YouTube Digest

March 31, 2026

Listen to this issue

English

Based on “7 free ways to get customers with AI” from Greg Isenberg Watch the original video

Beyond the Code: Why Distribution is the New Gold Rush in the AI Era

The world is abuzz with the transformative power of Artificial Intelligence. Every day, countless innovators are “vibe coding” revolutionary new tools, products, and services, leveraging AI to build faster and more efficiently than ever before. Yet, amidst this flurry of creation, a critical question often goes unanswered: How do you get customers to actually see and use these brilliant creations?

According to entrepreneur and investor Greg Isenberg, this is the defining challenge of our time. He boldly asserts, “The wealthiest people will be marketers over the next 10 years.” While countless tutorials focus on the intricacies of building with AI, the real scarcity — and thus the real opportunity — lies in understanding and mastering distribution.

Isenberg argues that a profound shift, which he calls “the great flip,” has occurred in Silicon Valley’s hierarchy. In 2014, engineers were at the top, then product managers, with marketing relegated to the bottom, often seen as the “laughingstock.” Today, AI has fundamentally commoditized code, making product development increasingly accessible. This has propelled those who understand how to get customers – the distribution experts – to the pinnacle.

The trap for many builders is the “build it and they will come” fallacy. They pour their energy into creating the perfect product, only to launch it into silence. The smart builders, Isenberg contends, flip this paradigm: distribution first, product second. They cultivate an audience, understand their needs, then rapidly build a solution, launching it to a warm, engaged community.

Consider the success of Peter Levels, the creator of Nomad List. With over $3 million in revenue and zero employees, his secret isn’t just a great product (which can be copied quickly), but a massive audience (750,000+ followers) and excellent SEO. He’s tweeted over 125,000 times in eight years, consistently building trust and iterating with his users. This dedication to audience building and distribution is his true moat.

For those ready to move beyond just building and start acquiring customers, here are seven powerful, often free, strategies to harness AI for distribution:


1. Let AI Sell for You: Build an MCP Server

Imagine your product having a 24/7 sales team that costs you nothing. This is the promise of building an MCP (Machine-to-Computer Protocol) server, which acts like an app or plugin for AI assistants.

How it Works: When a user asks an AI (like Claude or ChatGPT) a question, the AI discovers your MCP server, which is designed to provide data or functionality related to that query. Your product is then returned as a solution, delivering instant value to the user. This effectively turns the AI assistant into your sales team, leading to zero Customer Acquisition Cost (CAC).

Isenberg cites examples of friends incubating projects in this space, with one fintech company achieving over 150 installations in 30 days with $0 ad spend. He likens building an MCP server today to building for mobile in 2010 – a massive, untapped opportunity.

How to Start This Week:

  1. Identify Your Product’s Core Question: What specific problem does your product solve that an AI user might ask about?
  2. Build Your MCP Server: Create a server that returns the relevant data or functionality. This can often be “vibe coded” in 24 hours or less.
  3. Publish to Registries: Submit your server to MCP registries like Smithery, MCPT, and Open Tools.
  4. Automated Sales: Every AI assistant connected to your server becomes a perpetual salesperson for your product.

2. Scale with Speed: Programmatic SEO

Forget manually writing thousands of blog posts. Programmatic SEO leverages AI to create vast numbers of targeted, high-quality SEO pages in a fraction of the time, allowing you to dominate specific keyword patterns.

How it Works: The strategy revolves around identifying a keyword pattern like “Best X for Y” (e.g., “best CRM for dentists” or “roofing in Miami”). You then gather structured data using tools like Firecrawl (which scrapes and cleans data) or existing databases. A page template (built with tools like Next.js or Webflow) combined with AI-generated unique content for each page allows you to scale rapidly.

While the idea of pressing a button for 10,000 pages might sound like a recipe for “slop,” the key is optimization. Start with a smaller batch, refine the AI content to ensure it doesn’t feel AI-generated, and potentially incorporate human editing.

The math is compelling: 10,000 pages, each getting a modest 30 visits a month, translates to 300,000 monthly visitors. A 2% conversion rate at $10 per conversion generates $60,000 a month from pages built once. While SEO takes time to compound, the directional advantage is clear.

How to Start This Week:

  1. Pick a Keyword Pattern: Choose a niche-specific pattern (e.g., “product type for niche,” “service in city”).
  2. Build Your Data Set: Scrape data with Firecrawl or utilize existing databases.
  3. Create a Page Template: Use your preferred framework (Next.js, Webflow, WordPress).
  4. Generate AI Content: Use AI to create unique, high-quality paragraphs and information for each page, not just variable swaps. Consider a human in the loop for editing.
  5. Publish an MVP: Launch 100 pages, monitor indexation, and then scale up.

3. The Tool is Your Marketing: Free Tools as Top of Funnel

Free tools have long been a powerful marketing strategy, but AI has supercharged their potential. Now, you can build and ship valuable tools in days, not months, creating an organic, viral marketing loop.

How it Works: The concept is simple: create a free tool (a grader, analyzer, calculator, etc.) that offers instant value to users. A classic example is Ahrefs’ backlink checker, which gives users a taste of their full product, enticing them to subscribe for deeper insights.

  1. Build a Free Tool: Leverage AI and cloud code to rapidly develop a useful tool.
  2. Instant Value & Data Capture: Users get immediate results (e.g., “Your site scores 43/100”). Crucially, you capture their contact information (email/SMS) to deliver the results.
  3. Encourage Sharing: Design the tool’s output to be shareable, creating social proof and backlinks.
  4. Viral Loop: More users discover the tool through shares, driving organic growth.
  5. Upsell to Paid Product: Once users experience value and identify problems, offer your paid product as the solution.

Isenberg emphasizes that with today’s AI capabilities, it’s “not crazy to build a free tool a week.” This shifts marketing from an editorial calendar to a “free tool calendar.”

How to Start This Week:

  1. Brainstorm Ideas: Ask your LLM of choice to generate 10 free tool ideas based on your product.
  2. Prioritize & Validate: Select the best ideas, perhaps by asking your audience which they’d find most useful.
  3. Build & Ship: “Vibe code” your chosen tool in a day and launch it.
  4. Integrate with Your Product: Ensure the tool naturally leads users to your paid offering.

4. Be the Source AI Cites: Answer Engine Optimization (AEO)

As AI search and answer engines like Perplexity and ChatGPT grow, the game of SEO is evolving. Traditional keyword stuffing and 30,000-word blog posts are giving way to a new imperative: becoming the definitive source that AI cites.

How it Works: AEO focuses on providing structured, direct, and authoritative answers that AI can easily parse and reference. This means moving beyond getting clicks to your site and instead aiming for “zero-click searches” where AI directly answers a user’s question, citing your content as the source.

Strategies include:

  • Structured Direct Answers: Clear, concise answers to common questions.
  • FAQ Format: Presenting information in a Q&A style.
  • Schema Markup: Using code to tell search engines and AI what your content is about.
  • Comparisons and Tables: Presenting data in easily digestible formats.

Peter Levels reportedly saw his AI referrals jump from 4% to 20% in just one month, highlighting the immense potential of AEO. First movers in this space will own niches for years.

How to Start This Week:

  1. Identify Top Questions: Google the top 20 questions your customers ask.
  2. Write Definitive Answers: Create clear, direct, citation-worthy answers for each question – no fluff.
  3. Add Schema Markup & FAQ Blocks: Implement technical SEO elements to help AI understand your content.
  4. Publish on Authority Domains: Publish on a domain with existing authority, or build authority through other strategies.
  5. Monitor Citations: Track your mentions and citations on Perplexity, ChatGPT, and other AI engines using tools like Otterly or manual checks.

5. Make Outputs Sharable: Viral Artifacts

What do Spotify Wrapped, GitHub contribution graphs, Stripe Atlas incorporation milestones, and Duolingo streaks have in common? They are all highly shareable outputs that users are proud to display, generating massive free marketing.

How it Works: The core idea is to identify what your user wants to brag about and then make that thing beautiful and easily shareable. It taps into human psychology, allowing users to express identity, celebrate achievements, and connect with others.

  • Spotify Wrapped: Users share their year in music, reflecting their identity.
  • GitHub Contribution Graph: Developers proudly display their “green squares” of coding activity.
  • Stripe Atlas: Founders tweet about their incorporation anniversary.
  • Duolingo Streaks: Users boast about their continuous language learning.

This strategy works not just for consumer products but also for B2B. Business users might share achievements within Slack or Microsoft Teams. The key is to create an artifact that resonates with a desire to show off or validate effort.

How to Start This Week:

  1. Identify Shareable Milestones: What output or achievement would your user screenshot and share?
  2. Design for Beauty & Brand: Make the artifact visually appealing and branded subtly (logo present but not overwhelming).
  3. Add a Share Button: Implement a one-click share button that prefills the post with the artifact and a link to your product.
  4. Leverage Free Impressions: Every share by a user provides free impressions to their network, which is often your target audience.

6. Acquire Your Audience: Buy a Niche Newsletter

Building an audience from scratch is hard. It takes years, daily content, and comes with no guarantee of success. What if you could skip the grind and instantly acquire a direct channel to your target customers?

How it Works: Instead of starting from zero subscribers, you can purchase an existing niche newsletter. Many smaller newsletters (5,000-50,000 subscribers) are not effectively monetized and can be acquired for a modest sum, often between $5,000 to $20,000.

  1. Instant Trust & Audience: You immediately inherit a loyal, engaged audience and their trust.
  2. Direct Product Integration: You can plug your product into the newsletter immediately, driving traffic and sales.
  3. Repeat Across Niches: Once successful, you can repeat the process with other newsletters in related niches.
  4. Owned Channel: Unlike social media, which can suppress reach or change algorithms, a newsletter is a direct, owned channel to your audience.

This strategy is particularly effective if you have confidence in your product and understand its Lifetime Value (LTV).

