TLTD #29 - Vibe Coding Your Way to a Seed Round
A new founder archetype
As my colleagues will attest, I’m the first one to wince when someone mentions vibe coding next to me. But I can’t deny the impact it is having in our industry.
I have a friend who’s built a high-quality MVP vibe coding his way through PRDs like there’s no tomorrow! Seeing what someone without a coding background can do has been eye opening.
Using a combination of ChatGPT, Claude, and Cursor, he built a functioning prototype, including a backend and a native iOS app, in a ridiculously short amount of time. This would probably have taken me months to do without AI, or cost me thousands of dollars in outsourcing!
To him, it cost few hundred bucks in subscription fees and a lot of late nights learning to talk to AI. (you can see his app for yourself here).
And he is far from being an isolated case…
The Relentless March Toward Zero-Cost Prototyping
When you step back and look at the history of software development, you can spot a trend: the cost and barriers to building an MVP have been steadily approaching zero for decades:
The Physical Era (1980s-90s): Remember when software was distributed on physical media? You’d need to burn it onto floppy disks or CDs. Mistakes were expensive, with single bugs potentially costing millions in product recalls. Not to mention piracy was rampant!
The Web (Late 90s-2000s): The Web dramatically reduced distribution costs, but you still needed significant infrastructure investment. Companies had to purchase servers, configure them in office closets, or rent expensive data centre space.
Infrastructure as a Service (Mid 2000s): AWS and similar platforms arrived, eliminating the need for physical hardware. This was revolutionary, but still required substantial technical knowledge to properly configure and manage cloud resources.
Platform as a Service (Late 2010s): Services like Heroku emerged, abstracting away infrastructure complexity. Developers could focus on building products rather than configuring servers, a significant leap forward.
The AI Era (2023-Present): With ChatGPT, Claude, and AI-powered development platforms like Replit, Lovable, and v0, we’ve reached a new milestone: the cost of building an MVP has approached near zero, both in terms of money and technical expertise required.
Each wave democratised software development further, but this latest shift might be the most impactful yet.
Enter vibe coding: now founders can translate ideas into working prototypes by describing what they want to AI tools in natural language rather than code. It offers an alternate path in how early-stage ventures come to life, and it’s challenging our traditional notions of technical founders, creating MVPs, and early-stage funding.
The Vibe Coding Iceberg
Let me be clear upfront: vibe coding won’t replace software engineers. Anyone who’s shipped production systems used by thousands or millions of people knows this is absurd.
Besides the maintenance nightmare, the security implications are very real, with 170 out of 1,645 Lovable-created web apps suffering from the same glaring security flaw, allowing anyone to access information about the app’s users, including names, email addresses, and secret API keys.
vibe coding won’t replace software engineers
Robust engineering practices, performance optimisation, system architecture, security hardening, and scalability all require skilled engineers who understand software fundamentals. The jump from prototype to production is massive, and AI tools can only take you so far (at least for the foreseeable future).
What vibe coding does change, however, is the barrier to entry for testing ideas. And that’s huge!
Dirt Cheap Prototyping
Before generative AI, turning an idea into a testable product required you to have the chops to write code yourself, convince a technical co-founder to join you or forking out $30K-$100K hiring developers to do it for you.
But things have changed. Today, a determined founder with domain expertise can use AI tools to:
Create basic web and mobile applications
Create simple backend APIs
Design user interfaces
and much more…
But don’t take my word for it. According to recent research, 89% of AI startup founders are using generative AI tools in their development processes. More tellingly, AI app building platforms like Lovable are reporting crazy adoption rates, with 30,000 paying customers building 25,000 new product apps per day.
Case Study: From Idea to Seed in 60 Days
Take LOOK AI, a fashion search engine founded by a non-technical solo entrepreneur in Europe. Using Lovable, the founder created what they described as a “Perplexity-for-shopping” prototype that could search across fashion catalogs intelligently.
