What You'll Learn
The Great Migration: From Visual Builders to AI-Powered Development
The no-code movement promised to democratize software development, and for years it delivered on that promise. But in early 2026, something fundamental has shifted. AI-powered coding tools like Claude Code aren't just competing with visual builders—they're fundamentally changing how we think about building software.
Kieran's journey from no-code advocate to AI coding convert illustrates this transformation perfectly. Just six months after publishing a blog post explaining why he wasn't sold on vibe coding, he found himself building four production SaaS applications in six weeks using Claude Code.
The Tipping Point: When Skepticism Becomes Curiosity
What changed? Not a single eureka moment, but an accumulation of evidence that became impossible to ignore. After attending Create With meetups in London and Bristol, watching community members demonstrate increasingly sophisticated projects built with AI coding tools, Kieran reached his tipping point.
The catalyst came from seeing Tanbe's demonstration at a London meetup, showcasing complex applications built entirely through natural language prompts. The key difference from earlier attempts at AI coding? Claude Code's ability to debug itself.
Previously, encountering a terminal error meant hours on Stack Overflow, copying and pasting solutions without fully understanding them. Now, Claude watches for errors, identifies problems, tries alternative solutions, and iterates until things work. This autonomous debugging capability removes the primary barrier that kept non-traditional developers away from code.
Real-World Test: Replacing a $900/Year Platform
The first major project proved Claude Code's capabilities weren't just theoretical. Kieran received a renewal notice from Podia for $900 annually—a significant cost for hosting video courses. Rather than accept the expense, he asked Claude to build a replacement.
The resulting custom platform includes:
- Video hosting and streaming infrastructure
- User authentication and payment processing via Stripe
- Automatic video transcription
- AI-powered chat interface for course content
- Auto-generated quizzes from transcripts
- Full development, staging, and production environments
What's remarkable isn't just that these features were built, but how quickly. The AI chat feature took five minutes to implement. The development environment configuration—something that would have required extensive research and setup—was explained and implemented in a single session.
From Static Sites to Internal Automation
The learning curve continued with lower-stakes projects: rebuilding static websites previously hosted on Framer and a blog on Feather ($39/month). Claude Code pulled content from Notion via API, downloaded images, generated markdown files, and deployed everything to Vercel.
The result? Websites with 100/100 Lighthouse performance scores, zero backend complexity, and near-zero hosting costs. More importantly, complete control over every aspect of functionality.
This success led to more ambitious internal tooling. Kieran began building a comprehensive dashboard for Create With operations, automating newsletter curation by monitoring AI news sources and ranking stories by relevance. The system consolidates workflows previously spread across Zapier, n8n, Google Docs, and Notion into a single custom interface.
The New Paradigm: Marketing Over Building
This shift has profound implications for software businesses. When building becomes dramatically easier and cheaper, what becomes the differentiator? Marketing, positioning, and understanding customer needs.
The conversation reveals an emerging philosophy: the hard part of SaaS is no longer the engineering—it's finding product-market fit, communicating value, and building distribution. Technical execution, once the highest barrier, has become a commodity skill.
Andrew Bryce, featured at a Bristol meetup, exemplifies this new model. He built an extremely feature-rich internal system for his company entirely with Claude Code, replacing multiple enterprise tools with custom solutions that perfectly fit his workflow.
Understanding the Stack: Real Development Workflows
One crucial aspect of this transition: Claude Code doesn't eliminate the need to understand development concepts. Kieran had to learn about:
- Version control with GitHub
- Deployment pipelines with Vercel
- Environment management (local, staging, production)
- API integration patterns
- Database architecture decisions
The difference is the learning curve is compressed, and you learn by doing rather than studying. You understand these concepts in context because you're implementing them for your specific use case, with an AI assistant that explains as it builds.
The Security and DevOps Question
One area that remains challenging is security and DevOps. While Claude can implement authentication, set up SSL certificates, and configure deployment pipelines, understanding security implications requires human judgment. The podcast discussion acknowledges this as "the final frontier"—areas where AI assistance is valuable but human oversight remains critical.
Building Mobile Apps and Beyond
The capabilities extend beyond web applications. Vibe Code iOS enables building and testing mobile applications directly on an iPhone, further lowering barriers to multi-platform development. Combined with Shopify's development MCP (Model Context Protocol) tools, developers can create e-commerce integrations in hours rather than weeks.
What This Means for the No-Code Ecosystem
This isn't necessarily the death of no-code tools—it's an evolution. Visual builders will likely remain relevant for:
- Rapid prototyping and validation
- Non-technical team members making updates
- Specific use cases where visual logic is clearer
- Situations where platform ecosystems provide value
But for builders willing to invest a few weeks learning fundamental development concepts, AI coding tools offer unprecedented flexibility, lower long-term costs, and complete control over their technology stack.
The Learning Path Forward
For those considering this transition, Kieran's approach provides a template:
1. Start with static websites to learn deployment basics
2. Progress to simple applications with minimal backend logic
3. Build internal tools before customer-facing products
4. Accept that some projects will be learning experiences, not launches
5. Use community resources and meetups to see what's possible
The key insight: this is the worst AI coding will ever be. Models will continue improving, making complex features easier to implement and debugging more autonomous. The time to start learning is now, while the barrier to entry is already lower than it's ever been.
Conclusion: A New Era of Software Development
We're witnessing a fundamental shift in who can build software and how quickly they can do it. The combination of powerful AI models, improved tooling, and emerging best practices is creating a new category of builder—not quite no-code, not traditional developers, but something in between.
Kieran's journey from skeptic to advocate in six months demonstrates that this transition is accessible to anyone willing to push past initial uncertainty. The question isn't whether to learn these tools, but how quickly you can start applying them to real projects.
As we move through 2026, expect this trend to accelerate. The builders who adapt earliest will have significant advantages—not just in what they can build, but in how efficiently they can test ideas, iterate on feedback, and bring products to market.
