Software Engineering

A thing you build can be used by one person or a billion, and the marginal cost of the billionth copy is essentially zero. No other field has quite that property, and it is the reason a single person at a kitchen table can build something that reaches the whole world.

It is tempting to call the structural pull Creation, and Creation is genuinely half the life of the work. But the honest version of the field, the one a 17-year-old needs to hear, is that the dominant daily pull is Resolution. Software does not stay built. It breaks — constantly, in ways nobody predicted, across layers of systems other people wrote years ago and documented in a hurry. The running joke in the field is that you spend twenty percent of your time writing new code and eighty percent figuring out why existing code does not do what it claims. The defining act, the thing that fills most working hours and separates the people who last from the people who burn out, is the move from broken to functional: tracing a failure through a system you only partly understand, forming a hypothesis, testing it, and making the thing work again. The people who thrive are the ones who find the debugging as satisfying as the building, who like the puzzle as much as the product.

Creation, Organization, and Spread sit alongside that Resolution core. Creation is the build-from-nothing feeling, the first moment the thing you wrote does what you intended. Organization is the pull underneath good architecture, the discipline of taking a messy problem and giving it a clean structure that other engineers can reason about and extend. Spread is unusually strong in software because of that zero-marginal-cost property: a tool you make for yourself can become a tool for everyone, which is why the field produces both tiny personal projects and planet-spanning infrastructure, and why a flaw shipped at scale does not cause one bad outcome but millions at once. Most engineers move between all four pulls in a single week, and the ratio between them is one of the things that most distinguishes one software role from another.

The daily texture is quieter and more interior than the popular image suggests. The workspace is wherever the laptop opens, the light is screen light, and the primary sensory reality is visual-symbolic — text on dark backgrounds, scrolling code, terminal windows, browser tabs stacked like geological layers. Headphones are a boundary that means "I am in it, do not talk to me." Time runs on the deploy schedule and the sprint deadline rather than the clock on the wall, which is a real freedom and also a trap, because the work is always available and the boundary between working and not working is something you have to build yourself. The physical cost is sedentary and cumulative; the people who last figure out their ergonomics and their movement routine early, or they pay for it later.

Software engineering has one of the most open credentialing landscapes of any high-paying field, and that is both its promise and its complication. There is no single licensing gate. People enter through computer-science degrees, through coding bootcamps, and through years of self-directed learning backed by a portfolio of real projects, and the gap between what hiring processes say they want (specific degrees, specific frameworks) and what the work actually requires (the ability to reason under uncertainty, to read other people's code, to keep learning when last year's tools are already outdated) is wide. AI is now reshaping the field's own entry path in real time. Tools that generate and explain code are compressing some of what junior engineers used to do while expanding what a single skilled engineer can attempt, and the value is visibly shifting away from raw code-writing speed toward judgment about what to build and why. The field remains in genuine demand, the building-from-nothing feeling is real, and the work is more "patient detective who types" than "visionary architect" on most days.

People and jobs. Aggregate BLS projection positive (15% growth through 2034); entry-level contracting. Field restructuring (junior tier compressing, mid-level expanding) rather than net shortage or surplus at aggregate level. [happening now — BLS; Stanford/ADP]

The main limit on getting in. Entry-level demand contracting primarily because AI automates the tasks that justified junior hiring. The credential gate (CS degree) remains but is no longer sufficient differentiation. The binding constraint: what can you do that AI cannot, or cannot yet do reliably? AI fluency now in 42% of job descriptions (up from 8% in 2022).

Entry-level jobs. 28% decline in entry-level postings from 2022 peaks; Stanford/ADP -20% employment for ages 22-25; new grads 7% of Big Tech hires vs 32% in 2019; Handshake internships -30% since 2023. [happening now — multiple sources]

Cost of training. CS degrees carry moderate cost relative to expected income. Bootcamp and self-taught paths have low credentialing cost. Debt-to-income is not a structural burden. 43% salary premium for engineers with AI skills.

Unlike shortage fields where AI extends capacity with a positive augmentation narrative, SWE is a balanced-to-contracting field where AI compresses the entry ramp. This is the field where the AI displacement story is most directly relevant to PurPassion's audience — not as an abstraction but as a documented cohort-level effect.