Concept

Amjad's Law

concept coding-skills ai-leverage learning

Amjad’s Law

“The return on investment for learning to code is doubling every six months.”

A principle coined by Amjad Masad (co-founder of Replit) to describe the accelerating leverage that AI tools place on even basic coding literacy.

The argument

Pre-AI, the value of “knowing how to code” was roughly proportional to coding depth: a beginner could do little, an expert could do a lot. The investment required to reach productive competence was high.

With AI coding agents:

  1. The minimum viable skill threshold drops — prompting, reading code, and basic debugging are now sufficient to build functional software.
  2. AI amplifies whatever baseline exists — a user who can read and debug code guides an agent far more effectively than one who cannot.
  3. Each model improvement amplifies the existing skill base further.

The combination of a lower entry cost and rising amplification factor means the ROI curve steepens rapidly. Amjad frames this as doubling every six months — not a precise empirical claim, but a directional assertion that the compounding is faster than most people expect.

Practical implications

For non-engineers (PMs, designers, founders):

  • Learning to code via traditional bootcamp methods (Git, algorithms, frameworks from scratch) is not the right model.
  • The productive learning path is: prompt AI → read the output → try to unblock it when it gets stuck → learn what broke and why.
  • The skills that carry the most leverage: prompting, code reading, debugging mental models, understanding what a server / API / database does.

For engineers:

  • Deep coding skill remains valuable — at the ceiling (system architecture, database migrations at scale, security) AI still needs expert guidance.
  • The differential between an engineer who embraces AI tooling and one who does not is compounding rapidly.

Caveats

The “doubling every six months” figure is not derived from measurement. It is a rhetorical device to convey accelerating returns. The mechanism — AI leverage multiplies a skill baseline that keeps lowering its entry point — is the more defensible claim. The specific cadence should be treated as directional, not quantitative.

See also