The working prototype was compelling enough to secure $500K in pre-seed funding. The key? Investors could see, touch, and test a real product rather than evaluating a pitch deck.
Or consider Omnispay, a UAE-based fintech startup founded by domain experts without coding backgrounds. They used Replit to build a functioning payments platform for small businesses. This prototype secured them a $1.5M seed round, with investors noting how impressed they were that the team could demonstrate a working product from day one.
The End of Non-Technical Co-Founders?
While these tools don’t eliminate the need for engineers, they do challenge the concept of the “non-technical co-founder”. When any motivated founder can build a working prototype, the traditional technical/non-technical dichotomy no longer holds.
How effective you are at using AI is now a technical skill set
In fact, we’re seeing a new archetype emerge. These are domain experts who would otherwise not have considered starting a business but are now using AI as their technical co-founder during the earliest stages. They understand their industry in detail and can translate that expertise into product requirements that AI tools can implement.
This trend is already influencing how VCs evaluate early-stage companies. Anton Osika, CEO of Lovable, says VCs are increasingly scouting founders who effectively use AI tools despite not being traditional coders: “How effective you are at using AI is now a technical skill set”.
What Investors Are Looking For Now
Multiple VCs have publicly stated they're rethinking the importance of a founder having coding expertise, especially at the earliest stages.
The new profile they’re looking for:
Domain expertise: Deep understanding of the problem space
AI fluency: Ability to effectively use AI tools to execute on vision
Product intuition: Clear understanding of user needs and product requirements
Learning velocity: How quickly founders can adapt to new tools and approaches
As I’ve heard in some circles: “I’m less concerned about whether the founder can code and more interested in whether they can build.”
Enterprise Implications
This trend extends beyond startups. Within enterprises, product managers are already using these same tools to build business cases with functioning software rather than slides and wireframes.
Imagine a product manager at a large financial institution who wants to propose a new customer onboarding flow. Rather than creating mockups and documentation, they can now build a working prototype using AI tools. When they present to leadership, executives can click through a real interface rather than imagining how wireframes might function.
product managers are using these tools to build business cases with functioning software
This changes how internal product decisions are made:
Higher-fidelity communication: Working software speaks louder than requirements docs
Faster validation cycles: Ideas can be tested with users before committing significant engineering resources
Better resource allocation: Engineering teams focus on productionising validated concepts
This trend blurs the line between software engineering and product management somewhat. I believe product teams will be better for it, with product managers who are now closer to the medium engineers use on a daily basis.
A New Founder Toolkit
For aspiring founders, learning to effectively use AI coding tools is a must and just as important as understanding your market. It unlocks your ability to execute independently.
Here’s what your toolkit likely needs to include:
AI coding assistants: Tools like Cursor or GitHub Copilot for code generation, completion and general assistance
No-code/AI platforms: Solutions like Lovable, Replit, v0 or similar for full-stack applications
Prompt engineering skills: The ability to clearly articulate what you want AI tools to build
Basic technical literacy: Enough understanding to evaluate and adapt what AI spits at you
Most importantly, founders need to understand what these tools can and cannot do. They excel at creating functional prototypes but will struggle with optimisation, scalability, and security. Engineers aren’t going to be replaced anytime soon.
Moving Faster with AI
Trying new ideas, either as a founder or within an enterprise, has never been cheaper. This capability is only limited by our vision and perseverance. But more importantly, it should allow us to move faster and quickly weed out bad ideas.
we are only limited by our vision and perseverance
This doesn’t mean engineering talent is less valuable. If anything, it means human engineers can focus on what they do best: building robust, scalable, and secure systems. Besides, professional engineers are leveraging AI in their own way, as I’ve written about before.
Business leaders need to understand this shift and invest in the tools that will make their people more productive, creative and fulfilled . The tools to build have never been more accessible, and the ability to vibe code your way to a seed round is very compelling.