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

In the realm of startup development, the well-known wisdom shared by Paul Graham, “Do things that don’t scale,” can often feel challenging, especially when it comes to coding. As someone who has devoted the last eight months to building an AI podcast platform, I’ve devised a straightforward approach to implementing this principle effectively: the 3-Month Rule. Each unconventional, unscalable hack has a lifespan of three months. At the end of those three months, it either proves its worth and is developed into a robust solution, or it is discarded.

As engineers, we are often conditioned to focus on scalable solutions right from the outset. Concepts such as design patterns, microservices, and distributed systems are staples in our toolkit, created to accommodate vast user bases. However, this mindset often belongs to larger companies with substantial resources.

In a startup environment, emphasizing scalability too soon can be a form of costly procrastination. You’re often optimizing for users that don’t yet exist and resolving issues that may never arise. By adhering to my 3-Month Rule, I commit to creating straightforward and sometimes “imperfect” code that is actually deployable, allowing me to better understand the needs of my users.

My Strategic Infrastructure Hacks: Simplicity Meets Intelligence

1. Consolidated Resources on One VM

Currently, my entire stack—database, web server, background processes, and caching—runs on a single $40/month VM without redundancy and relies on manual local backups.

This setup may seem reckless, but it has enabled me to gain insights into my actual resource usage far beyond what any capacity planning document could provide. I discovered that my “AI-centric” platform peaks at only 4GB RAM. I was able to bypass setting up a complex Kubernetes instance, which would have required me to manage empty containers. Moreover, when system crashes happen (and they have, a couple of times), I glean real-time data on what goes wrong—often surprising me with the unexpected.

2. Direct Configuration Management

Instead of utilizing configuration files or environment variables, I define constants directly within my codebase:

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

While this may seem like a primitive method, it allows