This isn’t the article I expected to write.

I sat down to write a different version of it: the one where the funnel is broken and brilliant senior engineers like me are being filtered out by keyword-matchers and twenty-five-year-olds who’ve never used Objective-C. Most of that piece is true. None of it is the point.

The point is harder, and I want to say it plainly because I think a lot of my generation needs to hear it: the extra ten years past the five-year mark don’t make me ten years more valuable. They don’t even make me one year more valuable. Most of what I learned in those ten years has been compressed into the AI that any five-year engineer can now use.

Let me run the thought experiment honestly.

If I applied for an iOS role tomorrow, I’d fall out of the funnel at stage one. A keyword filter would notice that my recent years have been infrastructure and engineering leadership, not SwiftUI commits. It would move on. I’d never reach a human.

Is this unfair? I shipped Apple-featured apps. I built frameworks. I traced codebases like an archaeologist at a dig. I watched Objective-C give way to Swift, programmatic UI to UIKit to SwiftUI, CALayer to Metal, and shipped through every transition. Surely that has to count for something?

Here’s the uncomfortable answer: it counted for a great deal in 1995, in 2005, and in 2015. In 2026 it counts for almost nothing that an AI doesn’t now provide for free.

Let me go through the things people might expect to be “my edge” as a senior engineer:

Judgement on what to build. This is real. It’s also no longer scarce. I watched as other senior engineers embarrassingly told senior designers to go and start again, clearly they didn’t know “our field” well enough and tell management to “leave the coding to us”. Now, my CEO is building his own features — not because he learned to code in the last six months, but because he doesn’t need to. He understands what he wants, and AI handles the rest. The judgement layer senior engineers were protecting with their “knowledge and experience” turns out to be the layer non-engineers were already standing on. They just weren’t free to use the keyboard before.

Arguing with AI when it’s confidently wrong. I’m good at this. So is anyone with reasonable taste and a willingness to say “that’s not right, try again.” It takes a week to learn. Maybe a month. Not an extra ten years. One of my PM team wanted a drag and drop feature, he said it took a few hours to get it right, but he and Claude got there in the end.

Reading a room. Spotting what isn’t being said. Genuinely a skill I have. Genuinely not a programmer’s skill. A good PM, a good QA lead, a good designer all have it. It’s why I got promoted to mobile lead for a large corporate — and it’s also why, if I’m honest, I was being paid for something other than engineering even back then. The skill doesn’t justify a senior engineer’s salary today, and looking back, on its own it wouldn’t justify one then either.

Pattern recognition across codebases — knowing which mistakes I’ve already made. This sounds impressive. In practice it amounts to things like use a string constant instead of a literal here or this nesting will hurt you in six months. These are real, but they’re trivial. Any five-year engineer can learn them in an afternoon. Most of them are now flagged by the AI before the human even sees the code. The arguments my teams had over whether to use MVC, MVVM, MVP or VIPER proves the point that it doesn’t matter. With this many competing right answers, there was never a right answer. My CEO hasn’t heard of design patterns, doesn’t care and doesn’t need to. His code works, we will clean it up in the same way that he wrote it. None of us will be asking which design pattern Claude chose. Nobody will be auditing the code for pattern adherence. The question won’t come up. Our task will be to prepare his code for scaling, efficiency and security — things that the CEO doesn’t know how to ask for or how to verify.

So what’s left? Honestly: not much that justifies the premium the market used to pay for my CV or the reasons I was headhunted.

Looking back, I wonder how Claude would cope with the problem that introduced me to my CEO in the first place — a complex video player issue that three outsource teams had already failed to fix. They’d burned countless days between them and all admitted defeat one after the other. The POC was dead in the water without it. I solved it in three days. That’s the kind of problem my fifteen years used to be for. The honest questions now are: would Claude solve it in less and would my CEO have needed me at all?

The value has been redistributed. The PM who understands the user — they have it now. The QA who knows the system in production — they have it. The five-year engineer with good taste and basic social skills and who knows how to argue with Claude — they have most of what I have, at a fraction of the cost. My extra ten years didn’t disappear; they got unbundled, and the pieces went to people who were already doing parts of the job better than I was.

This is happening to thousands of engineers right now, and I want to be honest about how it feels from inside it.

For me — not much. I’ve seen this before. I was an assembler engineer. I was a Borland C++ Builder engineer. I was an Objective-C engineer. Each of those “was”, at the time, a real specialism with real scarcity value. Each of them got abstracted away, and I shrugged and learned the next layer. This is the same process, just faster. The shrug is the same shrug.

For my ex-colleagues, though — the senior engineers I came up with, the ones who specialised hard and stayed specialised — it’s much worse. They’re watching their CV become a list of skills the market no longer prizes. They’re seeing junior engineers do in a day what used to take them months. They’re being told, in coded language by recruiters who used to court them, that they’re “too senior” or “not the right fit” or “not what we’re looking for at this level.” Some of them are fighting back with arguments about the desperate need to keep them around, to oversee, to read every line of code like they always have — they are arguing in an ever-emptying room.

What they’re hearing is: the years you accumulated aren’t worth what you think.

That’s not a failure of the hiring system. That’s the hiring system finally telling the truth. The premium for senior years was always partly a bet that those years bought judgement that the market couldn’t get cheaper (and the hope that they were nice people) anywhere else. AI just made most of that judgement cheaper somewhere else.

The most painful version of this transition is the one where you keep insisting the market is wrong about you. It isn’t. The market has moved on. What we did used to be worth what we charged. It mostly isn’t anymore. Companies are paying far less for AI than an entire team of engineers, so why should they continue to pay us at our old rate? They aren’t a charity, it’s business. We took the high rates while we could, we’re just lucky it lasted as long as it did.

The path forward isn’t to argue with the funnel. It’s to find the part of the work that genuinely hasn’t been compressed — and most of us, if we’re honest, will discover that part is much smaller than we thought, and lives in places we didn’t train for. Closer to the customer. Closer to the business decision. Closer to the messy ambiguous work of figuring out what should exist, rather than how to build it.

That work isn’t called “senior engineer” anymore. Some of us will make the transition. Some of us will just retire and maybe train as a masseuse or yoga instructor (as I know some have already).

I’m telling you this because someone needs to, and because I’d rather hear it from someone who’s been through so many of these shifts already than from a recruiter who’s just stopped returning your calls.

The extra years aren’t worth what they used to be. Mine aren’t either. The sooner we accept that, the sooner we can work out what we are worth — which might still be a lot, but probably not for the reasons we’ve been telling ourselves.