I Have Survived 4 Tech Extinctions Since 2002 (And AI Is Number 4) 🦕

The Morning My Job Felt Small

This morning I woke up, poured coffee, and read a summary of the latest AI research.

I did not write that summary. An agent did. It ran while I was asleep, read more papers than I could read in a week, and left me a clean list of what mattered.

I sat there for a second with the cup in my hand.

Twenty four years ago I got paid to remember things. Now I get paid to read what a machine remembered for me.

That feeling has a name, and I think a lot of you have felt it lately. It sits somewhere between gratitude and grief. You are faster than you have ever been, and you feel smaller than you have ever been, and both of those things are true at the same time.

Here is the part that keeps me calm.

This is the fourth time.

I started writing cell phone software at Motorola in 2002. Since then I have watched my job get declared dead four separate times. Three of those times, the people announcing the death were right about the change and wrong about the people. I do not know yet if the fourth one is different. But I do know what the first three taught me, and nobody ever put that in a blog post, so I am putting it here.

Every extinction I survived looked exactly like the end of the world while I was standing inside it.

Extinction One: The Rule Was That You Had To Know It (2002)

When I started, the rule was simple and everyone said it out loud.

You had to know it. In your head. Not in a browser tab.

Looking up someone else's code online was treated as a character flaw. If you were caught searching the web for an answer, the reaction was not "nice, you found it fast." The reaction was "why do you not know this?" Your value as an engineer was measured by what you could produce from memory, at a whiteboard, with no help.

Stack Overflow did not exist yet. It would not launch for another six years, in September 2008. There was no place to go even if you wanted to cheat.

So we memorized. We read the chip documentation until we could see it with our eyes closed. We kept printed reference pages on our desks with coffee rings on them. Senior engineers were respected for the amount of the system they could hold in their skull at one time.

And then the web filled up with answers, and that whole system of value quietly collapsed.

Nobody sent an email announcing it. One year, searching was shameful. A few years later, an engineer who refused to search was just slow.

I watched good people take that badly. They had spent a decade building a reputation on memory, and the market stopped paying for memory. Some of them adapted. Some of them got quietly bitter and told younger engineers that they were not real programmers.

I have thought about those bitter engineers a lot this year. Because now I understand them. They were not wrong that something real was lost. They were wrong about what to do next.

The first skill I ever got paid for was memory. The market stopped paying for it, and I did not even notice the funeral.

Extinction Two: The Phone That Ended Our Party (2007)

In January 2007, Apple showed the iPhone.

I want you to understand what the world looked like from inside a phone company right before that. The Motorola RAZR sold around 130 million units. It was the best selling clamshell phone ever made. When you work inside a product that big, the building feels permanent. The hallways feel permanent. The way you build software feels like the only correct way to build software, because look at the numbers.

Then one product arrived and the ground moved.

You already know the ending. Motorola split into two companies on January 4, 2011. Google bought Motorola Mobility in May 2012 for 12.5 billion dollars, then sold it to Lenovo in October 2014 for about 2.9 billion.

I lived through the middle part, and the middle part is the part nobody writes about. The middle part is not dramatic. It is meetings. It is a roadmap that everyone in the room knows is wrong, and nobody says so, because saying so is not how you keep your job. It is watching smart people defend a decision they stopped believing in six months ago.

That is where I learned the thing I still believe most strongly.

Change is frightening. Change produces failure. And failure is the best data you will ever get.

Not the good kind of data that goes in a slide deck. The expensive kind. The kind that costs somebody their quarter and somebody else their team. That data tells you the truth about how a thing actually works, and you cannot buy it, and you cannot get it from a model, and you only get it by standing inside the failure and paying attention while it happens.

I got a full education in one industry collapsing. It cost me years and it was worth every one of them, because I now recognize the shape of it early. When I walk into a room today and feel that specific quiet, I know what it is. I have felt it before.

Failure is expensive data, and the receipt is your own career.

Extinction Three: The Quiet One Nobody Announced (2010s)

The third one had no launch event.

Mobile stopped being the frontier. Data and machine learning became the place where the interesting problems lived. There was no keynote, no single product, no clean before and after. It just slowly became true.

And I noticed a pattern in that one that I have seen repeat every single time since.

The people who understood the change were scared, and because they were scared, they moved. They learned the new thing badly, at night, with no support, feeling stupid the whole time. Fear made them move.

The people who did not understand the change were not scared at all. They were not interested. They had opinions about it, but no curiosity. They waited.

And then, later, when it was obvious to everyone, those same people were the loudest voices saying the whole thing had been a mistake all along. Not because they had evaluated it. Because they had missed it, and calling it a mistake is easier than saying you were not paying attention.

This still happens. It is happening right now, around you, about AI. I watch it in comment sections and in meetings and in my own team chat.

Here is the uncomfortable version of that observation, and it includes me.

Being scared is a sign that you understand the situation. If you feel nothing about AI right now, that is not strength. Calm can mean you have a plan. It can also mean you have not looked.

The scared ones move. The comfortable ones wait, then call it a mistake once it is too late to join.

Extinction Four: The One Where I Wake Up Behind

Which brings me back to my coffee.

Today I write AI and ML software. Hands on, in the code, most days. And the honest report from the inside is this.

You have to study every day. Every single day. There is no version of this where you learn it once and coast for three years.

