Welcome back!

This week we break down the data on why the AI founders that win stack organic content with strategic paid placements, walk through how Automateed went from a cloned prompt business to the top AI eBook creator, explore Perplexity’s partnership-powered growth to $200M ARR, and show how one solo founder turned a controversial tweet into six figures a month.

Have suggestions? Send them to us at [email protected].

Founder’s Intel

The best distribution stacks organic and paid together

Data Intel

• Content marketing costs 62% less than paid ads while generating 3× more leads per dollar spent.

• 62% of B2B buyers engage with 3-7 pieces of content before talking to a sales rep, and 11% consume more than seven.

• Companies with blogs generate 67% more leads per month and have 434% more indexed pages than those without.

Why It Matters

Content alone is slow. Paid ads alone are temporary. The founders growing fastest right now combine both.

Organic content builds trust.

Demo clips, build-in-public updates, product walkthroughs. Each piece becomes a permanent asset that compounds over time.

AI products are inherently visual, so a 30-second screen recording of your tool generating something impressive does more selling than any landing page copy.

But organic takes months to compound.

If nobody sees your content in month one, you’re publishing into a void. That’s where strategic paid placements change the math.

The difference is what you’re paying for.

Running ads on Meta or Google rents attention that disappears the moment your budget stops. But paying for a listing on a discovery platform gives you a permanent slot in front of people actively searching for tools like yours.

That traffic doesn’t stop when the campaign ends. It keeps compounding as long as the listing stays live.

This is why the smartest AI founders stack both channels.

They post content consistently to build trust and authority. Then they invest in placements where high-intent users are already looking, so the content has an audience from day one.

At TAAFT, the founders who grow fastest do exactly this.

Quick Tip

Don’t choose between organic and paid. Stack them.

Post content 3× per week to build trust. Then invest in a permanent listing where your target users are already searching.

The content converts the traffic. The listing supplies it. Together, they compound faster than either channel alone.

Behind the Tool

From Prompts to Product

The Spark

Stefan Mitrovic was one of the first people selling ChatGPT prompts.

His company had paying customers and real traction. But prompts are text.

Competitors copied his entire catalog and resold it as their own. There was no way to patent a prompt, no way to stop it.

So Stefan decided to build something harder to replicate: a software tool that did what the prompts did, but packaged inside an app nobody could clone overnight.

The Build

The first version of Automateed was a marketing task platform with around 100 AI-powered tasks (write a Facebook ad, generate a sub-niche, craft a summary).

But Stefan was also a book author and Upwork content creator who knew firsthand how frustrating the publishing process was.

He kept building toward that pain point until the tool evolved into what it is now an all-in-one AI eBook creator that takes users from idea to finished book (covers, chapters, images, 150+ pages) in minutes.

Stefan handles customer support himself through his personal email.

Every feature added in the past year came from direct user conversations. He ships updates almost daily, and his philosophy is straightforward: “the main purpose of the tool has to work perfectly; everything else is fixable.”

The Breakthrough

TAAFT’s newsletter gave Automateed its first traffic spike (600 users in the first two to three days).

Stefan wasn’t sure if the traffic was real at first. It was.

That initial feature generated $1,000 in sales against a $375 investment, a clear signal the product converted.

From there, Stefan focused on SEO and ranked Automateed near the top of Google for terms like “AI eBook Creator.”

He layered in Facebook and Instagram ads, with TAAFT providing steady passive traffic alongside the paid and organic channels.

Today, Automateed has generated over $200K in revenue, holds 300+ five-star reviews on Trustpilot, and supports output formats for Amazon KDP, EPUB, and PDF.

The Next Chapter

Stefan is expanding into adjacent tools like course creation (CourseAI), text simplification, and AI content detection removal.

He’s also building a publishing marketplace directly on the Automateed platform where users can publish and sell their books with 85% royalties.

His long-term bet is that as LLM context windows expand, standalone eBook tools will face pressure, so he’s diversifying now while the product leads its category.

Key Lessons

Build what can’t be copied in a day. Prompts, templates, and content get cloned overnight. Software with a real workflow behind it takes months to replicate, and that gap is your moat.

Discovery platforms beat broad ads for bootstrapped founders. TAAFT’s newsletter delivered 600 users and $1,000 in revenue for $375. Spending on ads before you know your product converts is burning runway.

Let users shape the roadmap. Every feature Stefan shipped in the past year came from direct user conversations. That feedback loop is faster and cheaper than any analytics dashboard, and it builds loyalty that compounds.

Diversify before the moat shrinks. LLM context windows are expanding, and standalone tools in every category will face pressure. The time to build adjacent products is when you’re ahead, not when you’re threatened.

Tool of the Week

Perplexity’s Distribution Partnership Playbook

What’s Perplexity?

An AI answer engine that searches the web and synthesizes cited responses. Reached $200M ARR by October 2025, 45M monthly active users, and 780M monthly queries at a $20B valuation.