How to Start This Week:

  1. Find Newsletters: Search platforms like Deuce.com, Newsletter Investor, or directly on Twitter/Substack for niche newsletters.
  2. Target Size: Look for newsletters with 5,000 to 50,000 subscribers in your target niche.
  3. DM the Owner: Initiate contact with a simple question: “Have you ever thought about selling?” You might be surprised by the response.
  4. Make a Fair Offer: Many owners making $0-$500/month would be open to an offer in the $5,000-$20,000 range.

7. Maximize Content Output: AI Repurposing Engine

In today’s content-saturated world, consistency and volume are key. An AI content repurposing engine allows you to take one “pillar” piece of content and transform it into dozens of different formats for multiple platforms, dramatically increasing your touchpoints.

How it Works: Start with a single, high-value piece of content – a 30-minute podcast episode, a YouTube video, or a long-form essay. Then, use AI to atomize and repurpose it:

  • Transcription: Convert audio/video to text.
  • AI Transformation: Feed the transcription into an LLM (like Claude, ChatGPT, or specialized tools) and prompt it to create:
    • 5-10 tweet threads
    • 3-5 LinkedIn posts
    • 2-3 short-form videos (using AI tools for re-motion/editing)
    • 1 newsletter edition
    • 1 blog post
    • 5-10 quote graphics (using creative LLMs)
    • Email sequences

Tools like OpenClaw, Cloud Dispatch, or Perplexity Computer can automate much of this. The goal is to generate “shots on net” across various platforms, leveraging the power of “for you” pages to reach new audiences without needing a large existing following.

How to Start This Week:

  1. Create Pillar Content: Record one 30-minute piece of content (podcast, video, or even just a voice memo). It’s often easier to speak than to write a long essay.
  2. Transcribe: Use an AI service to transcribe your content.
  3. AI Repurpose: Drop the transcription into an LLM and prompt it to generate various content formats (tweets, LinkedIn posts, newsletter).
  4. Optimize & Refine: While AI will generate content, you’ll need to optimize it to avoid “slop,” ensuring it aligns with your brand, incorporates reference images, and is high quality.
  5. Schedule & Distribute: Schedule the repurposed content across all relevant platforms.

The New Moat: Distribution

The era of “vibe coding” in isolation is over. While building incredible AI-powered products is more accessible than ever, it’s no longer the differentiator. As Greg Isenberg emphatically states, “Code used to be the moat… distribution is the new moat.”

Distribution is scarce, something AI cannot build for you. Product, while important, is becoming commoditized. Code, he argues, is fully commoditized. Therefore, the ability to consistently get customers and build a growth engine is the most valuable skill set in the AI economy.

Don’t just build in silence. Pick two of these seven distribution weapons and start implementing them this week. Get customers, make money, grow, reinvest, and have fun. The future belongs to those who not only build with AI but also master how to get their creations into the hands of the people who need them.


Based on “How to turn Claude code into your personal life operating system | Hilary Gridley” from How I AI Watch the original video

Reclaim Your Day: Hilary Gridley’s “Anti-System” Approach to Personal AI

In an age where every minute feels increasingly precious, and attention is a fractured commodity, the quest for productivity often leads to complex, time-consuming systems. But what if the secret to managing your life, career, and even new parenthood with AI isn’t about building the perfect system, but rather cultivating an “anti-system” that learns from you?

Meet Hilary Gridley, an entrepreneur and new mom who has mastered this philosophy, leveraging tools like Claude Code to create a personal life operating system that reduces friction, adapts to her real-world behavior, and prioritizes human joy over relentless optimization.

The High Cost of Time: A New Mom’s Revelation

For Gridley, the arrival of her baby amplified an already pressing challenge: the opportunity cost of her time had never been higher, nor had it been harder to achieve her goals. “My attention is super fractured,” she explains. “I’m tracking about a thousand more life admin things and logistics than I ever have been before, and I’m historically not great with logistics and life admin.”

This isn’t about hyper-productivity for its own sake. Instead, it’s about reclaiming time for what truly matters – whether that’s engaging work, spending time with her son, or simply enjoying a baguette and some cheese at the market. Her vision isn’t a post-AGI utopia where robots do everything, but one where AI frees her to pursue meaningful work and delightful leisure.

The “10x Impact” Automation Framework

How does one decide what to automate and what to pour human effort into? Gridley employs a surprisingly simple, yet profound, framework:

For any possible task, if I were 10 times better at it, would it have 10 times the impact?

  • If the answer is NO: Automate it. These are tasks where increased skill doesn’t proportionally increase value. Think of moving pixels around a PowerPoint deck. While necessary, becoming a pixel-moving virtuoso won’t make you 10 times better at your core job.
  • If the answer is YES: Invest human time and effort. These are the activities where genuine improvement yields exponential returns. For Gridley, this means crafting compelling narratives, generating innovative ideas, or deeply understanding user research to make better product decisions.

This framework extends beyond work. In her personal life, it helps distinguish between chores and truly enriching activities. Baking a loaf of bread, for example, might be automated if it feels like a chore, but deeply invested in if it brings genuine joy and fulfillment. It’s a dynamic assessment, acknowledging that what’s valuable to automate can also shift with one’s learning curve. A task that might be a crucial learning experience for a junior team member could be ripe for automation for a seasoned expert.

The “Anti-System System”: Learning by Doing

Many AI-powered productivity systems demand significant upfront setup, meticulous organization, and constant maintenance. Gridley calls this her “anti-system system” because it actively resists these demands.

“The whole point of this is that I don’t want to put work into maintaining a system. I don’t want to do a bunch of setup,” she asserts. Her philosophy is simple: you learn by doing. Instead of pre-defining rigid preferences or complex workflows, her AI, Claude Code, observes her real-world behavior and adapts over time. This approach drives the cost of setting up and maintaining the system “to zero” because Claude handles everything.

This contrasts sharply with what she playfully calls “Notion boys” – those who thrive on intricate databases and linked assets. Gridley, much like the legendary Al Gore whose famously chaotic desktop she once photographed, prefers a more “stream of consciousness” approach, where simplicity and adaptability reign supreme.

Building Your Life OS with Claude Code

Here’s a breakdown of how Gridley implements her low-friction, AI-powered system:

1. Reducing Friction: The Lock Screen Shortcut

The first step is surprisingly low-tech: a simple phone shortcut. With a double-tap on the back of her iPhone, a dictation box appears, allowing her to quickly capture any stray thought or to-do, like “reschedule pediatrician appointment.” This immediate capture mechanism ensures nothing escapes the “leaky sieve” of a busy brain, feeding into her system without any AI involvement at this initial stage.

2. Claude Code: The Terminal-Based Assistant

For those intimidated by command-line interfaces, Gridley demystifies Claude Code. Installing it is a quick copy-paste into the terminal, and then it’s ready to run. Her philosophy is pragmatic: “Don’t ask questions. Dangerously skip permissions.” (Though she admits to being a “scaredy-cat” and doesn’t actually skip permissions). The point is, the barrier to entry is far lower than many perceive.

3. “Plan My Day”: Intelligent Scheduling

With Claude Code running, Gridley simply types “plan my day.” The AI then springs into action, following a set of instructions it has been given, along with insights it has learned:

  • Reminders Integration: Claude pulls from the list of reminders captured via the phone shortcut, organizing them into a simple markdown text file. This file, living in a dedicated “Claude folder,” is edited solely by the AI, remaining “none of my business how Claude gets its work done.”
  • Learned Preferences: Crucially, Claude doesn’t rely on Gridley manually inputting her ideal schedule. Instead, it observes her over time. It learns about her daily constraints (e.g., pumping schedule, childcare needs, husband’s availability), and even her quirks (e.g., avoiding certain types of work before bed to prevent getting “in flow” and staying up late). This ensures schedules are based on real behavior, not aspirational ideals.
  • Calendar Integration & Task Breakdown: Claude reviews her existing calendar and then asks what she wants to accomplish. For a large task like “baby passport,” Claude intelligently breaks it down into the very first, actionable step: “make the post office appointment.” This small, 10-minute task is then slotted into her schedule, making daunting admin feel manageable. The AI even adds “hippo emojis” to her calendar entries, a personal touch that helps her track her commitments.

This “management by calendar” approach is transformative. By blocking out time for specific, small tasks, Gridley commits to getting things done and can later observe the delta between her intentions and reality.

4. The “Yappers API”: Human-in-the-Loop Feedback

Perhaps the most surprising aspect of Gridley’s system is her resistance to over-integrating tools. She doesn’t want Claude “connecting all a bunch of stuff in the background” without her involvement. Instead, she relies on what she playfully calls the “Yappers API.”

With Claude Code open in the terminal alongside her work, she simply talks to it throughout the day. If she writes a blog post, she tells Claude, “I wrote a blog post.” If she completes a part of the baby passport task, she informs Claude. This direct, verbal feedback allows Claude to observe and take notes on her actual activities.

At the end of the day, Claude compiles a “daily note” – another markdown file – logging her scheduled tasks versus what she actually did. Gridley can periodically check in, asking Claude, “What are you seeing in terms of the gap between what I’m trying to get done and what I’m actually doing?” Claude might then highlight patterns: “building is crowding out writing,” “you’re doing a bad job time boxing something,” or “most days list three priorities, but only number one gets real time.”

This continuous feedback loop is where the “learning by doing” truly shines. Claude isn’t just executing; it’s learning and improving its understanding of Gridley’s workflow, making its recommendations increasingly effective without her needing to meticulously define rules.

5. The “Mom Sharing News API”: Visual Context

Beyond talking, Gridley advocates for other low-friction input methods, like screenshots. Just as a mom might screenshot an article to share context, users can screenshot their calendar, a doctor’s note, or any visual information and feed it to their AI. This “Mom Sharing News API” is a powerful reminder that advanced AI doesn’t always require complex API integrations; sometimes, a simple image and a spoken instruction are all that’s needed to provide crucial context.

