Problem Decomposition
This is a member-only chapter. Log in with your Signal Over Noise membership email to continue.
Log in to readModule 3 · Section 1 of 6
Most people approach AI the same way they approach a search engine: fire off a single request and hope for the best. “Write my business proposal.” “Plan my team offsite.” “Summarise this report.” When the result comes back thin or off-target, they blame the tool — or assume they need better prompts.
The actual problem is usually upstream. Before you type anything, you need to have done the thinking the AI can’t do for you: breaking the problem into its real parts.
Decomposition is the skill of splitting a complex problem into smaller, manageable pieces. It’s one of the five foundations of computational thinking, and it’s arguably the most important one for working effectively with AI. Not because it makes prompts fancier — because it makes the task itself clearer. When you break “write my proposal” into distinct components — context, audience, problem being solved, section structure, desired tone — you’ve already done most of the hard work. The AI fills in the substance; you provide the shape.
This module explores decomposition through four different lenses: the way magicians structure illusions, the constraint-driven logic of tangram puzzles, the four-part vocabulary computational thinking gives you for any problem, and the engineering choices that took India to Mars on a fraction of the typical budget. Each one sharpens the same underlying skill. By the end, you’ll have a reliable way to approach any AI task — not by prompting better, but by thinking first.