What Worked

• Distribution partnerships over paid acquisition: Instead of spending on ads, Perplexity bundled free Pro subscriptions with device manufacturers and telcos. Samsung launched a Perplexity TV app across all 2025 models, giving every owner 12 months of Pro access free. Deutsche Telekom launched an AI Phone in 10 European markets with Perplexity Assistant built in and 18 months of Pro included. Airtel India bundled it with mobile plans for all 360M subscribers, driving 640% year-over-year user growth and 2.8M app downloads in Q2 2025 alone. Each partnership cost Perplexity nothing in ad spend while delivering millions of activated users.

• Samsung Galaxy S26 integration deepened the moat: Announced at Galaxy Unpacked on February 25, 2026, Perplexity is now embedded as a system-level AI agent across Samsung’s flagship phones. Users trigger it with “Hey Plex” or by long-pressing the side button. The agent connects to Samsung Notes, Calendar, Gallery, Reminders, and the Samsung Internet browser. Perplexity’s chief business officer called it the first time a third-party AI company achieved OS-level access on a Samsung phone.

• Free tier as flywheel: Unlimited free queries let users build a habit before paying. At $20/month for Pro and $200/month for Max, upsells happen naturally when usage deepens. Every free user improves the product through query data, creating a compounding data advantage.

• Publisher Program turned a liability into a moat: After copyright lawsuits from Dow Jones, New York Post, and the NYT, Perplexity launched a revenue-sharing program with publishers (Time, Fortune, Der Spiegel, LA Times). First AI company to pay creators at scale. What started as legal risk became a distribution and credibility asset.

• Computer launch expanded the product category: On February 25, Perplexity launched “Computer,” an autonomous agent that takes a project brief, breaks it into subtasks, and assigns each to the best-fit model (Nano Banana for images, Veo 3.1 for video, Gemini for deep research, Grok for lightweight speed tasks, ChatGPT 5.2 for long-context recall). The move positions Perplexity beyond search and into agentic work.

Founder Quote

“Perplexity is transitioning from just being an answer machine to an action machine. It’s going to start doing things for you, not just answering questions.” — Aravind Srinivas, Perplexity CEO

Key Lesson

You don’t need Samsung-level deals to apply this playbook. The principle scales down.

Find platforms, tools, or communities with large existing user bases and offer your product as a bundled value-add. A small AI tool for Shopify stores, a free integration for Notion, a plugin for a popular CRM.

The mechanism is the same: trade free access to your product for placement inside someone else’s distribution channel. Every user who tries you through a partner has zero acquisition cost and higher intent than a cold ad click.

Start with one integration partnership. Measure activation. Then stack more.

Fresh Out of the Lab

OpenSandbox

What Is It?

An open-source sandbox platform from Alibaba for running AI agent code safely. It provides multi-language SDKs (Python, Java/Kotlin, JavaScript/TypeScript, C#/.NET), unified APIs, and built-in Docker + Kubernetes runtimes.

What’s New

OpenSandbox solves a problem most AI agent teams hack around: secure execution of untrusted, dynamically generated code.

It ships with built-in support for coding agents (Claude Code, Gemini CLI, Codex), browser automation (Chrome, Playwright), VS Code desktop environments, and network isolation per sandbox instance.

The MCP (Model Context Protocol) server integration lets LLMs invoke sandbox capabilities as standardized tools. Fine-grained egress control via DNS proxy and nftables prevents data exfiltration.

Why It Matters

Founders building AI agents that write code, browse the web, or interact with file systems need isolation they don’t have to build from scratch.

Most teams either roll their own basic containers (fragile, insecure) or skip sandboxing entirely (risky).

OpenSandbox provides production-grade infrastructure out of the box: persistent state across multi-turn AI interactions, resource quotas via Kubernetes, Seccomp/AppArmor security modules, and elastic scaling.

Deploy locally with Docker for development, then move to Kubernetes clusters for production. No vendor lock-in. No per-execution fees.

Founder’s Edge

This Week’s Builder Toolkit

• Dev Tool: Mastra is an open-source TypeScript framework for building AI agents with built-in tool calling, RAG, and workflow orchestration. Think LangChain for TypeScript developers who want cleaner abstractions.

• Free Dataset: The CB Insights State of AI 2025 report breaks down funding trends, top investors, deal sizes, and emerging categories across 6,000+ AI deals. Filter by sector to find where capital is flowing and where gaps exist for new entrants.

• No-Code App: Cursor is an AI code editor where you describe what you want in plain English and the AI writes, edits, and debugs code. Built-in features like Command+K for instant edits and Tab for code prediction make it the fastest way for founders to ship MVPs without a full dev team.

• Productivity Hack: Tana combines notes, tasks, and databases into a single workspace with built-in AI. Tag anything, query across your entire knowledge base, and build custom views without switching between apps. Works like a personal operating system for research and project management.