Complexity Must Earn Its Keep

Gridley’s final piece of advice is a guiding principle: complexity has to earn its keep. She starts with the “jankiest version” of any workflow, often held together with “screenshots and prayers,” and only invests in deeper integrations (like connecting her Google Calendar) if the workflow proves consistently valuable over time. Her personal “hit rate” for successful workflows is around 20%, meaning a lot of ideas don’t pan out. By starting simple, she avoids wasted time building and troubleshooting systems that ultimately don’t stick.

This approach isn’t just for personal productivity. Gridley suggests it’s a vital strategy for professionals seeking to implement AI at work. Instead of demanding full company data access upfront, prove the value of a “janky” AI-powered workflow with limited data. Show the impact, and then you’ll earn the permissions needed to scale.

Hilary Gridley’s “anti-system system” offers a refreshing perspective on personal AI. It’s a testament to the power of low-friction tools, adaptive learning, and a clear understanding of what truly deserves our precious human attention. By letting AI handle the logistical plates, we can free ourselves to pursue meaningful work and, perhaps most importantly, carve out more joy in our increasingly complex lives.


Based on “Trump Says He’s Ready for Diplomacy. Iran? Not So Much.” from New York Times Podcasts Watch the original video

A Dangerous Dance: Trump’s Diplomatic Push Meets Iranian Defiance

Despite a month of escalating military threats and strikes, President Trump appears increasingly eager to de-escalate the conflict with Iran through diplomacy. His problem? Iran is not only unwilling to come to the negotiating table on America’s terms but seems to be actively using the ongoing conflict to strengthen its own hand. The result is a perilous standoff, a war of words and wills, where the path to peace remains elusive and fraught with unparalleled complexity.

The confusion surrounding the state of US-Iran relations is palpable. Last week, President Trump explicitly claimed, “I am pleased to report that the United States of America and the country of Iran have had very good and productive conversations regarding a complete and total resolution of our hostilities in the Middle East.” Yet, mere hours later, Iran’s government issued an unequivocal denial: “We deny what US President Donald Trump said regarding any negotiations between the United States and the Islamic Republic of Iran.”

According to David Sanger, a veteran foreign policy correspondent, both sides are engaged in a strategic “fibbing.” Trump, having overseen the striking of 11,000 targets, wants to project an image of having forced Iran to the negotiating table through sheer brute strength. Iran, conversely, driven by national pride and a shrewd understanding of the current dynamics, is determined not to appear as if it’s capitulating to American demands. They fundamentally believe they are “winning by virtue of the fact that we’re still alive,” as Sanger puts it. This leaves President Trump in the unusual position of having to pivot from demanding unconditional surrender to calling for talks, which, by their very nature, imply compromise.

Trump’s Diplomatic Imperative: A “Real Box”

The shift in Trump’s stance isn’t arbitrary. Despite the significant military action, the US has discovered the limitations of its force in Iran, coupled with severe blowback. The Iranian regime, though diminished, is far from collapsing; regime change is not on the horizon. More critically, the conflict has paralyzed the global energy industry by locking up the Strait of Hormuz, causing economic repercussions that grow week by week. Furthermore, US allies are conspicuously absent, refusing to join Trump’s call to arms.

These compounding issues have boxed the president in. His military might has failed to deliver the swift political results he envisioned, leaving him desperate for a diplomatic off-ramp to mitigate the economic and geopolitical fallout. The bombing, he has learned, doesn’t automatically lead to regime change, and Iran’s resilience means the economic impact continues to snowball.

The Non-Negotiation: A Chasm of Demands

The US “outreach” to Iran has been strikingly limited, primarily taking the form of a two-page document containing 15 items: 12 demands on Iran and three offers. The core of these demands is simple: Iran must abandon any hope of enriching nuclear fuel or developing nuclear weapons, and drastically limit its missile arsenal to a number incapable of reaching Israel. In return, the US dangles the incentive of lifting sanctions, offering Iran a pathway back into the global economy.

Iran’s response, however, was less a counter-proposal and more a non-response. They presented five of their own points, starting with a demand for “compensation for all of the damage you’ve done to Iran’s infrastructure and people over the past month.” They also asserted their sovereignty over the entire Strait of Hormuz, hinting at expanding their practice of charging hefty tolls for ships passing through. This exchange highlights a fundamental disconnect: the two sides are nowhere near agreeing on even the subject of negotiation, let alone the terms.

The Vance Gambit: A Signal of Seriousness

Behind the scenes, the US is planning for eventual negotiations. While the specific participants remain unconfirmed, the administration has floated Vice President JD Vance as the potential lead negotiator. This move is intended to signal seriousness to the Iranians: sending the country’s number two official, empowered to make decisions and with a direct line to President Trump, suggests a genuine desire for high-level talks.

Vance’s potential role is particularly intriguing given his widely reported skepticism about engaging in conflict with Iran from the outset. This perceived opposition to the war could lend him greater credibility in bringing it to an end, both with the Iranians and with the American public. As Sanger observes, it’s a multi-layered message: “The Iranians view him as probably the one they’d most like to talk to because they think he opposed this idea from the start… to President Trump, it’s an opportunity to show to the Iranians that he’s serious… And I think the third message that it sends is actually to the MAGA base. It’s don’t worry, this is under control. JD Vance, who was the most outspoken in favor of staying out of foreign wars, is the one we’ve put in charge of ending this one.”

Israel’s Divergent Agenda: A Partner with Other Priorities

Any discussion of US plans for negotiation with Iran must acknowledge Israel, America’s key partner in the region. However, Israel’s enthusiasm for ending the war may not align with Trump’s. With ground troops in Lebanon and continued assaults on Iran, Israel has a more expansive agenda.

The differing objectives have become increasingly evident. Trump is primarily concerned with the stock market, global oil prices, and the backlash from European and Asian allies. Israel, on the other hand, is largely unconcerned with these factors. For them, the paramount question is whether Iran emerges from the conflict “so defanged that it couldn’t even think of striking beyond its borders again.” While the US seeks a cohesive Iranian government capable of international negotiations, Israel might be perfectly content with “sheer chaos” that could lead to regime overthrow or internal upheaval, effectively neutralizing Iran as a regional threat.

Adding to the complexity, while this diplomatic dance unfolds, Israel continues to strike targets across Iran, including nuclear sites. Iran, in turn, has retaliated, successfully taking out two major American aircraft—an in-air refueling plane and a half-billion-dollar command center—even as messages about negotiations were being exchanged. This simultaneous escalation and discussion underscore the profound distrust and the high-stakes maneuvering at play.

Iran’s Calculated Resistance: Distrust, Unity, and Leverage

Why isn’t Iran eager to accept an end to the daily barrage of attacks that have decimated its military and infrastructure, and even killed its supreme leader? The answer is multifaceted, rooted in deep distrust, strategic calculations, and a desire to maximize leverage.

Broken Trust: “Why should we trust them in the negotiations?” is the question the surviving Iranian leadership likely asks. They point to a history of US and Israeli attacks occurring during periods of negotiation. In June of last year, talks were cut short by attacks on major nuclear sites (Fordo, Natanz, Isvahan). In February, broader agreements were being discussed when a “well-planned, well-coordinated attack on the entire country” was launched. From Iran’s perspective, negotiations have served as a “cover for the US to attack Iran and to buy the time to assemble the forces.” The ongoing deployment of thousands more US troops, including Marine Expeditionary Units, only reinforces this perception, leading Iran to believe the US is “building up to another attack while he’s urging us to negotiate.”

Internal Unity: Beyond distrust, the deeply unpopular Iranian government (with an estimated 80% of Iranians having complaints about economic mismanagement, corruption, and brutality) may see the external attacks as a way to unify a disaffected populace. When foreign actors attack the core of their society, terrorizing ordinary citizens, the leadership can strategically frame itself as the victim, rallying support against a common enemy. In this sense, the longer the war continues, the more likely alienated Iranians might look to their regime as a source of strength against perceived US aggression.

Economic Leverage: Time, from Iran’s perspective, is on their side. They gain leverage by observing “markets slide, watching oil prices go up, watching international anxiety increase, watching the Western Alliance fracture.” The “Trump slump” and rising oil prices only strengthen their position. They believe that if these trends continue, the Trump administration may be forced to relent on some of its more stringent demands, such as the permanent cessation of nuclear fuel enrichment. Iran, a signatory to the nuclear non-proliferation treaty, has consistently argued for its right to enrich uranium for peaceful purposes, a right the US has denied. They hope extended pressure will force the US to accept a deal more akin to the 2015 agreement, allowing for limited enrichment after a pause.

The Nuclear Shadow: An Unintended Consequence

Paradoxically, this entire conflict has likely increased Iran’s incentive to pursue a nuclear weapon. The strategic lesson is clear: nations with nuclear weapons are far less likely to be attacked by the US and Israel. North Korea, possessing 60 or more nuclear weapons and missiles capable of reaching the US, has not faced similar aggression from the Trump administration. As Sanger notes, “The Iranians cannot have missed the point that the Trump administration has not attacked North Korea… And that’s the very reason that the US wouldn’t even consider an attack like this on Pyongyang.” Depending on how this plays out, the US may have inadvertently given Iran every incentive to follow the North Korean model.

Echoes of 2015, A Different Reality

The US and Iran did engage in serious and successful negotiations in 2015, culminating in a historic deal that Iran largely adhered to until President Trump’s withdrawal in 2018. That diplomacy was intended to prevent a war like this, to keep the Iranian nuclear program in a box and provide a year’s notice if Iran raced for a bomb. Trump tore it up, and now the world faces the very conflict it sought to avoid.

However, the current situation is “wildly more complicated” than the 2015 negotiations.

