The AI Survival Story: What You Can Build With Zero Code Experience

Most people use AI for convenience. I used it for survival.

    This is not a Silicon Valley success story. There was no venture capital, no technical co-founder, no coding bootcamp, no safety net.

    Just a basement, a divorce, custody battles, and Claude running 14-16 hours a day.

    **This is what AI unlocks when you have no other option.**

What Got Built (By Someone Who Knows Nothing About Code)

      - Won two custody battles using AI to organize evidence and prepare legal arguments
      - Built a complete publishing platform with AI as teacher: 80+ Lambda functions, React app, Step Functions orchestration
      - Published 150+ books through the platform, helped network publish 100+ more
      - $0 in outside funding. Everything bootstrapped.
      - **Still don't know how to code.** AI walks me through everything.

2016: When Platforms Show You Who's Boss

    I built Ant Hive Media to $14,000/month in 7 months.

    **The model:**

    - 21 writers creating summary books of bestsellers
    - ~250 summary books across multiple platforms
    - Revenue: $6K from Amazon + $6K from Smashwords + $2K from Audible

    Then Amazon shut me down.

Three Copyright Strikes

    Summary books exist in a gray area. Some authors love them (free marketing). Some authors hate them (they want control).

    Tony Robbins has a team that hunts down anyone using his name. Some authors personally file takedowns. Amazon has a 3-strike policy: three copyright claims and you're out. No appeal. No negotiation.

    I got some strikes reversed by talking to authors directly. But eventually, three authors refused to withdraw their claims.

    **Amazon terminated my account.**

    $6,000/month gone instantly. Plus I lost the Audible revenue tied to that account (another $2K).

    **What I kept:** The Smashwords revenue ($6K/month) and Audible ($2K/month) = $8K total.

The Slow Death After the Shutdown

    For a couple years, I maintained ~$8K/month from Smashwords (now Draft2Digital) and Audible.

    Then the platforms kept chipping away:

    - **Year 3:** Audible changed their algorithm → lost $1K/month
    - **Year 4:** Audible banned summary books entirely → lost another $1K/month
    - **Years 5-8:** Draft2Digital revenue slowly declined to ~$4K/month as summary books fell out of favor
    - **Year 9:** Down to $1K/month
    - **Today (10 years later):** ~$500/month from Draft2Digital

    **From $14K/month to $500/month over a decade.**

    Not because the content got worse. Not because readers stopped wanting it. Because platforms changed policies, algorithms, and rules.

The Lesson I Should Have Learned

    Platforms don't just flip the switch once. They keep flipping it.

    Every policy change. Every algorithm update. Every new restriction. Each one takes a piece.

    You can follow the rules, stay in the gray areas they allow, and still watch your income die slowly over years.

    **Platform risk isn't just existential. It's erosive.**

September 2023: When AI Became Survival

    My wife asked for a divorce.

    Within 48 hours, I learned that custody battles run on documentation. Every conversation, every interaction, every piece of evidence matters. But I was drowning in thousands of text messages, court filings, and contradictory claims with no way to make sense of it all.

    **I needed a force multiplier.**

    I spent 30 days teaching myself how to use Claude effectively. Not for writing books. Not for content creation. For organizing evidence, finding contradictions, drafting legal responses, and preparing for depositions.

What AI Did in the First Custody Battle (October 2023)

      - Analyzed 2,000+ text messages to find patterns and contradictions
      - Generated timeline visualizations from scattered evidence
      - Drafted counter-motions in hours instead of days
      - Prepared me for depositions by running mock question-answer scenarios
      - Spotted inconsistencies in opposing counsel's arguments that I was too stressed to catch

Result: Won primary custody in October 2023.

The Realization That Changed Everything

    If AI could help me win a custody battle—the highest-stakes situation of my life—it could definitely help me build software.

    Even though I had no idea how to code.

    Even though I'd never built anything technical beyond some bad WordPress sites.

    **If AI could teach me legal strategy under pressure, it could teach me to build.**

March 2024: Starting Teneo

    I had a vision: a platform that could generate complete books using AI, but with quality controls and workflow automation.

    **The problem:** I didn't know how to build it.

    **What I did know:**

    - Publishing workflows (from 10 years in the industry)
    - What made books good vs bad (from editing hundreds)
    - How to use Claude effectively (from the custody battles)

Month 1: Figuring Out the Right Path

    My first version generated a book in **1 hour**.

    One chapter at a time, sequentially. Waiting for each to finish before starting the next.

    **Unusable.**

    I asked Claude: "How do I make this faster?"

