You did the hard part. You can build the thing. You ship features, you read a stack trace without panicking, you can pick up a new framework over a weekend. That competence is real and you earned it.
So here is the uncomfortable question: what happens when coding itself starts losing value? Not disappearing, losing value. AI now writes the boilerplate, the tests, the first draft of almost anything. The thing you trained for, the part that used to make you stand out, is becoming the baseline everyone has. The differences that decide where you end up sit somewhere else entirely, and nobody really sat you down and taught you where.
Is software engineering dead?
No. Software engineering is not dead, but the version of the job where writing code was enough is ending. Companies still hire engineers. Most of the open positions are for seniors, though, and the entry-level jobs are the ones that are shrinking.
Three numbers behind that:
- 71% of all open developer positions are for seniors (Indeed Hiring Lab, via Uvik).
- Entry-level software engineering employment fell close to 20% between late 2022 and mid-2025, while employment for experienced engineers held steady or grew (Stanford Digital Economy Lab).
- Writing working code is no longer scarce. On SWE-bench Verified, AI performance rose from about 60% to near the human baseline within a single year (Stanford AI Index).
We go through these numbers in detail in what the Stanford AI Index means for junior engineers. The short reading: the demand is still there, but it has moved toward experienced engineers.
Will AI replace software engineers?
AI will replace part of the work, not the profession. The part it takes first is routine coding: boilerplate, tests, small well-scoped tickets. Engineers whose whole contribution sits there are the most exposed.
The clearest sign of what stays is how the companies that still hire juniors describe the job. IBM is tripling its US entry-level hiring, but the new roles replace routine coding with judgment, oversight of AI output, customer interaction and business acumen.
That does not mean nobody loses a job. About 30% of organizations expect to reduce their workforce within a year because of AI, according to McKinsey's 2025 State of AI survey, and software and data engineering are among the most affected functions.
So the job changes more than it disappears. What the Stanford AI Index means for your role goes through that change in three concrete parts of the daily work, and the software engineer career ladder shows which skills each level adds on top of code.
The three skills that now decide your career
My co-founder put together a short on this when the book came out.
Here is the longer version, with the reasoning behind each point.
Addy Osmani spent more than a decade at Google before joining Anthropic, and he put this into a book called The Effective Software Engineer. Three patterns run through it, and the shape underneath them is the whole point.
Efficient is not the same as effective
Doing something well is not the same as doing the right thing, and mixing them up is expensive. You can spend three months building a technically flawless feature that nobody uses. Clean code, full test coverage, zero business value.
The skill that protects you here is not faster typing. It is judgment: looking at the backlog and asking which of these actually matters, instead of which one is the most fun to build. Most engineers optimize for the interesting problem. The effective ones optimize for the problem that moves something.
Your code only travels as far as you can explain it
Finding the right solution is half the battle. The other half is convincing the organization to actually build it. Your ability to communicate does not sit next to your technical skill as a nice extra. It sets the ceiling on how far that technical skill reaches.
A great design you cannot get anyone to agree to is a great design that never gets built. Communication is not the soft thing around the real work. It decides whether the real work happens at all.
Invisible work does not get counted
You might assume people notice what you are shipping. In a busy company, they mostly do not. Promotion conversations happen in rooms you are not in, and the only version of your work that exists in those rooms is the version someone can describe out loud.
This is not about self-promotion. If your work never enters the conversation, it does not exist for the people deciding who moves up. Making your impact legible is a skill nobody teaches you on the way in.
So here is the pattern
Three takeaways, one shape underneath all of them. Choosing what matters, explaining it, making it visible: these are exactly the skills most engineers treat as optional, and exactly the ones that decide where a career goes. They were never on the syllabus, so it is easy to assume they do not apply to you. As AI takes over more of the pure technical work, they are quietly becoming the part that is actually yours.
This is not about your code getting worse. It is about the ground shifting under what “being good at this job” means, and noticing it before the people around you do.
FAQ
These are the skills we train at Utterskills: communication, ownership, and judgment. The Senior Shortcut teaches them on purpose, so you don't have to pick them up by accident.