From speed to momentum: how AI reshaped my development workflow in 2025.
Over the past year, I've completely rebuilt how I work. Between AI coding partners like Claude Code, better tooling such as Linear, and a bit of structure (with plenty of help from ChatGPT), I've gone from half-finished side projects to shipping working products that actually make sense.
This post is a look at what my workflow looks like in 2025, a peek into how these tools fit together, and what I've learned about balancing speed and structure (while maintaining my sanity) along the way.
Slow Start
Over the past two years, I've seen how new advances in AI have unlocked the ability for hobby developers to build and develop tools and solutions at a speed previously unseen. This movement has resulted in a Cambrian explosion, a sudden burst of creativity and experimentation where it feels like everyone is building something new every week. As someone who started their career in engineering and then took a deviation into the world of business and entrepreneurship, I saw tremendous potential in these tools to bring my (admittedly half-baked) ideas to life.
In my journey through entrepreneurship, I always had a list of business ideas or projects of various scales and viability, bouncing around in my head or hastily jotted down across various notes apps. There was aways something blocking their pursuit, capital, time, scalability, you name it. Most often though it was development skills, which I didn't really have to the extent I needed to get things really moving. While I could speak engineering, when staring at pages of code I'd lose focus quickly and didn't have much patience for the minutiae needed to deliver something that scales and works.
However, since these tools allow me to abstract away some of that complexity, they help reduce the barrier to entry and enable me to build and experiment faster, with lower risks.
At first, I was chasing speed and trying to build faster — but I realised that this wasn't getting me very far since I was directionless. What I really needed wasn't just speed — it was momentum.
First Steps
Parenthood hit pause on many things: sleep, spontaneity, but it also opened space for a different kind of creation. After my first child was born, I took a sabbatical — a time that slowed me down in all the right ways and reshaped how I thought about time and focus. Between feedings, naps, and the general chaos, I carved out small pockets of time to tinker and learn. It wasn't glamorous, but it was the most fun I'd had building in years — and enough to rediscover how much I love creating things from scratch.
I initially tried the cohort of "vibe coding" apps; a term used for AI-assisted IDEs like Cursor, Lovable, and Replit that let you code by describing intent rather than syntax. It seemed miraculous to be able to make your ideas come to life, and playing with them really helped to stratch the development and building itch I had been feeling. And most of all, I was able to build quickly, and I found the speed exhilarating.
However, while I found these tools easy to use, they didn't really give me confidence that I could build something stable. Things would look correct, but as I kept working on the products, I'd run into issues I couldn't fix. The black box that my projects were being built in didn't really allow me to see what was happening or even understand why things were being done in certain ways. My code would eventually devolve into spaghetti and become unmanageable, which really put me off the idea that "vibe-coding" was the answer I was looking for.
Gaining Traction
Those first attempts were messy, but they helped me understand what I actually needed from an AI coding partner.
This all changed with the launch of Claude Code. All of a sudden, the AI was in my codebase, and I could instruct it with a much larger context window directly in my CLI. This was much closer to what I was looking for, and I jumped on it to see how far I could take my projects.
I loved it. It was concise, capable, and able to solve just about everything I threw at it. The code it produced was modular and manageable, to the extent that even I could understand it.
More importantly, it gave me direction. For the first time, my projects felt like connected steps toward a tangible product, and not just random experiments. The gap between my ideas and execution felt smaller, like I finally had some velocity.
Together with Claude Code, I built a few hobby projects and deployed real experiments. But with greater speed came a new problem: controlling the trajectory.
Overclocked
Once I found a setup that worked, I leaned in, but it turned out I might have leaned on it a bit too hard.
It was now incredibly easy to build new features. Whatever I could think of, I built. If I thought it might add value, I tried it which eventually pushed my codebases into longer and more convoluted solutions.
Then it would break.
And I was right back where I started. I realized I now had more power, but my scope was exploding since I didn't have a clear plan. My mindset was "can it be done?" not "should it be done?" And most importantly I didn't have a definition of what "success" or "done" meant.
Bringing Order to the Chaos
I needed to slow down, add some structure, and start defining what "done" actually meant.
Now that I had a better idea of what I could build, I went back to my idea backlog and started working through it with Claude Code to scope things out. I focused on answering key questions such as:-
- Could they be built?
- Did they make sense?
- Would they take too much effort to maintain and scale?
Initially, I turned to Claude for this, but found that its answers weren't as inventive as I'd like for brainstorming. On advice from a friend and colleague, I tried out ChatGPT 4.5 and found it to be a better thought partner for creative and strategic exercises. Together with the newer models from OpenAI, I could evaluate ideas, break them down into manageable pieces, and define clear scopes of work.
Laying Down Tracks
Once I knew how to evaluate ideas, the next challenge was managing execution.
I turned to Linear as my project management tool. With it, I scoped projects into epics and issues, set milestones, priorities, estimates, and dependencies, and tracked development velocity. This helped me understand complexity and how each piece impacted the larger system.
I started running short sprints to timebox effort and measure output. Claude Code's default five-hour context timeout windows actually worked as perfect sprint bursts — long enough to make progress, short enough to stay sharp.
To manage feature development, I moved to Git- and TDD-based development. I instructed Claude Code or OpenAI Codex to write tests before the code. Each change lived in a new branch, and I submitted work as pull requests for review in GitHub.
This prevented errors from spreading through the codebase and kept context clean, where each issue or branch started fresh, free from old noise.
Once I started layering in structure, I noticed something new: I could now go after larger projects without losing control. Structure gave my work more mass — weight and stability — which meant I could tackle bigger systems or juggle multiple projects concurrently without things collapsing under their own speed.
That's when momentum truly took off.**.
What I've Learned About Working With AI as a coding partner
Working with AI as a coding partner has been incredibly enlightening. I've learned that:
- Scope management matters more than ever. As the barriers to building come down, it's even more important to stay focused on the end goal.
- Context is a currency. Keeping sessions small and focused makes AI output exponentially better.
- Burnout hides in productivity. Just because I can build faster doesn't mean I should keep going nonstop.
- The most valuable thing AI gives me isn't speed; it's momentum.
Momentum is mass times velocity, which is speed with direction. And that direction is what turns raw effort into real progress.
Wrapping Up
So this is where I am now, juggling between bursts of writing code, reviewing pull requests, scoping future work, and quality time with my kids sandwiched in between. And it works for me.
What's next?
Playing around more with Claude Code and Codex on the web, exploring Claude Skills, and continuing to build and ship, this time with a clearer sense of direction and enough structure to keep the momentum going.
Tools in My 2025 Development Workflow
- AI Coding Partner:
Claude Code| Core development assistant for building and debugging - Brainstorming & Ideation:
ChatGPT| Scopes, refines, and challenges project ideas - Model Foundation:
OpenAI Codex| Generates and tests code during TDD - Project Management:
Linear| Planning, sprinting, and tracking progress - Codebase Management:
GitHub| Version control and pull request review
Last updated: October 2025.