The AI Development Revolution: Part 1 - The Awakening

Part 1 of a series about what it felt like to build software with AI while it was changing under me. Not the theory. The actual days.

When everything changed

I wrote this in my journal:

Working with a partner all day, whether human or not, is tiring. You really keep on going, and the pace is fast. Practically coding at the speed of thought, or somewhat slower. An aching back and body are the least of my concerns.

It is truly hard to balance the day's work, which has been much more productive in some senses, but also a bit jarring in terms of my overall flow and day. It really feels like everything has changed. And you can get sucked into code projects for days.

I did not set out to document anything. I was just trying to keep up. But reading it back, I can see I was watching the way I build software change while I was inside it, project by project, one late night after another.

The physical part

The thing I did not expect was how physical it got. When you can move at the speed you can think, your body is the slow part. The mind runs ahead and the back and the eyes are left trying to catch it.

The pull is real, though. From the same entry:

What it's like, working with the machine. The back and forth, the commits, the branches. The waiting and watching and learning. The AI is incredibly creative as a software engineer. Much in the same way we would put together tools, reduce problems, try different angles and approaches. And the surprising additional thing it picks up or decides to check.

It looks like typing prompts. It is not. You are holding several systems in your head at once, checking the work, watching for the moment it suggests something you would not have reached for. The late nights were not only excitement. They were the cost of running at that pace for hours.

As I wrote in "Vibe Coding with AI": "You can get sucked into code projects for days."

It was not the code that surprised me

I expected it to write code. That part was the point. What I did not expect was the sense that it was thinking like an engineer. Trying angles. Catching connections I had walked past.

That is not how automation usually feels. Automation does a task you handed it. This felt more like sitting next to someone. "The AI is incredibly creative as a software engineer," I wrote. "Much in the same way we would put together tools, reduce problems, try different angles and approaches."

Four projects at once

The same shift showed up across four projects, each one a different shape of it.

Enterprise infrastructure: AWS systems that should have taken months. I watched it move through CDK stacks with something like intuition, holding the dependencies between services while it offered architectures I had not thought to try.

Content automation: years of handwritten journals, turned into clean text without flattening the voice out of them. It understood more than the OCR. It understood the editing.

Business applications: retro design over modern guts, Flutter web pushed harder than it wants to go, visual decisions made fast.

Game development: this one taught me the most. Building a physics engine, structuring the architecture around the fact that a machine would be working in it too.

Each one took real attention. It was not prompting. It was leading the work while the machine moved fast under it.

The numbers

The numbers were hard to argue with. Hundreds of commits. Tens of thousands of lines. Timelines that used to mean something stopped meaning anything. But the numbers are not the part I remember.

What it actually felt like

From my journal:

1 week into the migration and almost there. 1 week. To build. To move to reimplement. To plan. To vision. Nuts. So fast. Wow.

The disorientation was the honest reaction. Infrastructure that should have been a quarter of work was done in days. Things that normally need a team got built alone, at night. I did not have a frame for it.

As I wrote in "Living Through the AI Revolution": "I tingle at times, knowing I'm part of something so historic, so grand...I've never seen something move and change so fast."

The learning curve

Working this way is not like picking up a new language. The skill is not the prompt. It is holding the whole thing in your head, the mental model across systems, while the machine suggests connections, and checking quality faster than feels comfortable, and steering the architecture in real time.

The late nights came from that. "Just one more feature" at 3 AM, when it is flowing and you cannot make yourself stop. You only feel the cost the next day.

And the questions came with it. Am I still a developer if the AI writes the code? Where does my idea end and its suggestion begin? I wrote about that stretch in "The Multithreaded Mind": living at machine speed and trying to keep my own judgment intact.

The honest part

None of this was clean. CDK stacks I had to tear down and redesign when the AI built circular dependencies into them. Long nights chasing permissions it had set wrong. Whole weekends lost to networking config that held until it didn't.

Five different OCR approaches before one of them kept the handwriting intact. Pipeline rewrites when a suggestion worked fine and then would not scale. "Am I still engineering if the AI does the implementation?" That one followed me through the debugging.

But something was happening that I did not have words for yet, and it was changing more than my workflow. It was changing how I think about working alongside a machine at all.

As I wrote in "Layers of Abstraction": "This is the age of grand abstraction. Of thought, of systems, of beautiful design, grand unified patterns."

The parts that follow are where I tried to make sense of it. The methodology, the infrastructure, the content pipeline, and where I think it goes.

This series is based on real development work, documented through daily journal entries and git histories.

Next: Part 2: The Methodology →

Series Navigation: