I teach my friend how to make his first internet $ with AI.

Business / Strategy Summary

• Core message: AI “coding agents” can reduce software/product creation from months to minutes, enabling first-time builders to ship working products quickly—then monetize with simple subscription pricing.
• The event is framed as a 3-level execution funnel:
  • Level 1: Build a personal landing page (Lovable) → publish → validate with a “first visitor”.
  • Level 2: Build a functional, logged-in web app with user data (Replit) → add account system, uploads, and a trial/monetization concept to target first revenue.
  • Level 3: Build a local iOS app (Xcode + Claude Code) → demonstrate end-to-end mobile functionality (with payments noted as the main remaining issue).

---

Frameworks / Playbooks Mentioned

Agent-based “prompt-to-build” workflow

• Describe the desired outcome in plain language.
• The agent generates the code/files and deploys the result.
• A human then reviews, requests changes, and asks the agent to fix issues—since agents are early/buggy.

Testing & iteration loop

• Replit runs automated app testing and supports agent-driven fixes using pass/fail signals.
• When something breaks (e.g., hallucinated links, Stripe checkout issues), you instruct the agent to debug and fix (e.g., “debug with agent”).

Simple monetization ladder

• Start with a basic paid plan (e.g., $5/month).
• Run pricing experiments later, such as:
  • Free for first 10 entries, then charge
  • Paywall for viewing history

Product validation through micro-metrics

• First confirm reach (“first visitor”).
• Then move to monetization attempts (e.g., Stripe sandbox → publish).

---

Concrete Examples / Mini Case Studies

Level 1: Lovable personal website (Landing page)

Inputs

• Copied LinkedIn and Twitter/X content.

Output

• A live personal landing page highlighting relevant projects.

Enhancements

• “Night mode/light mode” added via natural language edits.

Failure handling

• Acknowledged risk of hallucinated/made-up links.
• The agent can be instructed to fix broken sections.

Result

• The site was published live, and “visitors” were counted via a visitors widget.

---

Level 2: Replit web app — “Moments” (private visual journal)

Inputs

• An audience sketch and feature requirements.

Features built

• Private user accounts (login required)
• “New moment” creation: upload image + add text
• A Moments feed UI (calendar-like UI mentioned; later moved toward a scrollable feed/design)
• Persistence via an automatically created database
• Monetization attempt:
  • $5/month subscription using Stripe
  • Stripe setup + checkout page generation

Verification approach

• Replit performs AI-driven end-to-end UI testing (agent acts like a user; includes flow video analysis).
• Stripe verified using sandbox payments with a test card number: 4242.

Failure case

• Subscription/banner/checkout worked at one point, but later Stripe appeared broken/blank.
• A refresh/publish attempt failed with runtime errors.
• Conceptual resolution: agent can debug and republish, but time ran out.
  • Fixing required additional iteration (described as taking tens of minutes).

---

Level 3: Claude Code + Xcode iOS app — “Workout Tracker”

Build

• Implemented locally as a functioning iOS app (simulated on the computer).

Features

• Welcome/onboarding:
  • name, age, units
• Workout logging:
  • add workout → add exercise → reps/weight → save
• The app remembers workout history (database persistence implied).

Monetization

• Payments were not fully validated in the provided recording.
• Stripe key/config issues were surfaced (e.g., “I gave the wrong password / wrong keys”).

---

Game generation with agents

• Rapid creation of a “Flappy Bird” variant (“Gotham Flap”).
• Difficulty tuning via instructions (gravity, speed, gap size, spawn rate).
• Code generation cost cited:
  • $0.55–$0.58 range.

---

Key Metrics / KPIs and Targets

Personal app / website growth story (Farza’s earlier work)

• Claimed: $100,000/year by age 15
• Claimed traffic: first app reached ~1.5–2 million users in 2 weeks
• Modern comparison:
  • AI can generate a “same website” in ~6 minutes vs 6 months

Event-level “early validation” KPIs

• Audience site visitor count:
  • personal page: “three visitors” (including the event/test audience)

App monetization target

• Target: $5/month subscription
• “First dollar” motivation:
  • Unit economics example: if cost ~$3 to build payments and you get at least one sale at $5, then ~$2 profit.

Stripe testing

• Stripe sandbox verification using test card number 4242.

Code generation cost examples

• Game: $0.58, later $0.55
• “Whole website so far cost us $3” (Moments app, at that stage)

Scale logic (conversion math)

• Example revenue benchmark: $250/month (“happy” benchmark)
• Example split: 25 people paying $10/month (or similar)
• Encourages stepping stones:
  • “First one → then five → then it grows”

---

Actionable Recommendations (Implicit Lessons)

• Choose a narrow, personal utility product (avoid building a “huge idea” first).
  • Example: a journal app “made for me”.
• Start with a landing page to validate demand quickly.
  • Use LinkedIn/Twitter content to generate a polished pitch fast.
• Ship a working version fast, then iterate.
  • Expect agent failures (hallucinations, broken links, UI mismatches, mobile responsiveness gaps).
  • Treat debugging as normal workflow: instruct the agent to fix and republish.
• Use subscriptions for first monetization.
  • Start simple (e.g., $5/month).
  • Later variants:
    • free tier by number of entries
    • paywall for full history
    • recent-only free
• Test payments safely.
  • Use Stripe sandbox + test cards (4242).
• Leverage agent-driven QA.
  • Use platform automated testing and AI simulation of user flows.

---

Presenters / Sources Mentioned

People

• Farza (main speaker/teacher)
• Sid / Sidharth (student; co-builder during the live session)

Tools / Platforms

• Lovable (landing page agent)
• Replit (web app agent + database/auth + AI testing)
• Claude Code (advanced agent for more complex builds)
• Xcode (iOS app building environment)
• Stripe (subscriptions and checkout)
• AI/ChatGPT (referenced as baseline code-writing capability)

Additional named references (in biography/background)

• Paul Graham, Mark Cuban
• Sean Puri (example of a personal landing page)

Event / partners / credits

• Replit partnership (event sponsor)
• Nevermind (design/video agency mentioned)
• High school/Principal and tech support team (acknowledged)