That is not a motivational line. It is a complaint, and I mean it as one. It is tiring. My morning now starts with an agent's summary of what changed while I slept, because that is the only way I have found to not fall behind. The field moves faster than one human reading speed. So I use a machine to keep up with the machines.

Some mornings that feels like a superpower.

Some mornings it feels like being on a treadmill that somebody else controls the speed of.

Both feelings are correct, and I have stopped trying to pick one. I think demanding that you feel only one thing about AI is part of why this is so exhausting for people. You are allowed to open the agent output and think "this is incredible" and "I miss when I was the one who knew this" in the same breath.

What makes this extinction feel different from the other three is not the speed. It is the target.

The first three took my tools. This one aims at the part I thought was me.

I use a machine every morning to keep up with the machines. That is the actual job now.

The Grief Is Real And You Are Allowed To Feel It

I read something recently that put words on this better than I could.

A programmer named Andy Balaam published a piece about the grief he feels about AI, and what struck me was that it was not about job loss. It was about being told that the thing you love, the thing you spent your life getting good at, is now considered a commodity. He points out that his craft was also his identity and his proof that he was worth something, and watching the industry shrug at that is its own kind of loss. He also makes a point I did not expect, which is that plenty of other work has been treated this way for a long time. Farmers. Nurses. Drivers. Teachers. Cleaners. Caregivers. The shock some of us feel at being replaceable is partly the shock of finally being on that list.

That landed hard.

The discussion that followed it on Hacker News was worth reading too, mostly because people disagreed with each other so honestly. One commenter said the part that hurts is not today, it is the direction of travel, because if progress stopped right here the fear would stop with it. Another said they had tied their self worth to writing moderately clever code, and having that challenged was uncomfortable but maybe overdue. Somebody else described the version I dislike most, where the work turns into reviewing plans and output that a machine produced, and the thinking part that they actually loved gets handed off.

I have had that last day. The day where I approved more than I built. I did not enjoy it.

So no, I am not going to tell you to cheer up. Three extinctions did not teach me to skip the grief. They taught me that skipping the grief is exactly what turns a person into that bitter senior engineer from 2006, the one who never stopped telling younger people that they were not real programmers.

You have to actually feel it. Then you get to decide what to do.

Grief that you refuse to feel does not disappear. It turns into contempt for whoever comes next.

What Actually Survived All Four Times

Here is the contrarian part, and it took me four rounds to see it.

None of my technical skills survived. Not one.

The memory work from 2002 is gone. The embedded phone work is gone. Half of what I learned in the data years has been replaced by tools that do it better. If you measure my career by the specific things I knew how to do, I have been wiped out four times and I am sitting here at zero again.

But something did carry over, and it carried over every time.

I know what a failing project feels like about two months before the schedule admits it. I know which meeting is the real meeting. I know the difference between a demo that works and a system that works, and I learned that difference by shipping software to millions of phones where you cannot push a fix on Tuesday.

I know how to be wrong in public and keep going. That one came from extinction two and it cost the most.

I know how to start over while feeling stupid, because I have done it three times, and I now recognize feeling stupid as the first part of learning rather than proof that I am finished.

None of that is code. All of it came from doing code.

That is the strange part. You cannot get the judgment without doing the work. There is no shortcut where you skip twenty years of building things and arrive at the taste. And an agent can now write the code faster than I can, but the agent did not sit in the room in 2008 while a business that felt permanent came apart.

That is what I actually sell now. It just took me four extinctions to notice I was selling it the whole time.

I have lost every technical skill I ever had. I kept everything those skills taught me.

The Part That Still Scares Me

I promised myself I would not end this with a comfortable answer, so here is what I cannot resolve.

The first three extinctions replaced the tools and kept the human in the loop. This one is aimed at the loop.

If the trend continues at the current speed, I do not know how much judgment stays valuable, or for how long. Anyone who tells you they know the answer to that is guessing. The confident optimists are guessing and the confident doomers are guessing, and the loudest voices on both sides are the least useful people to listen to right now.

What I have is not a prediction. It is a position.

I keep my hands in the code, because judgment decays fast when you stop building. I study every day, because that is the entry fee now. I pay attention to failures, mine and other people's, because that is still the highest quality data available. And I try not to tie my whole sense of worth to whichever specific skill happens to be scarce this year, because I have watched that scarcity evaporate four times.

That last one is the hardest, and I am not good at it yet.

What I Would Tell You If We Were Having Coffee

If you are early in this and you are scared, you are reading the situation correctly. Being scared is not weakness here. It is comprehension. Use it to move.

If you have been doing this twenty years and you feel like the ground is gone, I believe you, and the grief is legitimate, and feel it properly so it does not turn into bitterness. The younger engineers are not your problem. They inherited this, same as you.

If you feel great about all of it, I am honestly a little envious, and I would still ask you to look one step further down the road than the step you are enjoying.

And if you are wondering whether it is worth learning to build things now, from scratch, in a world where a machine writes the code faster than you can type it, my answer is yes, and my reason is not sentimental.

You cannot get the judgment any other way.

I have been wiped out four times since 2002. Every time, the thing that saved me was built during the years that got wiped out.

That is a weird way to build a career. It is the only one I know that works.

If this hit something for you, please clap 👏 and leave a comment. Hearing from readers genuinely keeps me going, and it helps other people find this piece too.

Which extinction are you living through right now? I read every comment.

☕ If you want to support my writing: buymeacoffee.com/zirubakr

Thanks for reading, Suraj…

Enjoyed the read? Your support means a lot and helps others find this piece, too.