Buy the plumbing, vibe the rest
I keep seeing the same post on LinkedIn. Someone proudly announces they canceled their company’s software subscriptions. They dropped their CRM vendor and used an AI agent to build a custom internal tool over the weekend.
They claim the era of buying software ended because anyone can prompt a dashboard into existence.
I worry about this. Not because it threatens SaaS companies, but because it creates a massive trap for the teams writing the code.
You absolutely should build your own apps. AI gives you the power to bypass massive corporate packages that force you into their specific workflows. You can build something entirely bespoke.
The problem starts a month later.
The mile-wide, inch-deep trap
When you vibe-code an internal platform from scratch, you usually skip the architecture. The first version looks incredible. It does exactly what you asked. Then you put fifty employees on the system.
Performance collapses. Security flaws surface immediately. You quickly realize the AI built an application that is a mile wide and an inch deep. You end up with features that almost work, but the moment a user hits an edge case, the system panics.
By canceling a subscription, you accidentally became a database administrator and a security engineer.
Buy the plumbing, vibe on top
The most resilient teams right now adopt a hybrid approach. They buy extremely stable headless services to handle the invisible complexity, and they use AI to build the experience on top.
You buy a platform like Contentstack for the content infrastructure. You let the providers handle the global scaling and the enterprise-grade security. Then you construct the exact interface you need.
I saw the writing on the wall a while ago, which is why I built the Contentstack Platform SDK as a personal passion project and open-sourced it. I wanted to hide the complex backend mechanics entirely. The SDK scaffolds the application based on your product needs. It hands you a solid foundation with the right typing, routing, and connection protocols already established.
Once that foundation exists, the workflow changes completely.
Vibe engineering with Claude Code
Because the platform SDK handles the heavy lifting, you can just open Claude Code and vibe-engineer the rest of the application.
You connect Contentstack’s MCP and custom AI skills directly to Claude to work your way through the project. You tell the agent to fetch the content schema via the SDK and render a custom dashboard. The AI stops guessing how to build a database from scratch. It simply orchestrates UI components against a proven, stable service.
This is the sweet spot.
Do not spend your weekend trying to prompt a secure backend into existence. Buy the services that know how to scale, and vibe on top of them. The actual advantage is paying for the plumbing so you never have to think about it again.
On this page
Related articles
The Doer Economy - we are killing the translator class
This article argues that AI is collapsing the distance between vision and execution, ushering in a Doer Economy where the primary winners are those who can both imagine and build. The traditional corporate stack of translators (product managers, marketers, and multiple layers of management) is shrinking because executional tasks are increasingly handled by AI. Founders and CEOs are moving closer to product, validating ideas directly with AI-generated scaffolds and modern SaaS primitives. For individuals, the value has shifted from narrow, ticket-driven skills to end-to-end ownership, product thinking, and understanding users. The future belongs to people who ship, iterate quickly, and leverage AI and SaaS platforms to focus on business logic and user experience, rather than those who only manage the builders.
- Composable architecture
- AI engineering
- API design
The case for boring setups
Most developers overconfigure their machines and pay for it in ways they don't notice. Every custom alias, every remapped key, every hand-rolled config file is a small tax on your ability to work anywhere other than your own laptop. The same pattern is repeating right now with AI tooling, dozens of MCP servers and stacked skill files that look like leverage and behave like drag. After getting stranded by my own setups more times than I'd like to admit, I've come around to a quieter belief: portability is a skill, and defaults are how you practice it.
- Composable architecture
- AI engineering
- Cloud & infra
Skills over MCP explained: how the new extension delivers workflows alongside tools
How the new MCP Skills Extension works, from skill discovery and on-demand loading to content verification, workflow guidance, and managed profiles.
- Composable architecture
- AI engineering
- API design