Embracing the 3-Month Rule: A Pragmatic Approach to Non-Scalable Solutions in Tech Development
In the world of startups, the gold standard advice from Paul Graham resonates well: “Do things that don’t scale.” However, the vital question remains—how does one effectively put this advice into practice, especially in the realm of coding? After spending eight months building my AI podcast platform, I’ve crafted a succinct methodology that I like to call the “3-Month Rule.” This approach allows each non-scalable hack a trial period of three months to either demonstrate its worth and graduate to a robust solution or be phased out entirely.
Redefining Our Engineering Mindset
As engineers, we often start by envisioning scalable solutions, focusing on sophisticated design patterns, microservices, and distributed systems. While these concepts are essential for large organizations, they can become irrelevant in a startup scenario where the focus should be on immediate user needs rather than hypothetical future demands. My framework encourages me to adopt a straightforward coding style that risks imperfection but prioritizes learning and adaptation.
My Current Practical Hacks and Their Insights
1. Simplistic Infrastructure: One VM for Everything
I’m currently running my entire platform on a single virtual machine—web server, database, background jobs, and caching—all consolidated into a $40/month setup. While it sounds risky, this approach has revealed crucial insights about my platform’s resource consumption more effectively than traditional capacity planning strategies. With peaks at just 4GB of RAM, I’ve realized the complexity of Kubernetes would have been an unnecessary overhead.
When technical difficulties arise, I gain real-time understanding of failures, providing valuable data that helps inform future adjustments.
2. Utilizing Hardcoded Configurations
My current approach employs hardcoded constants for configuration management—think straightforward pricing tiers and user limits embedded directly within the codebase. This facilitates quick tracing and tracking through version control, eliminating the need for separate config files. The slight inconvenience of redeployment has proven vastly more efficient than dedicating an entire week to build a configuration service that may hardly be used.
3. Running SQLite in Production
Using SQLite as the cornerstone database in a multi-user web application was a deliberate choice. My entire database size is manageable at just 47MB, adeptly handling 50 concurrent users. This peculiar setup has unveiled usage patterns skewed heavily towards reads, affirming SQLite’s suitability. Had I opted for a more complex solution like Postgre
