Embracing the 3-Month Rule: A Pragmatic Approach to Unscalable Solutions in Tech

In the world of entrepreneurship, there’s a well-known piece of advice from Paul Graham: “Do things that don’t scale.” Though this advice is often cited, the challenge lies in effectively applying it, particularly within technical environments. After eight months of developing an AI podcast platform, I have crafted a straightforward framework that I like to call the “3-Month Rule.” This approach allows every unscalable solution a three-month testing period—after which, we either upgrade it because it has proven its worth, or we discard it entirely.

As engineers, our instincts often push us toward building scalable solutions right from the start—think microservices, fancy architecture, and vast systems designed to accommodate millions of users. However, this mindset can lead to costly delays, especially in a startup where resources are limited. By implementing my 3-Month Rule, I consciously focus on developing straightforward, albeit “bad,” code to quickly learn what my users genuinely need.

Key Infrastructure Strategies and Their Benefits

1. Streamlining on a Single VM

Currently, my entire system—database, web server, and background tasks—operates on a single, affordable $40/month virtual machine. While it may seem reckless with no redundancy, this approach has given me invaluable insights into my actual needs. In just two months, I’ve realized that my AI-driven platform uses a maximum of 4GB of RAM during peak times. Had I set up a complex Kubernetes infrastructure, I would have ended up managing containers that went unused.

Experiencing system failures (which have happened twice) has provided me with real data regarding the causes—insight that no preliminary setup could have predicted.

2. Utilizing Hardcoded Configurations

Instead of using configuration files or environment variables, I rely on hardcoded constants like:

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

This approach simplifies tracking changes—each modification is noted in git history and is reviewed, even if it’s just by myself. While a comprehensive configuration service might take weeks to develop, the reality is that I’ve only needed to change these constants three times over three months. That’s a mere 15 minutes spent on redeployment instead of 40 hours on engineering.

**3. Opting