Solo Founder. $80M. Six Months.

He built it alone, never raised a dollar, and did the impossible.

March 23, 2026

Welcome back!

This week we break down the data behind building in public, walk through how one solo founder went from zero to an $80M exit in six months without spending on marketing, explore Mirage’s weekly shipping playbook, and show how a non-tech data analyst quietly built a $2.6M/year AI tool from scratch.

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

Founder’s Intel

Is building in public part of your growth strategy yet?

Data Intel

  • Founders who share consistent build-in-public content report 3–5× more organic signups in their first 90 days compared to those who launch cold.

  • Products launched to an existing audience generate 3–10× more revenue in week one than cold launches with no prior presence.

  • LinkedIn and X posts about product milestones, revenue numbers, and shipping updates average 40–60% higher engagement than promo content.

Why It Matters

Build-in-public isn’t just content. It’s a distribution engine that compounds over time. Every progress update, revenue screenshot, or shipping announcement builds an audience that already trusts you before they ever pay you.

The compounding effect is the part most people underestimate. A founder who posts consistently for six months doesn’t just have six months of content. They have six months of trust, six months of search presence, and an audience that’s been watching the product get built from the start.

When those people finally sign up, they convert at much higher rates than cold traffic because they already know the story. That’s why build-in-public isn’t a marketing tactic, it’s a pre-loaded distribution channel you build while you’re still building the product.

Quick Tip

Post one build update per week (revenue milestones, features shipped, lessons from user feedback).

Don’t wait until things are impressive. Start when they’re messy, and the audience that finds you then will be your most loyal customers.

Behind the Tool

Solo Founder Does The Impossible

The Spark

In late 2023, Maor Shlomo finished a year of reserve military service. He’d spent a year working alongside the country’s best engineers on high-stakes problems. When he came out, he wanted to get back to building. Something small. Something he could move fast on alone.

He noticed a pattern during his service: every organization he helped needed simple internal tools, and every time the answer was an expensive agency quote and a six-month timeline. The idea clicked. What if anyone could describe what they wanted and an AI just built it?

Base44 wasn’t meant to be a big bet. He told his wife if it hit $1.5M ARR by end of 2025, they’d buy a nice car. It hit that in four weeks.

The Build

Shlomo built Base44 almost entirely alone. For roughly three months, he didn’t write a single line of HTML or JavaScript. AI wrote the front end. He focused on infrastructure, constraints, and the core product loop.

The key insight was that most no-code platforms were still too difficult for truly non-technical users. LLMs were already capable of writing the code.

The missing piece was the right infrastructure around them: built-in databases, user authentication, and storage so that the AI had everything it needed to produce something real.

He started with three close friends. Watched them use it. Fixed bugs in real time. The feedback loop was brutally tight. When something broke, he knew within minutes.

The Breakthrough

Growth compounded through two loops running at the same time.

  1. Users received bonus credits for sharing what they built on social media, which turned every successful app into a distribution object.

  2. Maor posted revenue milestones, build updates, and product decisions publicly on LinkedIn throughout the entire journey.

By the time most people heard about Base44, they’d already been watching it for months. That transparency became the primary growth channel, beating every paid alternative he tested.

By the time of the acquisition, Base44 had crossed 400,000 users and $3.5M ARR with zero outside funding. Wix acquired the company for over $80M in June 2025.

The Next Chapter

Base44 now runs as a distinct unit inside Wix, with Shlomo leading it.

The mission hasn’t changed: let anyone build software by describing what they want. The difference now is Wix’s distribution and enterprise infrastructure behind it.

Key Lesson

Solo doesn’t mean slow anymore.

AI-assisted development, tight user feedback loops, and build-in-public distribution let a single founder move faster than a funded team.

The advantage isn’t funding. It’s velocity.

Tool of the Week

Mirage’s Weekly Shipping Playbook

What’s Mirage?

An AI video creation and editing platform (formerly Captions) that lets anyone create, edit, and dub short-form videos using AI. $100M raised. Over 100,000 daily active users.

What Worked

  • Built a two-roadmap system. Mirage runs a public roadmap of features users explicitly request, and a secret roadmap of ideas the team believes could fundamentally change how users behave. The public roadmap keeps existing users happy and generates word-of-mouth. The secret roadmap is where the product actually leaps forward. Most teams only have the first one.

  • Every engineer ships something marketable every week. Misra’s core rule is that a marketable feature is something users would come to the app specifically for, or even pay for on its own. Not a polish update. Not a bug fix. An actual feature. Every week. Without exception.

  • Cut scope, not quality. When deadlines tighten, they ask what can be removed until removing any more would make the product useless. They ship the smallest valuable version.

  • Killed the CapCut vacuum. When TikTok’s ban looked likely in early 2025, Mirage went freemium to capture the audience that would need a replacement. Traditional video editing features, previously paywalled, opened up for free. AI features stayed behind the upgrade tier.

Founder Quote

"If nobody complains, it's almost a red flag. Complaints tell you exactly what matters most to your users." — Gaurav Misra, Mirage co-founder and CEO

Key Lesson

Shipping every week forces discipline.