  • Expanded Scope: The 2015 talks were narrowly focused on the nuclear program. Today, the war has unleashed far more issues: Iran’s demonstrated ability to disrupt the global economy and shut down the Strait of Hormuz will now be central to any negotiation.
  • Eroded Trust: While trust was low in 2015, it is “even less” today, given the recent history of attacks during diplomatic overtures.
  • Shifting Commitments: In 2015, Iran committed not to seek nuclear weapons. The current Iranian government, forged in the crucible of this conflict, may find it politically impossible to make such a commitment again.
  • Unleashed Political Forces: War shatters governments and traumatizes populations, unleashing new political forces whose long-term effects are unknown. The conflict could lead to an uprising, as President Trump and Benjamin Netanyahu might have envisioned, or it could trigger the opposite effect—a consolidation of power and further radicalization.

Ultimately, the 2015 negotiations concerned the status of Iran’s nuclear program. Now, the potential negotiations are “over their future survivability as a state.” This existential dimension for the Iranians dramatically amplifies the complexity of bringing this dangerous war to an end.

As the standoff continues, the US has deployed several hundred special operations forces to the Middle East. Iran’s response has been defiant, with the speaker of its parliament warning that if US ground troops enter Iran, they would “set them on fire.” Meanwhile, domestic protests against President Trump’s policies, including his “endless wars,” are gathering momentum across the US, adding another layer of pressure to an already volatile situation. The path to diplomacy, if it exists at all, is fraught with unprecedented challenges.


한국어

“7 free ways to get customers with AI”Greg Isenberg 기반 기사 원본 영상 보기

AI 시대, ‘만들기’만으로는 부족하다: 고객을 사로잡는 7가지 무료 AI 활용 전략

인공지능(AI)의 시대는 전례 없는 창조의 물결을 가져왔습니다. 누구나 AI를 활용해 놀라운 아이디어를 빠르게 구현하고 제품을 ‘바이브 코딩(vibe coding)‘할 수 있게 되었죠. 하지만 단순히 제품을 만드는 것만으로는 성공을 보장할 수 없습니다. 진정한 도전은 바로 ‘배포(distribution)’, 즉 잠재 고객에게 당신의 제품을 알리고 실제 사용자로 전환하는 데 있습니다.

유명 벤처 투자가이자 ‘Startup Ideas’ 팟캐스트의 진행자인 Greg Isenberg는 다가올 10년 동안 가장 부유한 사람들은 마케터가 될 것이라고 단언합니다. 모두가 AI로 제품을 만드는 방법에만 집중할 때, 그는 고객을 확보하는 것이야말로 새로운 시대의 ‘해자(moat)‘이자 핵심 경쟁력이 될 것이라고 강조합니다. 이 글에서는 Greg Isenberg가 제시하는, AI를 활용하여 고객을 무료로 확보할 수 있는 7가지 실용적인 전략을 깊이 있게 다룹니다.

패러다임의 전환: 배포가 새로운 해자다

Greg Isenberg는 실리콘밸리의 가치 체계가 급격히 변화하고 있다고 말합니다. 그가 2014년 실리콘밸리에 처음 왔을 때만 해도 엔지니어가 최상위 계층이었고, 그다음이 제품, 그리고 마케팅은 가장 아래, 심지어 ‘웃음거리(laughingstock)‘로 여겨졌습니다. 하지만 AI의 등장으로 이 순서가 완전히 뒤바뀌었습니다. 이제는 배포(고객 확보) 전문가가 최상위에 있고, 그다음이 제품 전문가, 그리고 개발자 순입니다.

이러한 변화는 배포에 대한 이해가 극히 드물기 때문에 더욱 중요합니다. 많은 사람이 브랜딩, 광고, 콘텐츠를 제대로 이해하지 못합니다. 하지만 이 분야를 이해하는 사람은 ‘불공평한 이점(unfair advantage)‘을 얻게 됩니다. 매일 수십만 개의 새로운 ‘바이브 코딩’ 프로젝트가 탄생하지만, 실제로 그 제품을 보는 사람은 극히 적습니다.

**피터 레벨스(Peter Levels)**의 사례는 이를 명확히 보여줍니다. 그는 300만 달러 이상의 매출을 올리면서도 직원이 단 한 명도 없는 1인 기업가입니다. 그 비결은 바로 75만 명 이상의 팔로워와 뛰어난 검색 엔진 최적화(SEO) 덕분입니다. Nomad List와 같은 그의 도구들은 기술적으로는 쉽게 모방할 수 있지만, 그가 구축한 오디언스, 신뢰, 그리고 SEO는 쉽게 따라 할 수 없는 해자 역할을 합니다.

많은 창업자들은 제품을 훌륭하게 만들면 고객이 저절로 찾아올 것이라는 환상에 빠집니다. 하지만 현실은 ‘침묵’뿐입니다. 그들은 더 많은 기능을 추가하고 다시 출시하지만, 여전히 침묵만 흐릅니다. 반면, 현명한 빌더들은 배포에서 시작합니다. 그들은 먼저 1,000명 정도의 오디언스를 구축하고, 그들에게 필요한 것이 무엇인지 묻습니다. 그리고 주말이나 며칠 안에 제품을 만들어 출시하고, 따뜻한(warm) 오디언스와 함께 제품을 개선하며 수익을 창출합니다. 핵심은 **“배포 우선, 제품은 그다음(distribution first, product second)“**입니다.

이제 AI 시대에 고객을 확보할 수 있는 7가지 구체적인 전략을 살펴보겠습니다.


1. MCP 서버를 영업팀으로 활용하기 (MCP Server as Your Sales Team)

개념: AI 어시스턴트가 당신의 제품을 자동으로 발견하고 고객에게 추천하며 판매하도록 만드는 전략입니다. MCP(Multi-Modal Communication Protocol)는 OpenAI 플러그인과 유사하게, 다양한 AI 모델이 정보를 교환하고 기능을 확장할 수 있도록 돕는 프로토콜입니다.

작동 방식:

  1. 사용자가 AI(예: Claude, ChatGPT)에 질문합니다.
  2. AI가 당신의 MCP 서버를 발견합니다.
  3. AI가 당신의 제품을 사용자에게 추천하고, 사용자는 이를 통해 가치를 얻습니다.

핵심 이점:

  • 제로 고객 획득 비용(Zero CAC - Customer Acquisition Cost): AI 어시스턴트가 당신의 영업팀이 되어 24시간 7일 내내 제품을 판매하므로, 마케팅 비용이 들지 않습니다.
  • 선점 효과: Greg Isenberg는 2026년 MCP 서버 구축이 2010년 모바일 앱을 구축하는 것과 같다고 비유합니다. 아직 초기 단계이므로, LLM(Large Language Model) 기반의 AI 어시스턴트 연결 및 플러그인 시장에서 엄청난 기회가 있습니다.

이번 주에 시작하는 방법:

  1. 제품의 핵심 질문 식별: 당신의 제품이 어떤 질문에 답을 제공하는지 명확히 정의하세요.
  2. MCP 서버 구축: 24시간 이내에 당신의 데이터를 반환하는 MCP 서버를 ‘바이브 코딩’하세요.
  3. MCP 레지스트리에 게시: Smithery, MCPT, Open Tools와 같은 MCP 레지스트리에 서버를 등록하세요.
  4. 자동 판매 시작: 이제 당신의 서버에 연결된 모든 AI 어시스턴트가 당신의 제품을 24/7 판매하기 시작할 것입니다.

2. 프로그래매틱 SEO (Programmatic SEO)

개념: AI를 활용하여 수만 개의 검색 엔진 최적화(SEO) 페이지를 자동으로 생성하고, 이를 통해 대규모 유기적 트래픽을 확보하는 전략입니다.

작동 방식:

  1. 키워드 패턴 분석: “최고의 X를 위한 Y”와 같은 키워드 패턴을 찾습니다 (예: “치과 의사를 위한 최고의 CRM”).
  2. 데이터 세트 구축: Firecrawl과 같은 도구를 사용하여 웹에서 관련 데이터를 스크래핑하고 구조화된 데이터를 확보합니다. 기존 데이터베이스를 활용할 수도 있습니다.
  3. 페이지 템플릿 생성: Next.js, Webflow, WordPress와 같은 프레임워크를 사용하여 페이지 템플릿을 만듭니다.
  4. AI 콘텐츠 생성: AI를 사용하여 각 페이지에 고유한 콘텐츠를 생성합니다. 단순히 변수만 바꾸는 것이 아니라, 실제 단락과 고품질의 내용을 만듭니다.
  5. 대규모 페이지 생성: 이 과정을 통해 수천, 수만 개의 SEO 페이지를 생성합니다.

핵심 이점:

  • 대규모 트래픽 잠재력: 페이지당 월 30회 방문이라는 적은 수치라도, 10,000개의 페이지가 있다면 월 300,000명의 방문자를 확보할 수 있습니다. 2%의 전환율과 $10의 전환 가치를 가정하면 월 $60,000의 수익을 창출할 수 있습니다.
  • 자동화된 성장: 한 번 구축하면 지속적으로 트래픽과 리드를 생성하는 강력한 성장 엔진이 됩니다.
  • 인간의 개입: AI 콘텐츠가 부자연스럽게 느껴지지 않도록 최적화하고, 필요에 따라 인간 검토자가 콘텐츠를 편집해야 합니다. 처음에는 소수의 페이지로 시작하여 품질을 확보한 후 확장하는 것이 중요합니다.

