Finding and Nurturing the 0.1% Who Could Be Great Software Developers
There is a persistent belief in the software industry that great developers are rare. Companies fight over the same pool of experienced engineers, salaries inflate, and hiring managers complain about the “talent shortage.” But the shortage is not what it appears to be. The world is not running out of people who could be great developers. We are failing to find them.
The Discovery Problem
Consider the path a typical software developer takes to their first job: exposure to computers at a young age, access to education (often a university CS degree), encouragement from people who recognize their aptitude, and the economic stability to invest years in learning before earning. Each of these prerequisites filters out enormous populations of potential developers.
A kid in a rural area with no CS program at their school. A teenager whose family needs them to earn immediately after high school. A career changer in their 30s who cannot afford a two-year gap in income. A person from a culture where programming is not presented as a viable career. These are not people who lack aptitude — they lack the specific sequence of opportunities that the industry’s standard pipeline requires.
If you accept that programming aptitude is roughly normally distributed across the human population — and there is no good reason to believe otherwise — then most of the people in the right tail of that distribution never encounter the circumstances that would lead them to programming.
What Aptitude Actually Looks Like
Great software developers share cognitive traits that are not specific to programming: strong abstract reasoning, pattern recognition, comfort with ambiguity, the ability to hold complex systems in working memory, and a specific form of stubbornness that keeps them debugging a problem for hours instead of giving up.
These traits show up in many places outside of CS departments. The mechanic who diagnoses intermittent engine problems through systematic elimination. The chess player who can think seven moves ahead. The musician who hears a complex piece and understands its structure. The farmer who optimizes irrigation systems through experimentation. The aptitude is there. The pathway to software development is not.
Why the Standard Pipeline Fails
University CS programs are effective at training developers who already have the background to get admitted. They are not effective at discovering aptitude in populations that never apply. Admissions criteria (test scores, prior coursework, economic access) filter on privilege as much as potential.
Bootcamps expanded access significantly but still require students to have savings, time, and geographic proximity (or reliable internet for remote programs). The typical bootcamp student is a middle-class career changer in their 20s or 30s — a broader demographic than CS programs but still a narrow slice of the potential talent pool.
Self-teaching is the most accessible path but also the most uneven. Free resources exist (freeCodeCamp, The Odin Project, countless tutorials), but self-teaching requires motivation without external structure, persistence without feedback, and the ability to navigate a vast and often contradictory landscape of information. Many people with high aptitude lack the specific meta-learning skills required to teach themselves programming from scratch.
What Effective Discovery Looks Like
The organizations that are best at finding untapped talent share common approaches:
Screen for aptitude, not experience. Instead of requiring prior programming knowledge, test for the underlying cognitive abilities: logic puzzles, pattern matching, problem decomposition. Candidates who score high on these assessments but have no coding experience are exactly the people the standard pipeline misses.
Provide structured pathways with economic support. The biggest barrier to entry is not aptitude or interest — it is economics. Effective programs provide stipends, living expenses, or income share agreements that allow participants to invest time in learning without financial ruin.
Pair newcomers with mentors. Self-teaching fails for many people because they get stuck and have no one to ask. A mentor who can answer questions, explain concepts, and model professional practices dramatically accelerates the learning curve.
Create on-ramps, not gates. Traditional hiring processes (whiteboard interviews, algorithm challenges, pedigree screening) are gates that filter based on preparation rather than ability. On-ramps — apprenticeships, trial projects, pair programming sessions — give candidates with non-traditional backgrounds a chance to demonstrate capability in context.
The Economic Argument
This is not purely an equity argument, though the equity case is compelling. It is an economic argument. If the software industry’s talent pool is drawn primarily from CS graduates of selective universities in wealthy countries, and the actual distribution of programming aptitude is much broader, then the industry is leaving enormous value on the table.
A company that finds and develops a great developer from a non-traditional background does not just fill a headcount — it gains a team member with problem-solving skills forged in a different context, a perspective that the rest of the team lacks, and (often) a loyalty born from having been given an opportunity that others withheld.
The Scale of the Opportunity
Global population: 8 billion. If 0.1% have the aptitude to be great software developers, that is 8 million people. The total number of professional software developers worldwide is approximately 27 million, and most of them are not in the “great” category. The math suggests that the majority of potentially great developers are not writing code.
Finding even a small fraction of those people — and giving them the pathway that circumstances denied them — would do more to solve the “talent shortage” than any amount of salary inflation or immigration policy. The talent exists. The discovery mechanism is broken.