊ฒ ๊ตฐ์ ์ ๋ํ ํ๋ ์น ์นผ๋ผ๋์คํธ์ ์ฌ๋ก๋ฅผ ๋ค๋ฉฐ, 17, 18์ธ์ ์ ์์ด๋ค์ ์ต์ ์ ์ ํ์ ํ ๋งํผ ์ฑ์ํ์ง ์์ ๊ฒฝ์ฐ๊ฐ ๋ง๋ค๊ณ ์ง์ ํ์ต๋๋ค. โ๋ชจ๋ ์ฌ๋์ด ์ฌ๋ฐ๋ฅธ ๋ต์ ์ฐพ์ ๋๊น์ง ๊ธฐ๋ค๋ฆฐ๋ค๋ฉด ๋๋ฌด ๋ง์ ์๊ฐ์ด ๊ฑธ๋ฆฐ๋คโ๋ ๊ฒ์ ๋๋ค. ๋งฅํฌ๋ฆฌ์คํธ ์ฅ๊ตฐ์ ์๋ฌด์ ์ธ ๊ตญ๊ฐ ๋ด์ฌ๊ฐ โ๋ฏธ๊ตญ ์ฌํ์ ํ๋ฅญํ ํ์คํ ์ฅ์นโ๊ฐ ๋ ๊ฒ์ด๋ผ๊ณ ๋ฏฟ์ต๋๋ค. ๋ชจ๋ ๋ฏธ๊ตญ์ธ์ด ๊ฒช์ด์ผ ํ ๊ณตํต๋ ๊ฒฝํ์ด ์๊ธฐ๋ฉด, ๋์ค์ โ์ด๋์ ๋ด์ฌํ๋๋?โ๋ ์ง๋ฌธ์ ํตํด ๋ํ๊ฐ ์์๋๊ณ ์ฌํ์ ๋ถ์ด์ ํด์ํ๋ ๋ค๋ฆฌ ์ญํ ์ ํ ์ ์๋ค๋ ๊ฒ์ ๋๋ค. ๋ถํธํ๊ฑฐ๋ ๋ถ์พํ ์ ์๋ ๊ฒฝํ์ ํตํด ์ฌ๋๋ค์ ๋ ๋์ ์กด์ฌ๋ก ์ฑ์ฅํ ์ ์๋ค๊ณ ๊ทธ๋ ๊ฐ์กฐํ๋ฉฐ, 71์ธ์ ๋์ด์๋ ๋ถ๊ตฌํ๊ณ ์ ์์ด๋ค์๊ฒ โ์ง๊ธ์ด ์๋๋ฉด ์ธ์ ์ด๋ฐ ๋ง์ ํ ์ ์๊ฒ ๋๊ฐโ๋ผ๋ฉฐ ์์ ์ ์์ ์ ํผ๋ ฅํ์ต๋๋ค.
์ด๋ฒ ๋ํ๋ ์ด๋๊ณผ์ ๋ณต์กํ ๊ฐ๋ฑ์ ์ดํดํ๊ณ , ๊ตฐ์ฌ ์ ๋ต์ ํจ์ ์ ๊ฒฝ๊ณํ๋ฉฐ, ๊ตฐ์ ๋ฆฌ๋์ญ๊ณผ ์ฌํ์ ์ญํ ์ ๊น์ด ์ฑ์ฐฐํ๋ ๊ณ๊ธฐ๋ฅผ ์ ๊ณตํ์ต๋๋ค. ๋งฅํฌ๋ฆฌ์คํธ ์ฅ๊ตฐ์ ํต์ฐฐ์ ๋จ์ํ ๊ตฐ์ฌ์ ๋ฌธ์ ๋ฅผ ๋์ด, ๋ถ์ด๋ ์ฌํ๋ฅผ ํตํฉํ๊ณ ๋ฏธ๋ ์ธ๋๋ฅผ ์ก์ฑํ๋ ๋ฐฉ์๊น์ง ์์ฐ๋ฅด๋ ํญ๋์ ์์ผ๋ฅผ ์ ์ํ์ต๋๋ค. โ์ง๊ธ ์ดํ๋ ๋ชจ๋ ๊ฒ์ด ๋ ์ด๋ ค์ธ ๊ฒโ์ด๋ผ๋ ๊ทธ์ ๊ฒฝ๊ณ ๋ ์ฐ๋ฆฌ๊ฐ ์ง๋ฉดํ ํ์ค์ ๋ํ ๋์ฒ ํ ์ธ์๊ณผ ํ๋ช ํ ๋์์ ์ด๊ตฌํ๋ ์ธ๋ฆผ์ผ๋ก ๋ค๊ฐ์ต๋๋ค.