이번 주에 시작하는 방법:

  1. 키워드 패턴 선정: “틈새시장을 위한 제품 유형” 또는 “도시의 서비스”와 같은 키워드 패턴을 선택하세요 (예: “마이애미의 지붕 공사”).
  2. 데이터 세트 구축: Firecrawl로 데이터를 스크래핑하거나 기존 데이터베이스를 활용하세요.
  3. 템플릿 생성: Next.js, Webflow, WordPress 등 익숙한 프레임워크로 템플릿을 만드세요.
  4. AI로 고유 콘텐츠 생성: AI를 사용하여 페이지별로 고유하고 고품질의 콘텐츠를 생성하고, 필요시 인간 검토를 거치세요.
  5. MVP 출시 및 모니터링: 100페이지를 MVP(최소 기능 제품)로 게시하고, 검색 엔진 인덱싱을 모니터링한 후 확장합니다.

3. 무료 도구(Free Tool)를 퍼널 상단으로 활용하기

개념: 사용자가 즉각적인 가치를 얻을 수 있는 무료 AI 기반 도구를 만들어, 유료 제품으로의 전환을 유도하는 마케팅 전략입니다. ‘도구 자체가 마케팅’이 됩니다.

작동 방식:

  1. 무료 도구 구축: 평가기(grader), 분석기(analyzer), 계산기(calculator) 등 사용자의 문제를 해결하거나 통찰력을 제공하는 AI 기반 무료 도구를 만듭니다.
  2. 즉각적인 가치 제공: 사용자가 도구를 통해 즉시 결과(예: “당신의 웹사이트 점수는 100점 만점에 43점입니다”)를 얻게 합니다. 이때 사용자의 이메일이나 전화번호를 수집하는 것이 좋습니다.
  3. 사용자 점수 공유: 사용자가 자신의 점수나 결과를 소셜 미디어에 공유하도록 유도하여 바이럴 효과를 만듭니다.
  4. 더 많은 사용자 유입: 공유를 통해 더 많은 사용자가 도구를 발견하고 유입됩니다.
  5. 유료 제품으로 업셀링: 도구가 발견한 문제점(예: “이러한 문제를 해결하고 싶다면, 저희 유료 제품을 사용해보세요”)을 해결할 수 있는 유료 제품으로 자연스럽게 전환을 유도합니다.

핵심 이점:

  • AI로 빠른 개발: 과거에는 몇 달이 걸렸을 무료 도구 개발이 이제 클라우드 코드(cloud code)와 AI 덕분에 하루 만에 구축하고 점심시간에 배포할 수 있을 정도로 빨라졌습니다.
  • 지속적인 마케팅: 한 번 구축하면 도구 자체가 끊임없이 마케팅을 수행합니다.
  • 강력한 리드 생성: 사용자에게 가치를 제공하면서 잠재 고객의 데이터를 확보할 수 있습니다.

이번 주에 시작하는 방법:

  1. 아이디어 브레인스토밍: 당신의 LLM(거대 언어 모델)에 당신의 제품/서비스에 대한 10가지 무료 도구 아이디어를 요청하세요.
  2. 우선순위 선정 및 오디언스 투표: 가장 유망한 아이디어를 선택하고, 가능하다면 오디언스에게 어떤 도구가 가장 흥미로운지 물어보세요.
  3. 구축 및 배포: 선택한 도구를 빠르게 구축하고 배포하세요. 마치 편집 캘린더처럼 ‘무료 도구 캘린더’를 만들어 지속적으로 새로운 도구를 출시하는 것을 고려해보세요.

4. 답변 엔진 최적화(AEO - Answer Engine Optimization)

개념: AI 챗봇(예: ChatGPT, Perplexity)이 질문에 답할 때 당신의 웹사이트를 출처로 인용하도록 콘텐츠를 최적화하는 전략입니다. 기존 SEO가 구글 1페이지 노출에 집중했다면, AEO는 AI가 직접 답변을 제공하는 ‘제로 클릭 검색(zero-click searches)’ 시대에 맞춰 AI의 ‘인용원(source)‘이 되는 것을 목표로 합니다.

작동 방식:

  1. 구조화된 직접 답변: AI가 쉽게 파싱하고 이해할 수 있도록 명확하고 간결하며 구조화된 답변을 제공합니다.
  2. FAQ 형식 및 스키마 마크업: 자주 묻는 질문(FAQ) 형식과 스키마 마크업(schema markup)을 사용하여 콘텐츠의 맥락을 AI에 명확히 전달합니다.
  3. 비교표 및 테이블: AI가 데이터를 추출하기 쉬운 비교표나 테이블 형식을 활용합니다.
  4. 권위 있는 도메인에 게시: 검색 엔진 및 AI로부터 신뢰를 얻을 수 있는 권위 있는 도메인에 콘텐츠를 게시하거나, 다른 전략을 통해 도메인 권위를 구축합니다.

핵심 이점:

  • AI의 직접 추천: AI가 사용자에게 당신의 제품이나 웹사이트를 직접 추천하여 트래픽을 유도합니다.
  • 빠른 성장: 피터 레벨스의 경우 AI 추천 트래픽이 한 달 만에 4%에서 20%로 급증했다고 합니다. 이는 전자상거래, SaaS, 앱 등 모든 분야에서 증가할 추세입니다.
  • 선점 효과: AEO는 2026년의 SEO와 같으며, 먼저 시작하는 기업이 해당 틈새시장을 몇 년 동안 장악할 수 있습니다.

이번 주에 시작하는 방법:

  1. 고객 질문 20가지 구글링: 고객이 가장 많이 묻는 상위 20가지 질문을 구글에서 검색하세요.
  2. “인용 가치 있는” 답변 작성: 각 질문에 대해 3,000단어의 불필요한 내용이 아닌, 명확하고 직접적이며 AI가 인용할 만한 답변을 작성하세요.
  3. 스키마 마크업 및 FAQ 블록 추가: 콘텐츠에 스키마 마크업과 FAQ 블록을 적용하세요.
  4. 권위 있는 도메인에 게시: 권위 있는 도메인에 콘텐츠를 게시하거나, 꾸준히 도메인 권위를 구축하세요.
  5. AI 인용 모니터링: Otterly, Profound와 같은 도구나 수동 테스트를 통해 Perplexity 및 ChatGPT에서 당신의 사이트가 얼마나 인용되는지 모니터링하세요.

5. 바이럴 아티팩트(Viral Artifacts) - 제품의 결과물을 공유 가능하게 만들기

개념: 사용자가 자신의 정체성이나 성취를 나타내는, 제품에서 생성된 ‘아티팩트(artifact)‘를 자랑스럽게 공유하도록 만들어 바이럴 효과를 유도하는 전략입니다.

핵심 질문: “내 사용자는 무엇을 자랑하고 싶어 할까?” 그리고 그 ‘무엇’을 아름답고 공유 가능하게 만드세요.

사례:

  • Spotify Wrapped: 매년 1억 회 이상 공유되는 연말 결산으로, 사용자의 음악 취향을 보여주며 정체성의 일부가 됩니다.
  • GitHub 기여 그래프: 개발자들이 녹색 사각형으로 이루어진 기여 그래프를 트위터에 공유하며 자신의 코딩 활동을 자랑합니다.
  • Stripe Atlas: 법인 설립 마일스톤을 기념하는 게시물로, 창업자들이 회사를 설립한 지 몇 년이 되었는지 공유합니다.
  • Duolingo: 연속 학습 기록(daily streak count)을 통해 사용자들이 학습 성과를 자랑합니다.

작동 방식:

  1. 공유할 만한 결과물/마일스톤 식별: 사용자가 스크린샷하거나 공유하고 싶어 할 만한 제품의 출력물이나 중요한 성취 지점을 찾아냅니다.
  2. 아름답고 브랜드화된 디자인: 해당 아티팩트를 시각적으로 매력적이고, 당신의 브랜드가 미묘하게 드러나도록 디자인합니다. 로고는 너무 크지 않게, 사용자가 자신을 드러내는 느낌을 주도록 만드세요.
  3. 공유 버튼 추가: 아티팩트와 함께 공유 버튼을 추가하여, 사용자가 클릭 한 번으로 미리 채워진 게시물과 함께 공유할 수 있도록 합니다.
  4. 무료 노출: 사용자의 공유는 당신의 정확한 타겟 오디언스에게 무료 노출로 이어집니다. 사용자가 스스로를 자랑하면서 당신의 마케팅을 대신 해주는 셈입니다.

B2B 적용: B2B 제품에서도 슬랙(Slack)이나 마이크로소프트 팀즈(Microsoft Teams)와 같은 내부 채널에서 팀원들이 공유하고 싶어 할 만한 성과 지표나 보고서를 만들 수 있습니다.

이번 주에 시작하는 방법:

  1. 핵심 출력물/마일스톤 식별: 당신의 사용자가 스크린샷하여 공유할 만한 제품의 결과물이나 중요한 이정표를 찾아내세요.
  2. 디자인 및 브랜딩: 해당 결과물을 아름답고 공유 가능하며, 당신의 브랜드가 미묘하게 드러나도록 디자인하세요.
  3. 공유 기능 추가: 결과물과 함께 공유 버튼을 추가하여, 사용자가 손쉽게 게시물을 공유할 수 있도록 만드세요.

6. 틈새 뉴스레터(Niche Newsletter) 인수하기

개념: 맨땅에서 오디언스를 구축하는 대신, 특정 틈새시장의 기존 뉴스레터를 인수하여 즉시 신뢰와 독자를 확보하는 전략입니다.

기존 오디언스 구축의 어려움:

  • 시간 소모: 몇 년이 걸릴 수 있습니다.
  • 꾸준한 콘텐츠: 매일 콘텐츠를 발행해야 합니다.
  • 성공 보장 없음: 노력해도 원하는 결과를 얻지 못할 수 있습니다.
  • 제로 구독자: 시작은 언제나 0명부터입니다.

