Stop Letting AI Agents Pick Your Tech Stack

Key Concepts

• Vibe Coding: A modern, AI-assisted approach to software development where the developer uses AI agents to generate code and architecture.
• Tech Stack: The combination of programming languages, frameworks, databases, and cloud services used to build an application.
• CLI (Command Line Interface): A text-based interface that allows AI agents to interact directly with cloud infrastructure, automating complex deployment and management tasks.
• Wrapper Services: Third-party platforms that provide a simplified interface for underlying cloud infrastructure (e.g., Firebase as a wrapper for Google Cloud Platform) often at a significant price markup.
• Open Source vs. Paid API: The distinction between using free, community-maintained code (like FFmpeg) versus proprietary, paid services (like 11 Labs or Alpaca).
• Ecosystem Synergy: Choosing tools and languages that are natively compatible or owned by the same parent company (e.g., Flutter and GCP) to ensure smoother integration.

---

1. Identifying the Target Consumer

The first step in selecting a tech stack is defining the user experience.

• Web vs. Mobile: Determine if the application is "web-first" (URL-based) or "mobile-first" (app-based).
• Case Study (5 Cent Club): The creator identified that a brokerage platform requires a mobile-first approach for trading, while a web presence is primarily for marketing and landing pages.
• Strategic Decision: Avoid building for platforms that don't fit the user's workflow. If the design process is desktop-centric, don't force a mobile-native version unless necessary.

2. Simplifying the Ecosystem

AI models often suggest fragmented stacks with too many disparate services.

• The "Simpler is Better" Rule: Aim to centralize your backend on one or two major platforms (e.g., Google Cloud Platform and Cloudflare) rather than managing 10+ different logins.
• The CLI Requirement: Always verify that your chosen tools have a robust CLI. This allows AI agents to perform actions on your behalf, reducing the need for deep manual cloud infrastructure knowledge.
• Gut-Checking AI: When an AI suggests a stack, ask it to "simplify and centralize." If it suggests a service, research if it is merely a "wrapper" for a cheaper, more foundational service (e.g., using AWS SES instead of Resend).

3. Free vs. Paid APIs

Distinguish between "value-add" services and "infrastructure" costs.

• Open Source Alternatives: Tools like FFmpeg provide powerful functionality (e.g., audio/video transcription) for free. The only cost is the "rented computer" (cloud server) time required to run the code.
• Necessary Paid APIs: Some services, like Alpaca, are worth the premium because they handle complex legal, compliance, and regulatory requirements that would be prohibitively expensive or difficult to build from scratch.
• Actionable Insight: Always ask the AI, "Is there an open-source version of this?" before committing to a paid API subscription.

4. Framework Selection and Bias

AI models are often biased toward popular frameworks like React Native/Expo because they have been trained on more data regarding these tools.

• The Flutter Advantage: For the 5 Cent Club, the creator chose Flutter over React Native.
  • Cross-Platform: Allows deployment to both iOS and Android from a single codebase.
  • Ecosystem Alignment: Flutter is part of the Google ecosystem, making it highly compatible with Google Cloud Platform (GCP).
• Native vs. Cross-Platform: While native languages (like Swift for iOS) offer the best UI performance, they require separate codebases for Android, which increases development overhead.

5. Synthesis of the Workflow

To build a robust tech stack, follow this logical progression:

1. Define the Consumer: Decide between web or mobile based on user behavior.
2. Brainstorm with AI: Use prompts to suggest a stack, but explicitly instruct the AI to "stop planning" and "just talk" to avoid biased, fragmented suggestions.
3. Verify Infrastructure: Ensure the chosen tools have a CLI and are not just expensive wrappers for foundational services.
4. Check for Synergy: Ensure the frontend language (e.g., Flutter) communicates efficiently with the backend (e.g., GCP).
5. Document: Create an .md (Markdown) file to formalize the tech stack, ensuring a clear, unified foundation for the project.

"The less platforms you have when building out your software, the better." — Corbin