It keeps the product feeling alive, generates new content from users, and gives you constant signal on what’s working. Most products update quarterly and die quietly between releases.

Fresh Out of the Lab

DeerFlow

What Is It?

An open-source SuperAgent harness from ByteDance that researches, codes, and builds. Built on LangGraph and LangChain, it runs inside an isolated Docker container with a persistent filesystem, long-term memory across sessions, and a lead agent that decomposes complex tasks and spawns parallel sub-agents to execute them.

What’s New

DeerFlow 2.0 is a complete ground-up rewrite.

The original version was a deep research tool. Devs used it to build data pipelines, spin up dashboards, and automate content workflows instead. ByteDance took that signal and rebuilt it as a task-agnostic runtime.

Give it one high-level prompt, “research the top AI startups this week and build me a slide deck”, and it plans the workflow, delegates research, coding, and image generation to parallel sub-agents, then assembles everything into a finished deliverable.

The agent actually executes code in a real bash terminal rather than suggesting it. It works with any OpenAI-compatible model: GPT, Claude, Gemini, DeepSeek, or local models via Ollama.

Why It Matters

Most agent frameworks hand you back a string of text. DeerFlow hands back a chart, a slide deck, or a deployed web page.

For founders who need multi-step research, competitive analysis, or automated content pipelines without paying per-call API costs, this is the closest open-source equivalent to having an AI employee with its own computer.

Self-hostable, fully extensible, no vendor lock-in.

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 pipelines, and workflow orchestration. The cleanest path from idea to production agent for TypeScript devs.

  • Free Dataset: Stripe's Atlas Year in Review covers 23,000 companies that incorporated in 2025 (time-to-revenue trends, AI startup rates, and global founding patterns).

  • No-Code App: Tally is the simplest way to build free, beautiful forms. No paywalls for logic, file uploads, or payment collection. Ideal for user research and waitlists without the Typeform tax.

  • Productivity Hack: Use Granola to capture meeting notes locally without any bot joining your call. Notes enhance automatically after the meeting. Nobody knows it’s running.

  • Learning Resource: Paul Graham's Essays are the most referenced free writing on startup thinking ever published. If you haven't read "Do Things That Don't Scale" or "Founder Mode," start there.

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

$2.6M/Year Without Writing Code

David Bressler was a data analyst. Not a dev. He knew Excel and Google Sheets the way most people know their own names, but the idea of building software felt far away.

In 2022, David went down a rabbit hole on GPT-3 during paternity leave and had one question: could he use it to automatically generate spreadsheet formulas from descriptions?

He tested the concept in a weekend. It worked well enough to ship.

FormulaBot launched shortly after with one feature (describe what you want your formula to do, get the formula back).

The product spread through data teams and finance departments without any paid spend. People who used spreadsheets all day discovered they could do in seconds what previously took twenty minutes of searching Stack Overflow.

They shared it with colleagues. Those colleagues shared it with others.

But the real breakthrough came from David watching where his users struggled. They weren’t just generating formulas. They were trying to analyze data, explain errors, and automate workflows they couldn’t code themselves.

So he expanded FormulaBot into a full data analysis suite: formula generation, error explanation, a SQL generator, and an AI layer that worked across both Excel and Google Sheets.

The Result:

FormulaBot generates over $220,000 per month in revenue, all bootstrapped, with a non-technical founder who used no-code tools to build the initial version and the OpenAI API to power the core feature.

The tactic that made it work isn’t complicated.

David built for one specific type of user, the person who lives in spreadsheets but can’t code. He didn’t try to serve everyone.

Every feature he added solved a problem that specific person had. That focus made word-of-mouth almost automatic, because when you fix something someone hates doing every day, they tell everyone.

Key Takeaways

You don’t need to be a dev to build an AI tool.

Pick one painful, specific task for one specific type of user. Use available APIs to automate it. Then expand into the adjacent problems that same user has. The niche compounds.

— David Bressler, FormulaBot founder

Weekly Challenge

One Experiment. One Week. One Win.

The Goal

Post one build-in-public update about your product this week and track how many profile visits, signups, or replies it generates.

How It Works

  1. Share ONE milestone, lesson, or feature you shipped recently. Include a number if you have one (users, revenue, requests handled, time saved).

  2. Post it on X, LinkedIn, or both.

  3. Track what happens (profile visits, signups, DMs, and replies over 7 days).

Why It Works

Most founders skip this because it feels uncomfortable. That discomfort is the point.

The founders who grow fastest remove the gap between building and telling people they’re building. One post takes 15 minutes.

The compounding effect takes months, but it starts today.

Spotlight

Share your results in the TAAFT community by end of week. We'll showcase the highest-performing build-in-public post in an upcoming newsletter.

AI Market Watch

Deals, Discoveries, and Demand

Megadeals
Top Research
Search Trends

Top “best AI for…” searches:

  1. Video Generation

  2. All-in-one Editor

  3. AI Content Detection

  4. Websites

  5. Agents

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