뉴스레터 인수의 이점:

  • 즉각적인 신뢰 확보: 첫날부터 구축된 오디언스의 신뢰를 물려받습니다.
  • 제품 연결: 인수 즉시 당신의 제품을 오디언스에게 소개하고 판매할 수 있습니다.
  • 반복 가능성: 여러 틈새시장에서 이 전략을 반복하여 적용할 수 있습니다.
  • 낮은 인수 비용: 많은 소규모 뉴스레터는 수익화가 잘 되지 않아, 1만 명의 구독자를 가진 뉴스레터를 $5,000~$20,000에 인수할 수 있습니다.
  • 안정적인 채널: 소셜 미디어 계정처럼 언제든 사라질 수 있는 플랫폼과 달리, 뉴스레터는 당신이 소유한 직접적인 채널이므로 도달률(reach)이 제한될 염려가 없습니다.

이번 주에 시작하는 방법:

  1. 뉴스레터 검색: Duuce.com, Newsletter Investor와 같은 뉴스레터 매매 플랫폼을 방문하거나, 트위터나 서브스택(Substack)에서 당신의 틈새시장에 맞는 뉴스레터를 검색하세요.
  2. 타겟 뉴스레터 선정: 5,000명에서 50,000명 사이의 구독자를 가진 뉴스레터 중 당신의 타겟 틈새시장에 맞는 것을 찾아보세요.
  3. 소유자에게 DM: 뉴스레터 소유자에게 “뉴스레터 판매에 대해 생각해본 적이 있나요?”라고 DM을 보내세요. 놀랍게도 많은 사람이 대화에 응할 것입니다.
  4. 합리적인 제안: 대부분의 소규모 뉴스레터는 월 0~$500 정도의 수익을 올리므로, $5,000~$20,000의 합리적인 제안에 긍정적으로 반응할 수 있습니다.

7. AI 콘텐츠 재활용 엔진(AI Content Repurposing Engine) 구축

개념: 하나의 ‘기둥(pillar) 콘텐츠’ (예: 긴 팟캐스트 에피소드, 유튜브 비디오, 긴 블로그 게시물)를 AI를 활용하여 여러 형식의 콘텐츠로 재활용하고, 다양한 플랫폼에 배포하여 접점(touch points)을


“How to turn Claude code into your personal life operating system | Hilary Gridley”How I AI 기반 기사 원본 영상 보기

클로드 코드(Claude Code)로 나만의 ‘안티-시스템’ 구축하기: 바쁜 워킹맘을 위한 AI 기반 시간 관리

새로운 기술의 물결 속에서, AI는 우리의 삶과 일하는 방식을 근본적으로 변화시키고 있습니다. 특히 시간 관리와 생산성 측면에서 AI는 단순한 도구를 넘어 개인 비서이자 삶의 운영 체제(Life Operating System)로 진화하고 있습니다. ‘How I AI’ 채널의 인기 에피소드에서, 기업가이자 워킹맘인 힐러리 그리들리(Hilary Gridley)는 클로드 코드(Claude Code)를 활용해 자신만의 ‘안티-시스템 시스템(anti-system system)‘을 구축하여 바쁜 일상 속에서 시간과 에너지를 되찾는 방법을 공유했습니다. 그녀의 이야기는 복잡한 설정 없이도 AI를 통해 삶의 질을 향상시킬 수 있다는 강력한 메시지를 전달합니다.

1. 시간의 기회비용, 육아 그리고 AI

힐러리는 새로운 엄마이자 기업가로서 “시간의 기회비용이 그 어느 때보다 높아졌습니다. 동시에 그 어느 때보다 시간이 부족하다고 느낍니다”라고 고백합니다. 육아와 사업이라는 두 가지 큰 책임 속에서 그녀의 주의는 산만해졌고, 일상적인 행정 업무(life admin)와 물류 관리에 엄청난 에너지를 쏟아야 했습니다. 역사적으로 물류 관리에 서툰 그녀는 수면 부족과 흐릿한 정신 상태 속에서 이 모든 것을 감당하기가 더욱 어려웠다고 말합니다.

힐러리가 AI를 통해 얻고자 하는 것은 단순한 ‘초생산성 최적화’가 아닙니다. 오히려 “시간을 되찾아 시장에 가서 바게트를 사고 맛있는 치즈를 먹으며 삶을 즐기는 것”입니다. 그녀는 “내 안에는 두 마리의 늑대가 있는데, 둘 다 행정 업무는 아니다”라며, 행정 업무를 처리하는 데 드는 시간과 노력을 최소화하여 진정으로 의미 있는 일과 삶의 즐거움에 집중하고 싶다고 강조합니다.

하지만 많은 AI 시스템이나 생산성 도구들이 복잡한 설정과 지속적인 유지 관리를 요구한다는 점이 문제였습니다. 힐러리는 “시스템을 유지하는 데 노력을 들이고 싶지 않다. 많은 설정을 하고 싶지 않다. 그냥 시작해서 문제를 해결하고 싶다”고 말하며, 이러한 복잡성 자체를 거부하는 ‘안티-시스템 시스템’ 철학을 제시합니다.

2. ‘안티-시스템 시스템’의 철학: 단순함이 핵심

힐러리는 개인 생산성 스펙트럼을 “노션 보이(Notion boys)“와 “알 고어(Al Gore)의 데스크톱”으로 비유합니다. 노션 보이들은 테이블, 페이지, 연결된 자산, 칸반 보드 등 복잡한 시스템을 선호하는 반면, 알 고어의 데스크톱은 극도로 혼란스러워 보이지만, 그 역시 엄청난 성과를 이뤄낸 사람의 예시입니다. 힐러리는 자신을 후자에 가깝다고 설명하며, “의식의 흐름(stream of consciousness)“에 따라 하루를 구조화하는 방식을 선호한다고 말합니다.

그녀의 철학은 다음과 같습니다:

  • 복잡성은 그 가치를 증명해야 한다(Complexity has to earn its keep): 처음부터 모든 것을 복잡하게 구축하기보다, 가장 간단한 버전으로 시작하여 가치를 확인한 후 점진적으로 기능을 추가하고 개선해나갑니다.
  • ‘야퍼스 API(Yappers API)‘와 스크린샷 활용: 구글 드라이브나 캘린더에 복잡한 API 연동을 구축하기보다, 화면을 보며 AI에게 직접 말하거나 스크린샷을 찍어 정보를 제공하는 방식을 선호합니다. “블로그 게시물을 작성한 후 클로드에게 ‘블로그 게시물을 작성했어’라고 말하는 식이죠.” 이는 기술 사용의 마찰을 최소화하는 가장 효과적인 방법입니다.
  • AI를 ‘직원’처럼 대하기: 힐러리는 AI 에이전트를 직원처럼 대하며, 그들이 어떻게 일을 처리하는지는 크게 신경 쓰지 않습니다. “폴리(가상의 AI 에이전트)가 일을 처리하는 방식은 내 알 바가 아닙니다. 그녀가 일을 처리하기만 하면 되죠.” 2% 정도의 오류율은 괜찮으며, AI가 제공하는 인지적 부담(cognitive effort) 경감 효과가 훨씬 크다고 말합니다.

3. 클로드 코드(Claude Code)로 하루 계획하기

힐러리는 클로드 코드(Claude Code)를 활용하여 하루를 계획하고 관리하는 구체적인 단계를 시연했습니다. 클로드 코드(Claude Code)는 터미널(terminal)에서 실행되는 AI 에이전트로, 컴퓨터가 할 수 있는 거의 모든 작업을 수행할 수 있습니다.

1단계: 아이디어 및 할 일 포착 가장 먼저 할 일은 “내 뇌의 새는 체(leaky sieve)를 통해 아무것도 빠져나가지 않도록 하는 것”입니다.

  • 아이폰 백탭 단축키(iPhone Backtap shortcut): 아이폰 잠금 화면에서 휴대폰 뒷면을 두 번 탭하면 받아쓰기(dictation) 상자가 나타납니다. 여기에 “소아과 진료 예약 변경”과 같은 할 일을 말하면, 해당 내용이 미리 알림(Reminders) 앱의 받은 편지함에 자동으로 추가됩니다. 이는 AI가 개입하기 전, 마찰 없는 정보 포착의 첫 단계입니다.

2단계: 클로드 코드 설정 및 실행

  • 클로드 코드 설치: 클로드 코드는 설치가 매우 간단합니다. 터미널을 열고 설치 명령어를 붙여넣기만 하면 됩니다.
  • 하루 계획 명령: 힐러리는 터미널에 plan my day라고 입력하고 엔터키를 누릅니다. 그러면 클로드 코드가 작동하여 하루 계획을 시작합니다.

3단계: AI 기반 일정 제안 및 세분화 클로드는 다음과 같은 정보들을 종합하여 하루 일정을 제안합니다.

  • 미리 알림(Reminders): 힐러리의 잠금 화면에서 포착된 미리 알림 목록을 클로드가 백그라운드에서 특정 문서(마크다운 파일)로 정리하여 관리합니다.
  • 선호도(Preferences): 클로드는 힐러리의 실제 행동을 관찰하며 그녀의 선호도(예: 모유 유축 시간, 주말 육아 시간, 특정 작업 후 몰입하면 잠을 못 자는 경향 등)를 학습하고 기록합니다. 힐러리가 직접 입력한 선호도가 아닌, AI가 관찰을 통해 파악한 선호도라는 점이 중요합니다.
  • 캘린더(Calendar): 이미 캘린더에 있는 일정을 파악하여 비어 있는 시간을 확인합니다.

이러한 정보들을 바탕으로 클로드는 힐러리에게 “아기 여권 진행”과 같은 큰 과제를 제시하고, 이를 “우체국 예약하기”와 같은 10~15분짜리 작은 첫 단계로 세분화하여 제안합니다. 이는 힐러리가 평소에 큰 과제를 미루는 경향이 있다는 것을 AI가 학습했기 때문입니다. 이렇게 세분화된 작은 단계는 바쁜 일상 속에서 과제를 완수할 가능성을 크게 높여줍니다. 클로드는 캘린더에 자동으로 이 일정을 추가하고, 귀여운 ‘하마 이모티콘’을 사용하여 힐러리가 일정을 쉽게 파악하도록 돕습니다.

