~/blog/tag/career
Career
What generation getting cheap does to the job itself. Who gets hired, what is still worth learning, and which parts of the work nobody automated.
What I write about here
The posts under this tag are about the job rather than the code. What happens to a career in software when the part that used to take the longest stopped taking long.
The thread running through them is that the change has not been evenly distributed. Typing got cheap. Deciding what to type did not, and neither did reading code you did not write, holding the context nobody wrote down, or being the person who can tell that something plausible is wrong. Those were always the senior half of the work. They are now most of it, which is fine for people who already have them and a real problem for anyone who was going to acquire them by doing the cheap half for a few years.
That is the junior question, and it gets several posts here because it is the one I find genuinely difficult. The pipeline that produced experienced engineers ran through work that is now automated. Nobody has proposed a replacement that survives contact with a budget.
The rest is practical. What is still worth learning first. Which tasks I still do by hand and why. How estimation changes when the typing was never the expensive part. Who is accountable when the code was generated and it goes wrong.
Read these expecting a position rather than a forecast. I do not know what this profession looks like in five years, and the posts that claim to are selling something.
best entry points
- Hiring junior developers in 2026: the collapse is a hiring decision
The employer side, argued from the payroll data rather than the headlines. The collapse runs through hiring that stopped, not people who were let go.
- Learning to program in the age of AI: what I would learn first
The other side of the same bargain. Four things I would still learn first, and the list I would happily skip now.
- The most resilient job is eating its seed corn
Why a profession that stops making juniors has a problem it will not feel for years, and cannot fix quickly when it does.
Software estimation with AI: the typing was never the estimate
Software estimation with AI keeps missing. One METR survey asked 349 technical workers how much faster they had become and how much more value they produced, and got 3x and 1.4 to 2x. The gap between those two answers is the estimate you are actually making.
read →Hiring junior developers in 2026: the collapse is a hiring decision
Junior developer hiring stopped, and the number everyone quotes to explain it is not in the paper it gets credited to. What the research found makes the collapse a decision, and one large employer is now betting the other way.
When not to use AI for coding: the tasks I still do by hand
Knowing when not to use AI for coding is already a skill you practise, unconsciously and badly. It is also what wrecked METR's follow-up study. Here is the dividing line I actually apply, and why difficulty is the wrong axis.
Who is responsible for AI-generated code? You are, and 2026 wrote it down
Accountability for AI-generated code stopped being a philosophical question this year. In January California barred "the AI did it" as a defence, Europe makes software a product in December, and no vendor indemnity covers your code being wrong.
Learning to program in the age of AI: what I would learn first
Every junior skills list for 2026 says the same thing: learn the agent. That advice is worth about two weeks. Here are the four things I would actually learn first if I started developing today, and why the market is short on exactly those.
How to get better at reading code: a practice routine
Everyone diagnoses AI skill atrophy. Nobody trains the fix. Reading code is a muscle you can drill: a concrete practice routine with a merged PR a day, predict-before-you-run, and explaining code back without the AI.
The most resilient job is eating its seed corn
New data says AI did not kill engineering jobs, engineers are the most resilient function in tech. Read the footnote: the same report shows the junior on-ramp has collapsed. Resilience now is borrowed against a senior shortage later.