The 3-Month Rule: A Practical Approach to Non-Scalable Solutions in Tech

The well-known advice from Paul Graham, “Do things that don’t scale,” is often discussed, yet its practical application in coding remains a mystery for many developers. After dedicating the last eight months to creating my AI podcast platform, I’ve established a straightforward framework based on this principle: any unscalable practice is given a three-month trial period. At the end of this timeframe, if it proves its worth, it gets a proper build-out; if not, it’s eliminated.

As engineers, we’re typically trained to devise scalable solutions from the outset. We gravitate towards sophisticated design patterns, microservices, and distributed systems—architectural marvels capable of managing millions of users. However, this perspective often aligns more with large corporations rather than the nimble, adaptive approaches required in startups.

In a startup environment, focusing too much on scalability can lead to costly delays, as it involves planning for users and challenges that may never materialize. Implementing my three-month rule compels me to write simpler, more direct code that can be deployed promptly, enabling me to gather crucial insights about actual user needs.

Current Infrastructure Practices: A Smart Strategy

1. Consolidated Operations on a Single Virtual Machine

I run my entire stack, including the database, web server, background processes, and caching, on a single $40-per-month virtual machine—without redundancy and handling backups manually.

This approach has proven more insightful than any extensive planning document. In just two months, I’ve gained a clearer picture of my resource usage—my “AI-heavy” platform peaks at 4GB of RAM. The complex Kubernetes architecture I contemplated would have only required me to manage empty containers.

When downtime occurs (twice so far), I receive data on the actual points of failure, which is never what I anticipated.

2. Hardcoded Configuration Settings

python
PRICE_TIER_1 = 9.99
PRICE_TIER_2 = 19.99
MAX_USERS = 100
AI_MODEL = "gpt-4"

With constants scattered throughout my code, there are no separate configuration files or environment variables. To make adjustments, I simply redeploy.

This method’s true advantage lies in its simplicity: I can search my entire codebase for any configuration value swiftly, and every change is meticulously tracked in git history. Over three months, I’ve updated these values just