4. 실행과 회고: 학습하는 AI

힐러리는 하루 종일 컴퓨터 화면의 오른쪽 3분의 1에 클로드 코드 터미널을 열어두고, 자신이 하는 일들을 클로드에게 말해줍니다. 클로드는 이를 관찰하고 메모하여 ‘일일 노트(Daily Note)‘라는 마크다운 파일에 기록합니다. 이 노트에는 계획된 일정과 실제로 수행된 작업 로그가 함께 담겨 있습니다.

  • 성찰 및 패턴 파악: 힐러리는 주기적으로 클로드에게 “내가 하려는 일과 실제로 하는 일 사이의 격차에 대해 무엇을 관찰하고 있니?” 또는 “어떤 패턴을 발견했니?”라고 질문합니다. 그러면 클로드는 “구축 작업이 글쓰기를 방해하고 있다”거나, “시간 관리를 잘 못하고 있다”, “세 가지 우선순위를 정하지만, 첫 번째만 제대로 처리된다”와 같은 통찰을 제공합니다.
  • ‘배우면서 실행하기(Learn by doing)’ 원칙: 힐러리는 이러한 피드백을 통해 자신의 계획과 실제 행동 사이의 불일치를 파악하고, AI 시스템이 스스로를 개선하도록 돕습니다. “만약 이 모든 것을 처음부터 계획했더라면 대부분 틀렸을 것”이라고 말하며, AI가 실제 행동을 관찰하며 학습하기 때문에 시스템 유지 및 설정 비용이 ‘제로’에 가깝다고 강조합니다.

5. 무엇을 자동화할 것인가? 10배 임팩트 프레임워크

힐러리는 AI를 통해 무엇을 자동화할지 결정하는 데 있어 “10배 임팩트 프레임워크(10x Impact Framework)“를 제시합니다.

질문: “어떤 작업에서 내가 10배 더 잘한다면, 10배 더 큰 영향력을 가질 수 있을까?”

  • 답이 ‘아니오’라면: 그 작업은 자동화합니다. (예: 파워포인트 슬라이드에서 픽셀을 옮기는 작업)
  • 답이 ‘예’라면: 그 작업에 더 많은 시간과 노력을 투자합니다. (예: 사용자 연구에서 중요한 통찰을 도출하여 제품 의사결정에 기여하는 작업)

이 프레임워크는 단순히 업무 맥락을 넘어 삶의 전반에 적용됩니다. 힐러리는 “우리는 기쁨을 만들어내야 하고, 인간으로서 살아가는 것을 즐겁게 만드는 모든 것을 적극적으로 보호해야 한다”고 말합니다. 빵 굽는 일처럼 자신을 풍요롭게 하는 일이라면 직접 투자하고, 단순한 잡무라면 자동화하는 식입니다. 심지어 하나의 작업 내에서도 아이디어 구상(인간의 영역)과 슬라이드 제작(자동화 가능)처럼 세분화하여 적용할 수 있습니다.

또한, 이 프레임워크는 학습 곡선(learning curve)에 따라 달라질 수 있습니다. 힐러리에게는 슬라이드 제작이 더 이상 가치 없는 반복 작업일 수 있지만, 팀원에게는 숙련도를 높이는 중요한 학습 기회가 될 수 있습니다. 따라서 AI 자동화는 개인의 기술 수준과 경력 단계에 맞춰 전략적으로 접근해야 합니다.

6. 라이트닝 라운드: ‘녹화 모드’ 스킬

힐러리는 마지막으로 라이트닝 라운드에서 라이브 데모 중 개인 정보 노출을 방지하기 위해 개발한 ‘녹화 모드(Recording Mode)’ 스킬을 소개했습니다. 이는 AI가 문제 해결의 ‘고도(altitude)‘를 완전히 바꿀 수 있다는 것을 보여주는 예시입니다. 그녀는 “개인 정보가 가려진(blinded) 컨텍스트 디렉토리 사본을 만드는 것”을 통해 데모 중에도 안전하게 개인 워크플로우를 시연할 수 있게 되었습니다. 이는 AI가 단순히 기존 작업을 빠르게 처리하는 것을 넘어, 완전히 새로운 방식으로 문제를 해결할 수 있게 해준다는 점을 강조합니다.

결론

힐러리 그리들리의 ‘안티-시스템 시스템’은 AI를 통해 삶의 복잡성을 줄이고, 시간을 되찾아 진정으로 소중한 일에 집중할 수 있음을 보여줍니다. 복잡한 설정에 매달리기보다, 단순한 시작과 ‘배우면서 실행하기’의 철학, 그리고 ‘10배 임팩트 프레임워크’를 통해 무엇을 자동화하고 무엇에 인간적인 노력을 쏟을지 현명하게 결정하는 것이 핵심입니다. 클로드 코드와 같은 AI 에이전트를 활용하여 개인의 삶을 위한 운영 체제를 구축하는 것은 더 이상 먼 미래의 이야기가 아니라, 지금 바로 시작할 수 있는 실용적인 현실이 되고 있습니다.


“Trump Says He’s Ready for Diplomacy. Iran? Not So Much.”New York Times Podcasts 기반 기사 원본 영상 보기

트럼프의 ‘외교’ 선언, 이란은 왜 냉담한가? 중동 갈등의 딜레마

뉴욕타임스의 인기 팟캐스트 ‘더 데일리(The Daily)‘는 최근 중동을 둘러싼 미국과 이란의 복잡한 외교적 대치 상황을 심층 분석했습니다. 마이클 바바로(Michael Barbaro) 진행자는 뉴욕타임스 기자 데이비드 생어(David Sanger)와 함께 도널드 트럼프 미국 대통령이 외교적 해결을 모색하고 있음에도 불구하고, 이란이 협상 테이블에 앉기를 주저하는 이유를 파고들었습니다. 겉으로는 외교의 문이 열린 듯 보이지만, 양측의 불신과 전략적 계산이 얽히면서 상황은 그 어느 때보다 복잡하게 전개되고 있습니다.

트럼프의 ‘외교’ 구상과 현실의 벽

트럼프 대통령은 확전 위협에도 불구하고 이란과의 갈등을 협상을 통해 끝내려는 의지를 강하게 내비치고 있습니다. 지난주, 그는 “미국과 이란은 중동에서의 적대 행위를 완전히 해결하기 위한 매우 훌륭하고 생산적인 대화를 나눴다”고 공개적으로 선언하며 협상이 진행 중임을 암시했습니다. 하지만 불과 몇 시간 뒤, 이란 정부는 “도널드 트럼프 미국 대통령이 미국과 이란 이슬람 공화국 간의 어떠한 협상에 대해서도 언급한 내용을 부인한다”는 성명을 발표하며 정면으로 반박했습니다.

데이비드 생어 기자는 이에 대해 “두 쪽 모두 어느 정도 ‘거짓말’을 하고 있다”고 분석합니다. 트럼프 대통령은 11,000개에 달하는 이란 목표물에 대한 무력 시위(brute strength)를 통해 이란을 협상 테이블로 끌어냈다는 인상을 주려 합니다. 반면 이란은 국가적 자존심과 현재의 전략적 판단으로 인해 미국의 조건에 따라 협상 테이블에 앉지 않을 것임을 보여주려 한다는 것입니다. 결국 양측은 전쟁을 외교로 전환하는 시점을 놓고 미묘한 대치 상태에 놓여 있습니다.

왜 트럼프는 ‘외교적 출구’를 찾는가?

트럼프 대통령이 외교적 해결책을 모색하는 배경에는 여러 현실적인 문제가 얽혀 있습니다. 미국은 이란 내 11,000개 목표물을 타격했음에도 불구하고, 군사력의 한계뿐만 아니라 예상치 못한 역풍(blowback)과 그로 인한 문제점들을 발견했습니다.

  • 정권 교체 실패: 이란 정권은 크게 약화되었지만 붕괴 직전은 아닙니다. 현재로서는 정권 교체(regime change)가 임박하지 않았습니다.
  • 경제적 파급 효과: 전쟁으로 호르무즈 해협(Strait of Hormuz)이 봉쇄되고 글로벌 에너지 산업이 마비되면서 경제적 파급 효과는 매주 커지고 있습니다.
  • 동맹국의 외면: 유럽과 아시아 동맹국들은 트럼프 대통령의 전쟁 참여 요청에 응답하지 않고 있습니다.
  • 국내외 압력: 주식 시장의 하락, 유가 상승, 국제적 불안감 증가는 트럼프 대통령에게 정치적 부담으로 작용하고 있습니다. 특히 미국 내에서는 ‘트럼프 슬럼프(Trump slump)’ 또는 ‘이란 슬럼프(Iran slump)‘라는 용어까지 등장하며 경제적 우려가 커지고 있습니다.

생어 기자는 “트럼프 대통령은 자신이 예상했던 정치적 결과를 가져오지 못했으며, 특히 그가 염두에 두었던 기간 내에는 더욱 그러하다는 것을 발견했다”고 지적합니다. 폭격만으로는 정권 교체를 이끌어낼 수 없고, 이란이 현명하게 대응하면 경제적 파급 효과는 계속 눈덩이처럼 불어날 것이라는 분석입니다.

협상 테이블로 가는 길: 미국의 제안과 이란의 거부

미국은 이란을 협상 테이블로 이끌기 위해 ‘15개 항목으로 구성된 두 페이지짜리 문서’를 제시했습니다. 이 중 12개는 이란에 대한 요구 사항이었고, 3개는 이란에 대한 제안이었습니다. 핵심 내용은 다음과 같습니다.

  • 미국의 요구: 이란이 핵연료 농축에 대한 모든 희망을 포기하고, 핵무기로 이어질 수 있는 모든 단계를 중단하며, 미사일 무기고를 이스라엘에 도달할 수 없는 극히 제한된 수로 제한할 것.
  • 미국의 제안: 이란이 글로벌 국제 경제에 재진입할 수 있는 길을 열어주는 제재 해제.