    *Claude:* "You need parallel processing. Generate all chapters simultaneously using AWS Lambda functions."

    *Me:* "What's AWS? What's a Lambda function?"

    **I restarted 3-4 times** before figuring out the right architecture:

    - AWS Lambda for serverless compute (no servers to manage)
    - Step Functions for workflow orchestration (managing the sequence)
    - Parallel execution for chapters (all at once instead of sequential)

    **Result:** Book generation time dropped to 3-5 minutes.

Months 2-6: Building the Generator

    Every single day, I'd sit in my basement and collaborate with Claude:

    - **Me:** "I need to save user data. How?"
    - **Claude:** "Use DynamoDB. Here's how to set it up..."
    - **Me:** Follows instructions, breaks it, pastes error
    - **Claude:** "That error means... here's the fix..."
    - **Me:** Fixes it, documents what I learned, moves to next feature

    **I still don't know how to code.**

    But I know how to:

    - Describe what I want in plain language
    - Let Claude translate it to code
    - Run it and capture errors
    - Show Claude the errors
    - Understand the explanations
    - Build knowledge by doing

The Outline Generator Moment

    After weeks of work, I got the outline generator working.

    My girlfriend watched me type a book idea and get back a complete 15-chapter outline with descriptions, frameworks, and structure.

    **She was blown away.**

    She got so excited—like she suddenly realized my wild idea would actually work.

    To me, it was just an outline. But she was right.

    **Eventually, it became a whole book.**

September 2024: V1 Works

    After 6 months of 14-16 hour days, the generator worked:

    - Generate complete book in 3-5 minutes
    - Outline → chapters → formatting
    - Quality evaluation system
    - Export to multiple formats (Word, Kindle, IngramSpark)

    **Technical stack built with zero code experience:**

    - React frontend
    - 80+ Lambda functions (Python and Node.js)
    - Step Function workflows for book generation
    - DynamoDB for data storage
    - S3 for file storage
    - API Gateway for routing
    - Cognito for authentication

    **AWS costs during development:** Free tier for first few months, then ~$20/month when I exceeded limits.

Publishing Network: 4 Months of Testing

    I spent the next 4 months publishing books with a network of about 8 people:

    - **I published ~150 books** across multiple niches
    - **Network published ~100 books** total
    - **Revenue from my books:** Building to ~$800/month

    **What I learned:** The generator worked, but quality was inconsistent. Positioning was off. Brands weren't dialed in.

    I needed to improve the quality of the output and build better brand systems.

Next 10 Months: Building V2

    I spent 10 months rebuilding from scratch:

    - Better quality evaluation (3-layer checking system)
    - Style training (learns your voice from samples)
    - Coherence checking (maintains consistency across chapters)
    - More control over output (50+ customization options)

    **Tradeoff:** V2 generates books in 8-12 minutes (slower than V1's 3-5 minutes) because it does more quality checks. I can optimize it back to ~5 minutes if I have time.

September 2025: Public Release

    Released V2 to users. Started building a Skool community. Shifted focus from pure generator work to UI improvements and marketing.

    **Marketing experiment:** Ran $1,000 in Facebook ads → 2 sales. 😬

    **Lesson learned:** My positioning was wrong. The marketing messaging wasn't dialed in.

End of October 2025: Back to Fundamentals

    Went back to work on marketing and positioning:

    - Started publishing books again (testing positioning)
    - Started posting on YouTube
    - Focused on clarity: what problem does this actually solve?

November 5, 2025: Building the Brand Generator

    **Current focus:** Almost done with Brand Generator v1.

    **The insight:** Books alone aren't enough. You need a complete brand—positioning, voice, audience, messaging. The Brand Generator will help users build that foundation before they write a single word.

    **Goal:** Fix the positioning problem that's keeping book revenue at ~$800/month instead of $3K-5K where it should be with 150 books.

Current Status

    - **Paying users:** 3 (~$100 MRR)
    - **Book revenue:** ~$800/month (positioning off, brands not dialed)
    - **Still working:** 14-16 hour days when kids are at their mom's or at school
    - **Still learning:** AI teaches me every single day

March 2025: Custody Battle #2

    Just as the business was stabilizing, I faced a second custody battle. This time, the allegation was that I was "working too much" (those 14-16 hour days building Teneo) and neglecting the children.

    **The pattern repeated:**

    Use Claude to organize evidence → analyze text messages → draft counter-motions → prepare defense.

