AI Will Not Fire You. It Will Just Stop Promoting You.

Sometime around 2010, the forums I used to follow started filling up with the same conversation. Steve Jobs had written something about Flash, and everyone had an opinion. I never actually read the letter. I didn’t need to, the noise was everywhere. Flash community sites, developer forums, comment threads that ran for pages. People were angry, then dismissive, then logical about it.

The arguments made sense at the time. HTML5 was young and couldn’t do half of what Flash did. The Flash developer community was massive, battle-tested, and deeply embedded in how the web worked. And the idea that one person could just decide that a technology used across most of the internet was finished? That felt like arrogance more than prophecy.

But looking back, the arguments were defending the present. Nobody was really arguing about where things were going. I kept working ActionScript jobs through all of it, but the options were visibly shrinking. By 2013 I started picking up HTML skills on the side, and in 2014 I took a salary dip just to get into an HTML5 role. It wasn’t a panic move. But I won’t pretend the timing was perfectly calculated either. The floor was thinning and I could feel it.

Every shift cost something

I started my career as a graphic designer, which was not what I had planned. I had studied animation and multimedia, imagined myself doing something more technical, but the jobs I wanted did not materialise after college. I will blame the 2008 recession for some of that. So I took what I could get, and within six months I knew the work was not right for me.

I went to my manager and asked to move into Flash scripting within the same company, at a smaller salary. It was not a hard decision intellectually because I had always been comfortable with programming, something I was good at in school before I chose animation. But it was still a dip I had to choose with no guarantee it would work out.

It worked out. ActionScript felt like the right kind of work and I got reasonably good at it. Then the opportunities shrinking, I made the shift to HTML5, and after that came JavaScript frameworks, React, Next.js, fullstack development, and now AI is the thing everyone is watching. From the outside it probably reads like a clean progression. It did not feel that way from inside it. Each move meant starting at the bottom of a new curve, with less credibility than I had just built in the thing I was leaving behind.

What I did not fully appreciate early on was how much easier it is to make those moves when the stakes are low. As your career grows, so does everything around it. The work gets more complex, responsibilities outside work pile up, and the mental load of both leaves you with genuinely less capacity to absorb new things. Learning still happens, but it does not come as freely or as fast.

The people who stayed

The colleagues I watched get left behind during the Flash to HTML5 shift knew that feeling well. These were not junior or careless people. Some of them had taught me things in my early days, senior and respected engineers who were genuinely good at their work. When the shift became impossible to ignore, they were given opportunities to transition. The door was not closed on them.

But transitioning meant learning something new while still delivering on current work, at a stage in their careers where time, energy and mental load were already stretched. Some tried and could not sustain it. Others did not try at all. Slowly, without any single moment you could point to, they became the people who maintained the old systems while all the new work moved elsewhere. Still employed, still useful, just no longer at the center of things the way they once were.

I felt lucky more than anything. Lucky that I had made the move when I did, even if it meant taking a dip to get there. I tried to help them where I could, sharing what I was learning and what I was seeing in the industry. It did not make much difference, and I do not think that was about awareness. They could see what was happening.

For some of them it went deeper than skill. They were known for that stack. Their reputation inside the team, the complex projects they owned, the depth they had built over years, all of it was tied to it. Walking away did not just mean learning something new. It meant giving up the thing people associated them with before they had anything else to replace it with.

The cost of moving just felt too high at that point in their lives.ima

The pattern does not wait for you to be ready

I see the same thing playing out with people I know who are deep in a legacy technology that has been quietly losing ground for years. The signals are familiar. Work is shrinking, new projects are not coming in on that stack, and the people who built deep expertise in it are slowly becoming the ones who keep old things running. The arguments are familiar too. The community is large, the technology still works, there is plenty of existing code that needs maintaining. All of that is true. It was true about Flash as well.

But the more important shift is happening around AI, and this one is different in a way that makes it harder to respond to. Flash took years to fade and gave people enough time to see what was coming, even if some still did not move. AI is not moving at that pace. The gap between early adopters and everyone else is widening faster than any technology shift I have seen in my career, and there is no single dramatic moment to react to.

That is what makes it more dangerous for mid-senior engineers specifically. Nobody is losing their job tomorrow because they have not adopted AI. The work is still there, the expertise still valued, the paycheque still arriving. But the ceiling is quietly coming down. The first sign is usually a promotion that does not come, or an interesting project that goes to someone else. The interesting work, the bigger roles, the meaningful growth are increasingly going to people who have figured out how to bring their experience and AI fluency together. The window where existing expertise subsidises that transition, where years of system design and domain knowledge give a head start on the new curve, does not stay open indefinitely.

The ceiling you do not see coming

When I think about the AI shift, I am not just watching from outside it. I am in it too. I have made every transition so far. Some by instinct, some by necessity, some by taking a dip I did not fully want to take. Each time it worked out, though not always immediately and not always cleanly. And each time I told myself the next shift would be easier because I had done this before.

It is not easier. The weight of it is more familiar but no lighter. I am somewhere in the middle with AI right now, learning, experimenting, figuring out where my experience is an asset and where it is just noise. I am not where I want to be with it yet.

What I do know is that every time I made a move, the hardest part was not the learning. It was accepting that what I had built had an expiry date, that the next move meant starting at the bottom of a new curve again, and then slowly finding ways to plug your experience in and make it count.

I am aware enough of the pattern to know that the feeling of having time is usually the first sign that you do not.