그러나 이란의 반응은 사실상 ‘비(非)반응’이었습니다. 이란은 5가지 자체 제안을 내놓았는데, 주로 지난 한 달간 이란의 인프라와 국민들에게 가한 모든 피해에 대한 보상과 호르무즈 해협 전체에 대한 이란의 주권 인정을 요구하는 내용이었습니다. 생어 기자는 “이는 양측이 무엇을 협상할지에 대해서조차 합의하지 못하고 있음을 시사한다”고 평가합니다.

미국은 그럼에도 불구하고 협상 가능성에 대비하고 있으며, 제이디 밴스(JD Vance) 부통령이 협상을 이끌 것이라고 밝혔습니다. 이는 이란에게 진지한 대화에 임하고 있음을 알리는 신호로 해석됩니다. 밴스 부통령은 전쟁 회의론자로 알려져 있어, 그의 참여는 이란에게 신뢰를 줄 수 있고, 동시에 트럼프 대통령의 핵심 지지층인 마가(MAGA) 지지층에게는 ‘외국 전쟁 개입에 반대했던 밴스가 이 전쟁을 끝내는 책임자가 되었다’는 메시지를 전달하는 효과도 있을 것이라는 분석입니다.

이란은 왜 협상을 주저하는가? 불신과 전략적 계산

트럼프 대통령은 이란이 왜 미국의 제안을 선뜻 받아들이지 않는지 의아해하지만, 이란의 입장은 복합적인 불신과 전략적 계산에 기반하고 있습니다.

  1. 미국에 대한 뿌리 깊은 불신:

    • 협상 중 공격 이력: 지난해 6월, 이란은 핵 시설(포르도, 나탄즈, 이스파한) 공격으로 인해 협상이 중단된 경험이 있습니다. 지난 2월에도 광범위한 합의를 위한 협상이 진행되던 중, 미국과 이스라엘이 이란 전역에 대한 잘 계획된 공격을 감행했습니다.
    • ‘트로이 목마’ 전략: 이란은 협상이 이란을 공격하고 병력을 집결할 시간을 벌기 위한 미국의 위장술(Trojan horse)일 뿐이라고 생각합니다. 현재 미국이 수천 명의 병력을 더 이 지역으로 보내고 있는 것도 이란의 관점에서는 또 다른 공격을 위한 준비로 비칠 수 있습니다.
    • ‘힘의 사용’에 대한 관점 차이: 트럼프 행정부는 군사력 사용을 협상의 한 요소로 봅니다. “다시 한번 그런 공격을 원하느냐? 더 심한 것을 원하느냐?”라고 묻기 위해 협상 전에 힘을 사용하는 것을 주저하지 않습니다. 그러나 이란을 비롯한 상대방은 이러한 접근 방식을 동일하게 받아들이지 않습니다.
  2. 전쟁 장기화를 통한 전략적 이득:

    • 국민 단결 유도: 이란 정부는 극도로 인기가 없는 정권입니다. 경제적 mismanagement, 부패, 정부의 잔혹성에 대한 불만이 팽배합니다. 그러나 사회 핵심부를 공격하고 평범한 이란인들을 공포에 떨게 하는 외세의 침략은 오히려 국민들을 단결시켜 정권에 대한 지지를 끌어내는 역설적인 효과를 낳을 수 있습니다.
    • 협상력 강화: 이란은 시간이 자신들의 편이라고 믿습니다. 시장이 하락하고 유가가 오르며 국제적 불안감이 증가하고 서방 동맹이 분열되는 것은 도널드 트럼프에게 가장 치명적인 타격을 입힐 수 있습니다. 이란은 시장이 계속 붕괴하고 유가가 계속 오르면, 협상이 시작될 때 더 많은 협상력을 가질 수 있다고 판단합니다.
    • 핵무기 개발 동기 증대: 이 전쟁은 이란이 핵무기를 추구할 동기를 의심할 여지 없이 증가시켰습니다. 북한의 사례에서 보듯, 핵무기를 가진 나라는 미국과 이스라엘의 공격을 받을 가능성이 훨씬 적습니다. 생어 기자는 “우리는 이란에게 북한 모델을 따를 모든 동기를 부여했을지도 모릅니다”라고 경고합니다. 이란은 협상에서 핵연료 농축 권리를 영구히 포기하라는 미국의 요구를 철회시키고, 제한된 농축을 허용하는 방향으로 협상을 이끌어낼 수 있다고 보는 것입니다.

이스라엘의 복잡한 방정식

이스라엘의 전쟁 종식 열망은 트럼프 대통령과 다를 수 있습니다. 이스라엘은 레바논에 지상군을 배치하고 이란에 대한 공격을 계속하면서 이 갈등에서 더 큰 의제를 가지고 있습니다. 트럼프 대통령이 주식 시장, 유가, 동맹국의 반발 등에 관심을 두는 반면, 이스라엘은 이란이 국경 밖을 공격할 엄두도 내지 못할 정도로 무력화(defanged)되기를 원합니다.

미국은 이란 정부가 적어도 의사 결정을 하고 국제 협상에 참여할 만큼 응집력 있는 형태로 남기를 바라지만, 이스라엘은 이란 내부의 혼란(chaos)이 현재 정권을 전복시키고, 더 나아가 이란이 이스라엘에 대해 생각할 여유조차 없도록 만들기를 바랄 수도 있습니다. 흥미로운 점은 협상 여부와 시기에 대한 ‘춤’이 진행되는 동안에도 이스라엘은 이란 전역의 목표물을 계속 타격하고 있다는 것입니다. 최근에는 여러 핵 시설을 공격했으며, 이란 역시 이에 맞서 걸프만 내 미국 목표물을 공격, 미국 항공기 두 대를 격추하는 성공을 거두기도 했습니다.

2015년 합의와 현재: 비교할 수 없는 복잡성

생어 기자는 2015년 이란 핵 협상(JCPOA)을 직접 취재했던 경험을 바탕으로 당시와 현재 상황의 극명한 차이를 설명합니다. 2015년 합의는 길고 고통스러운 과정이었지만, 이란 핵 프로그램에만 초점을 맞춘 것이었습니다. 이란은 트럼프 대통령이 2018년 합의를 일방적으로 파기하기 전까지 약속을 지켰습니다. 생어 기자는 “트럼프의 그 결정이 현재의 공격으로 이어지는 직접적인 선을 그을 수 있는지 역사가들은 오랫동안 논쟁할 것”이라고 말합니다. 당시 외교는 이란 핵 프로그램이 통제 상태를 유지하고, 이란이 핵무기를 향해 달려갈 경우 최소 1년의 경고 기간을 확보하기 위함이었습니다.

그러나 현재의 상황은 2015년과는 비교할 수 없을 정도로 복잡합니다.

  1. 확대된 의제: 이제는 이란 핵 프로그램 외에 훨씬 더 많은 문제들이 야기되었습니다. 이란은 글로벌 경제를 교란하고 호르무즈 해협을 봉쇄하는 엄청난 능력을 보여주었으며, 이는 이제 협상의 주요 쟁점이 되었습니다.
  2. 더 깊어진 불신: 2015년에도 양측 간의 신뢰가 부족했지만, 지금은 그보다 훨씬 더 적습니다.
  3. 이란의 생존 문제: 2015년 합의에서 이란은 핵무기를 추구하지 않겠다고 약속했지만, 전쟁이 시작된 지금 새로운 이란 정부가 그러한 약속을 할 수 있을지 의문입니다. 이란에게는 이제 국가로서의 미래 생존 가능성(future survivability as a state)이 걸린 문제입니다.
  4. 새로운 정치적 힘: 전쟁이 시작되고 정부가 무너지고 국민들이 폭격을 당한 후에는 우리 누구도 진정으로 이해할 수 없는 일련의 새로운 정치적 힘이 촉발됩니다. 이란 정부에 대한 봉기(uprising)로 끝날 수도 있지만, 정반대의 효과를 가져올 수도 있습니다.

생어 기자는 “이전에는 그들의 핵 프로그램 상태에 대해서만 협상했습니다. 이제는 국가로서의 그들의 미래 생존 가능성에 대해 협상하는 것입니다. 그리고 이것은 이란인들에게 완전히 실존적인 문제이며, 이 전쟁을 끝내는 복잡성을 극적으로 증가시킬 것입니다”라고 강조합니다.

결론: 더욱 복잡해질 미래

2015년 협상이 아무리 복잡하고 섬세했더라도, 앞으로 이란과 미국 사이에 놓인 상황은 훨씬 더 복잡할 것입니다. 양측의 뿌리 깊은 불신, 전략적 이득에 대한 상반된 계산, 그리고 이스라엘의 독자적인 의제까지 얽히면서 외교적 해결의 길은 요원해 보입니다.

한편, 팟캐스트 말미에는 최근의 추가적인 소식이 전해졌습니다. 미 국방부는 수백 명의 특수작전 부대(Army Rangers, Navy Seals 포함)가 중동에 도착했다고 보도했으며, 이란 의회 의장은 미국이 이란에 지상군을 배치할 경우 “그들을 불태워버릴 것”이라고 경고했습니다. 또한 미국 내에서는 트럼프 대통령의 국내 정책과 군사력 사용에 반대하는 ‘노 킹스(No Kings)’ 집회가 전국적으로 열려 수백만 명이 참여했으며, 상원의원과 브루스 스프링스틴(Bruce Springsteen) 같은 유명인사들이 연사로 나섰습니다.

이 모든 상황은 중동 갈등이 단기간에 해결되기 어려울 뿐만 아니라, 외교의 과정 자체가 극심한 대결과 불확실성 속에서 진행될 것임을 시사하고 있습니다.