    **Result:** Won again in March 2025.

What AI Actually Unlocked

    **I didn't build Teneo. Claude and I built it together.**

    Here's what collaboration with AI made possible:

1. Building Without Knowing How to Code

    **Traditional path:** Learn to code → build something simple → learn more → build something complex → 2-4 years minimum

    **With AI:** Describe what you want → AI generates code → Run it → Fix errors with AI → Learn by building → 6 months to working product

    **The difference:** I still don't know how to code. But I can build production software by collaborating with AI.

2. Learning by Doing, Not Studying

    I never read a programming book. Never watched a tutorial series. Never took a course.

    **Every error message became a teaching moment:**

    - "Undefined is not a function" → Claude explains JavaScript scope
    - "Access Denied" → Claude explains AWS IAM permissions
    - "Cannot read property of null" → Claude explains error handling

    I learned concepts in context, when I actually needed them, solving real problems.

3. Restarting Is Part of the Process

    **I restarted Teneo 3-4 times** before finding the right architecture.

    First version: 1 hour per book. Terrible.

    Each restart taught me something:

    - Restart 1: Learned sequential processing doesn't scale
    - Restart 2: Learned I needed serverless architecture
    - Restart 3: Figured out parallel execution
    - Restart 4: Got it working in 3-5 minutes

    **Without AI, each restart would have taken months. With AI, each restart took days.**

4. Surviving While Building

    **The reality:** I was building Teneo while fighting two custody battles, managing publishing revenue, and raising my children.

    **What AI did:**

    - Helped me win custody battles (freeing mental bandwidth)
    - Taught me to build software (during nights and weekends)
    - Never got frustrated when I asked the same question three times
    - Was available 24/7 when I had time to work (not 9-5)

What This Means for You

    **You don't need to know how to code either.**

    If you have:

    - A problem you understand deeply (from living it)
    - Access to AI collaboration tools (Claude, GPT-4, etc.)
    - Willingness to work when others stop
    - Tolerance for restarting until you get it right

    **You can build things that used to require teams, funding, and years of coding experience.**

The Pattern That Works

    - **Describe what you want** in plain language (be specific)
    - **AI generates code** (you don't need to understand every line)
    - **Run it and capture errors** (they're learning opportunities)
    - **Show AI the errors** (paste them exactly)
    - **AI explains and fixes** (you build understanding over time)
    - **Document what you learned** (helps AI in future sessions)
    - **Repeat daily** (compound learning beats intensity)

Start Before You're Ready

    I didn't learn to code, then build Teneo.

    I started building Teneo, then learned as I went.

    **Every error taught me something:**

    - 500 Internal Server Error → learned API error handling
    - Database corruption → learned about data validation
    - 1-hour book generation → learned about parallel processing
    - Inconsistent quality → learned about evaluation systems

    Waiting to be "ready" is procrastination with better branding.

The Real Lesson

    **AI is a force multiplier, not a replacement.**

    Claude didn't win my custody battles. I did.

    Claude didn't build Teneo. We did.

    Claude didn't publish 150 books. I did.

    But I couldn't have done any of it **without** Claude.

    **The pattern:**

    - I provide judgment, context, and domain expertise
    - AI provides speed, technical knowledge, and tireless execution
    - Together, we achieve things neither could do alone

What Comes Next

    I'm still in the basement. Still working 14-16 hour days when my kids aren't here. Still learning every single day with AI as my teacher.

    **Current focus:**

    - Finishing Brand Generator v1 (almost done)
    - Fixing positioning so book revenue scales from $800 → $3K-5K/month
    - Improving marketing messaging (so ads actually convert)
    - Building in public on YouTube

If You're Building Something

    If you're trying to build something with no technical background, using AI to learn skills you "shouldn't" be able to learn, or working on something while fighting battles in other parts of your life:

    **This is proof that it's possible.**

    Not easy. Not guaranteed. But **possible**.

    You can build it.

    You can learn it.

    You can survive it.

    And if AI can help win custody battles and build publishing platforms, it can definitely help you build whatever you're trying to build.

See What AI and Determination Can Build

      Try Teneo—the platform I built with AI from a basement. No credit card required.

      [Start building →](/signup)

    **Travis Eric**

    Founder, Teneo

    *Still learning, still building, still in the basement*

Related Reading

    - [The Basement Year](/learn/the-basement-year) — Month-by-month breakdown of building with AI
    - [The Algorithmic Publishing Stack](/learn/the-algorithmic-publishing-stack) — Technical architecture details
    - [AI Book Quality](/learn/ai-book-quality) — How the quality system works