Embracing the 3-Month Rule: A Strategized Approach to Non-Scalable Solutions

In the tech community, the advice from Paul Graham to “do things that don’t scale” resonates profoundly, yet few discuss the practical implementation of this concept in coding. After dedicating over eight months to developing my AI podcast platform, I have established a straightforward framework: every unscalable hack is given a lifespan of exactly three months. At the end of this period, it must either demonstrate its value and be transformed into a more robust solution or be discarded.

A Shift in Perspective on Scalability

As engineers, we often approach problems with a mindset geared towards scalability from the outset. We’re drawn to design patterns, microservices, and distributed systems—impressive architectures designed to support millions of users. However, this mindset can lead us to focus on hypothetical issues rather than the immediate needs of our users. I’ve found that applying my 3-month rule compels me to write straightforward, albeit inefficient, code that delivers results and helps me understand user behavior more effectively.

Current Infrastructure Strategies: A Smart Approach

Let’s delve into my current infrastructure hacks that, while seemingly suboptimal, have proven to be quite strategic:

1. Consolidation on a Single VM

I manage my entire platform—database, web server, background jobs, and even Redis—on a single $40-per-month virtual machine with no redundancy and manual backups to my local environment.

Why this approach is valuable: Over the past two months, I have gleaned insights about my resource requirements that no planning document could have provided. My platform’s peak memory utilization hits just 4GB. Had I opted for a complex Kubernetes setup, I would have devoted time to managing excess infrastructure rather than addressing actual needs.

In the few instances of crashing (two so far), I gathered real data about what went wrong—often surprising me.

2. Simplistic Configuration Management

My configuration settings are hardcoded directly in the codebase:

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

No config files or environment variables—just constants.

The advantage: I can quickly search my entire codebase for any configuration value. Changes are tracked through version control, which allows for an easy review process, albeit solo. Creating