• Learning Resource: Anthropic’s free prompt engineering course walks you through building effective prompts for production systems. Covers system prompts, tool use, and multi-step reasoning patterns.

Note: If you’ve found a tool that’s sped up your build process, hit reply and share it. We’ll feature the best submissions in a future issue.

AI Founder’s Journal

How Pieter Levels Turned Controversy Into Income

In July 2023, Pieter Levels posted a tweet that broke developer Twitter.

He shared that PhotoAI.com was “almost 14,000 lines of raw PHP mixed with inline HTML, CSS in style tags and raw JS in script tags” with no TypeScript, no flexbox, no frameworks except jQuery.

At the time, the product had 1,872 paying customers making $61,808 per month.

The tweet hit 4.9 million views. Devs lost their minds. Some called it brilliant. Others called it reckless.

The debate raged for weeks across Twitter, Reddit, Hacker News, and LinkedIn.

And signups spiked.

Here’s the backstory:

In late 2022, Levels started experimenting with Stable Diffusion and stumbled into AI-generated portraits. He launched AvatarAI first, a cheesy avatar generator that went viral and made $150K in a single week.

But it was a hype product.

Users came, generated their cartoon avatars, and left.

So he pivoted.

In February 2023, he launched PhotoAI.com with a sharper positioning of “Fire your photographer.”

Upload selfies, train a personal model, and generate photorealistic headshots for LinkedIn, dating profiles, and social media.

Week one: $5.4K in revenue.

By month six: $61.8K MRR.

By September 2025, a new record: $150K/month with 2,573 active subscribers and an 87% profit margin.

Solo founder. No full-time employees. No venture funding.

Tech stack: PHP + jQuery + SQLite on a Hetzner VPS with Nginx and Ubuntu.

Levels’ philosophy is simple: build fast, validate through payments (not signups), and kill projects that don’t make money within weeks.

He’s launched over 40 startups. Most failed. But by keeping each one lean and shipping rapidly, the winners pay for all the losers combined.

The controversial tech stack wasn’t an accident.

Levels uses PHP, jQuery, and raw HTML because that’s what he knows. No frameworks. No TypeScript. No build tools.

He ships features in hours instead of days, and the simplicity means fewer things break.

But here’s the part most people miss…

PhotoAI’s first-week revenue wasn’t random. Levels spent 10+ years building an audience of 600K+ followers on Twitter.

He posted build-in-public updates, revenue screenshots, and Stripe dashboards daily. When PhotoAI launched, he had customers waiting.

Compare that to launching without an audience. Most products make $500 to $2K in their first month.

Levels did $5.4K in his first week.

That’s a 3-10× advantage from day one.

Key Takeaways

• Building in public is a distribution strategy, not a vanity project.

• Share revenue numbers, ship publicly, and let controversy drive attention.

• Polarizing opinions about your tech stack, pricing, or approach generate 3-10× more engagement than polished marketing.

• The audience you build over years becomes the launch engine for every product that follows.

— Pieter Levels, solo founder (@levelsio)

Weekly Challenge

One Experiment. One Week. One Win.

The Goal

Create 3 pieces of founder-led content and measure which one drives the most signups in 7 days.

How It Works

1. Post ONE product demo video (screen recording, under 60 seconds).

2. Post ONE build-in-public update (what you shipped, what broke, what you learned).

3. Post ONE contrarian take on your industry.

4. Track clicks and signups from each post separately.

Why It Works

Most founders overthink content.

This challenge forces you to test three formats fast, measure what resonates, and build a repeatable content motion from real data instead of assumptions.

Spotlight

Share your user contribution system and early results in the TAAFT community by end of week. We'll showcase the highest-converting content strategy in an upcoming newsletter.

AI Market Watch

Deals, Discoveries, and Demand

Megadeals

• Wayve – $1.2B Series D (autonomous driving)

• OpenAI – $110B round ($730B pre-money valuation)

• Goodfire – $150M Series B (AI interpretability)

• Anthropic – $30B Series G (frontier AI research)

• World Labs – $1B round (spatial intelligence AI)

Top Research

• MiroFlow (open-source deep research agent framework)

• OmniGAIA (benchmark for omni-modal AI agents)

• PyVision-RL (reinforcement learning for multimodal tool use)

• SkyReels-V4 (unified video-audio generation model)

• SkillOrchestra (multi-agent orchestration with cost optimization)

• AgentDropoutV2 (pruning multi-agent systems for efficiency)

Search Trends

Top “best AI for…” searches:

1. Voice

2. Coding

3. Presentations

4. Shopping

5. Video editing

Thanks for reading!

Got a project in motion? Hit reply and share what you’re working on with us.

Keep an eye on your inbox. We’ll be back soon with more founder stories, fresh tools, and data worth your time.

Signing off,
— AI Empires