Part One: Read the market honestly · Chapter 2
What AI took away.
AI reduced the value of some traditional junior tasks. It also made a different kind of small, outcome-focused work commercially viable.
I am not going to tell you that AI will not affect your career. You would be right to stop trusting me if I did.
AI changed some things. To make sensible decisions about your next move, you need to be precise about what those things are.
What has gone
The boilerplate tier stopped being a reliable way in.
The work was never exciting, but it was billable—and for years it was how junior developers joined teams.
01
Boilerplate implementation
Project skeletons, standard forms, CRUD screens, routine markup and tests for functions that already exist.
02
Remembering syntax
The old advantage of remembering an API, looking things up quickly or being the person who had read the documentation is now close to free for everyone.
03
Some basic client jobs
One-page websites, basic contact forms and simple scripts are increasingly handled with templates and an AI chat window.
You used to do this basic work while you learned, and the company received value from you immediately. Now, a skilled developer can complete much of it in minutes.
Some small jobs did not move to cheaper developers. They disappeared because clients could handle them themselves.
If you were counting on that work to be your way in, it is a genuine loss. Grieve it briefly, then keep reading.
What became more valuable
The value moved beyond typing the code.
Generating code became easier. Deciding what should be built, judging whether it works and taking responsibility for the result did not.
Judgement
Knowing what to build
Clients often cannot explain what they need. They say “the website is bad”, “make it like theirs” or ask for the wrong solution entirely.
Turning those requests into a workable plan requires questions, judgement and sometimes the courage to say an idea is too expensive.
Responsibility
Being accountable
If a payment system breaks on a Saturday and the business is losing sales, somebody must take responsibility for diagnosing and fixing it.
A large part of what clients pay for is knowing that a capable person will answer and own the outcome.
Understanding
Evaluating the output
Anyone can generate code. Far fewer people can tell whether it is correct, secure, maintainable or genuinely solves the intended problem.
Your fundamentals still matter. They have been repriced, not written off.
The part tools struggle to finish
The last ten per cent of a real project.
Making software work inside an actual business, alongside its existing systems, data and staff, is rarely just a coding problem.
The generated feature still has to fit the business’s process. Existing data must be handled safely. Staff may resist changing how they work. Somebody must test the awkward cases, decide what happens when things fail and support the result after launch.
That final stretch is where technical implementation meets judgement, communication and accountability.
The opportunity underneath the loss
AI also changed which small jobs can be profitable.
AI removed some low-value work. It also removed much of the time and cost that made smaller projects unattractive to established agencies.
Small jobs were rarely unprofitable because they were technically difficult. They were unprofitable because an agency’s overheads do not shrink for a £600 project.
Meetings, project management, sales and account handling cost about the same whether the job is worth six hundred pounds or sixty thousand. Agencies therefore set minimum prices and ignore the smaller work.
You do not have that overhead
With AI-assisted production, one capable individual can sometimes complete in two days what previously required a small team for two weeks.
The ignored market becomes viable
Work that is too small for an agency can still be commercially worthwhile for an individual with low overheads and a repeatable process.
Reliability becomes the advantage
This market needs somebody who understands the problem, answers the phone and turns up when they say they will.
The honest summary
AI didn’t take the work. It took the tasks and left the outcomes.
Outcomes are what people pay for.
The task market is closing. The outcome market remains open and is short of suppliers who can understand a problem, deliver the result and take responsibility for what happens next.
The rest of this book will show you how to enter that outcome market.
Continue reading
Next: Evidence beats credentials
Chapter 3 explains why one small paid outcome can provide stronger professional evidence than another certificate or tutorial project.