There are 200,000 new vibe coded projects going live on Lovable every single day.
How many of them get customers? Almost none.
Greg Isenberg recorded an episode about this at night because it was bugging him that much. Every YouTube tutorial teaches you how to vibe code the app. Almost nobody teaches you how to get customers for your app once it exists.
His argument: the wealthiest people over the next 10 years will be marketers, not engineers. Code is commoditized. Distribution is the moat.
He laid out 7 plays you can start this week. I pulled them apart below, with the numbers.
Watch the original episode first
Everything below comes from this video on Greg's channel. Worth the full watch:
The great flip: marketers over engineers
When Greg moved to Silicon Valley in 2014, engineers sat at the top of the hierarchy. He studied computer science because of it. Marketing people were, in his words, the laughing stock of the Valley.
That hierarchy just flipped.
AI made building cheap. So the scarce skill is no longer "can you ship it", it's "can anyone find it". Distribution people first, product people second, developers third.
His proof is Pieter Levels: $3M+ in revenue, zero employees. Most of his tools could be cloned in a weekend. Nobody clones his 750,000 followers, or the SEO authority he built by tweeting 125,000 times over 8 years.
The trap Greg sees everywhere: you vibe code something, launch it, silence. So you build more features, launch again, more silence. Product first, distribution never.
Smart builders flip it. Grow an audience to even 1,000 people, ask them what they need, build it in 72 hours, launch warm. That's the whole playbook behind building in public too (I wrote a full guide on how to build in public if you want the tactical version).
Now the 7 plays.
1. Turn an MCP server into your sales team
This is the one most founders haven't touched yet.
An MCP server is basically an app that AI assistants can call. A user asks Claude or ChatGPT a question, the AI discovers your MCP server, returns your product, and the user gets value. Zero CAC. The AI assistant is your sales team, working 24/7.
Greg has a friend in fintech who vibe coded an MCP server and got 150+ installations in 30 days with $0 ad spend.
His comparison: building an MCP server in 2026 is like building for mobile in 2010. You're early, and the platforms want you there.
Start this week:
→ Identify the one question your product answers
→ Build an MCP server that returns that data (realistically 24 hours of work)
→ Publish it to registries: Smithery, MCPT, Open Tools
Not for everyone. If you run a boring local business, skip it. If you run a SaaS, this is free pipeline.
2. Programmatic SEO: 10,000 pages in 48 hours
The math on this one is worth staring at.
10,000 pages × 30 visits a month each = 300,000 monthly visitors. Convert 2% at $10 = $60,000 a month. From pages you built once.
Those 30 visits per page won't show up overnight, SEO compounds slowly. But directionally, that's the machine you're building.
The pattern: "best X for Y". Best CRM for dentists. Roofing in Miami. Pick a keyword pattern, scrape a clean dataset with Firecrawl, build a page template in Next.js with Claude Code, and generate unique content per page.
Greg is blunt about the catch: you can't press one button and ship 10,000 pages of slop. The content has to not feel like AI, which means real editing passes and a human in the loop.
So start with 100 pages as an MVP. Monitor indexation. Scale only when Google is eating them.
This works whether you're a SaaS, an agency, or a boring business. Still undertapped.
3. The free tool IS the marketing
Ahrefs has run a free backlink checker for years. You paste your domain, get instant value, and the moment you want more, you're staring at a paid plan. The tool is the funnel.
What changed: Claude Code collapsed the build cost. A free tool used to take months. Now you can vibe code a grader, analyzer, or calculator in a day and ship it by lunch.
The loop:
→ User gets instant value ("your site scores 43/100")
→ You capture their email to unlock the result
→ They share the score (social proof + backlinks)
→ More users discover it
→ You upsell the fix
Greg's move, and I like this one a lot: everyone has a content calendar, nobody has a free tool calendar. He's shipping free tools across all his businesses as fast as he can.
Ask your LLM: "Here's what I'm building. Give me 10 free tool ideas for my top of funnel." Pick the best one. Build it this week.
One thing Greg skips past: capturing the email is half the job. If you don't have a follow-up sequence behind the tool, those leads rot. TwiLead has a solid breakdown of lead nurturing strategies if you want the second half of that funnel.
4. Answer engine optimization: be the source AI cites
Old way: rank on Google page one with 3,000-word posts and backlink grinding.
New way: get cited by ChatGPT and Perplexity when your customer asks a question. Zero-click search is growing, and Greg's take is that AEO in 2026 is where SEO was in 2010. First movers own their niche for years.
Pieter Levels again: his AI referrals jumped from 4% to 20% of traffic in one month.
The playbook:
→ Google the top 20 questions your customers ask
→ Write the definitive, structured answer for each (direct and citation-worthy, not fluff)
→ Add schema markup and FAQ blocks
→ Publish on a domain with authority, or build authority with the other 6 plays
→ Monitor your citations with Otterly, Profound, or manual testing
I went deep on this exact play for SaaS specifically: how to get your SaaS recommended by ChatGPT.
5. Make your product's output shareable
Spotify Wrapped gets 100 million shares every December. Your mother shares it. It says something about who you are.
The question for your product: what does your user want to brag about?
→ GitHub: the contribution graph. Devs post their green squares unprompted.
→ Stripe Atlas: the incorporation milestone. "5 years since I incorporated, here's what I learned."
→ Duolingo: the streak. People defend a 365-day streak like it's their child.
Find your artifact. Make it beautiful, make it branded (logo subtle, not screaming, nobody shares someone else's logo), and add a share button that pre-fills the post.
"But I'm B2B." B2B is people too. Maybe the share happens in Slack instead of on X. Same psychology.
Every share is free impressions to your exact target audience, and the user does the marketing because they're bragging about themselves.
6. Buy a newsletter instead of building an audience
Building an audience takes years of daily content with no guarantee it works. Greg's shortcut: acquire one.
You can buy a 10,000-subscriber niche newsletter for $5,000 to $20,000. You inherit the trust on day one, plug your product in immediately, and repeat across niches.
Most small newsletter operators make $0 to $500 a month. A $10K offer is life-changing money for an asset they were about to abandon.
Where to look:
→ Duuce.com (newsletter marketplace)
→ Newsletter investor communities
→ Just search "[your niche] newsletter" on X or Substack and DM: "Ever thought about selling?"
Target 5,000 to 50,000 subscribers in your exact niche. And unlike a social account, nobody can suppress your reach or take it away. It's a direct line you own.
If you know your LTV, the math is not hard.
7. Build an AI content repurposing engine
One pillar piece of content should become 20 to 75 assets.
Record one 30-minute thing per week: a podcast episode, a video, even a voice memo you never publish. Talking for 30 minutes is 10x easier than writing a 3,000-word essay.
Then AI turns it into:
→ 5-10 tweets
→ 3-5 LinkedIn posts
→ 2-3 short-form videos (Remotion skill + Claude Code)
→ 1 newsletter edition
→ 1 blog post
→ 5-10 quote graphics
Greg's warning applies here too: you get slop by default. You have to feed it your brand references and set up real routines so the output doesn't scream AI.
But do this weekly and in 3 months you have more content than your competitors, because they are simply not producing at this volume. With for-you pages, you don't need followers to go viral in your niche. You need shots on net.
Pick two, start this week
Greg's closing point is the one to sit with: code used to be the moat. Now distribution is the moat. It's the one thing AI can't build for you.
The 7 plays:
1. MCP server (let AI sell for you)
2. Programmatic SEO (10,000 pages)
3. Free tool (the tool is the marketing)
4. AEO (be the source AI cites)
5. Viral artifacts (make output shareable)
6. Buy a newsletter (acquire the audience)
7. AI repurposing engine (one pillar, seven channels)
Don't do all seven. Pick the two that fit your product and give them a real month.
FAQ
What does "distribution first, product second" actually mean?
Build the audience before the product. Grow to even 1,000 followers in your niche, ask them what they need, then build it in a weekend and launch to a warm audience. The old way (build for months, launch to silence, add features, launch to silence again) is how most vibe coded projects die.
Is programmatic SEO still worth it in 2026?
Yes, but the bar moved. Google punishes thin AI slop, so the plays that work pair a real dataset (scraped with something like Firecrawl) with genuinely unique content per page and a human editing pass. Start with 100 pages, confirm indexation, then scale.
What's the fastest strategy on this list?
The free tool. With Claude Code you can build a grader or calculator in a day and ship it by lunch. It generates leads immediately, while MCP servers, AEO, and programmatic SEO compound over months.
How much does it cost to buy a newsletter?
Roughly $5,000 to $20,000 for 5,000 to 50,000 subscribers in a specific niche, based on the ranges Greg cites. Many small operators earn under $500 a month from their list, so fair offers get taken more often than you'd expect. Check Duuce.com or DM operators directly.
Do I need a big audience for the repurposing engine to work?
No. For-you-page algorithms distribute content on merit, not follower count. The strategy is volume: one weekly pillar becomes 20+ assets across platforms, which means 20+ chances per week to catch the algorithm. Shots on net.
Steal playbooks like this every week
I break down how real bootstrapped founders get customers and grow to $100K MRR and beyond, every week on the Profitable Founder Podcast. Real numbers, no fluff, founder to founder.
