After 15 years of shipping code, leading teams, and watching paradigms come and go, I've learned one thing: the tools change, but the fundamentals don't. AI is the most powerful tool I've encountered, and like all powerful tools, it demands respect, intention, and clear boundaries.
The Moment Everything Changed
I remember the first time I pair-programmed with an AI. It felt like having an infinitely patient junior developer who'd read every Stack Overflow answer ever written. Impressive. Useful. And immediately, I sensed the trap.
The trap isn't that AI will replace us. It's that we'll slowly stop doing the hard cognitive work that makes us good at what we do. We'll outsource our thinking, one prompt at a time, until we wake up unable to architect a system without asking a machine first.
This manifesto is my line in the sand.
Core Principles
1. I Am the Architect, AI Is the Draftsman
AI can generate code faster than I ever could. But generating code was never the hard part. The hard part is knowing what to build and why. I use AI to explore possibilities and accelerate execution, but I own the vision, the trade-offs, and the consequences.
Every line that ships is mine. Every decision that gets made is mine. The credit and the blame stay with me. "The AI said so" is not a defense. Accountability cannot be delegated.
And yes, AI lets me move fast. But velocity without direction is just thrashing. I use AI to accelerate iteration, to get something tangible, then I slow down and make it right. Speed serves quality, never the other way around.
In practice:
- I define the intent before I open a chat
- I question AI's suggestions against my domain knowledge
- I never ship code I don't understand
- I review AI-generated code like it came from an enthusiastic but inexperienced developer
- I test assumptions, especially when AI seems confident
- First draft is fast; final draft is mine
2. My Voice Stays Mine
AI can write in my style. It can mimic my tone. But the moment I let it speak for me, in emails, in documentation, in decisions, I've surrendered something I can't get back: authenticity.
In practice:
- AI drafts, I edit and own
- Client communication stays human
- My writing reflects my thinking, not a model's average of everyone's thinking
3. Curiosity Is Non-Negotiable
The best engineers I've worked with share one trait: relentless curiosity. AI makes it dangerously easy to get answers without understanding. I refuse to let convenience kill my drive to learn deeply.
In practice:
- I use AI to go deeper, not to skip the depth
- When AI gives me an answer, I ask myself: "Could I have reasoned my way here?"
- I treat AI as a sparring partner, not an oracle
4. Sometimes I Close the Chat
Not every problem benefits from AI assistance. Some thinking needs to happen in silence. The moments I struggle alone, without a prompt to save me, are the moments I grow. I practice deliberate disconnection, not out of ideology, but to maintain skill. Like a pilot who practices manual flight, I need to know I can still do this work without assistance.
In practice:
- I build side projects without AI to test my raw skills
- When I'm stuck, I sit with the problem before reaching for a prompt
- I write code from scratch regularly to remember what it feels like
- I recognize when AI would rob me of valuable struggle
5. AI Should Feel Uncomfortable
If using AI feels too smooth, too easy, too frictionless, I'm probably over-relying on it. Some friction is signal, not bug. The discomfort I feel when AI suggests something I don't fully understand is valuable information. Comfort is the warning sign.
In practice:
- I pay attention to the feeling of "just accepting" an answer
- When AI makes something too easy, I ask: "What am I not learning?"
- I welcome the friction of verification and understanding
- Ease without comprehension is a red flag, not a win
6. Context Is King, and I Control It
AI has no persistent understanding of my codebase, my team's conventions, or my business constraints. Every interaction starts from zero unless I provide context. This is a feature, not a bug, it keeps me in control.
In practice:
- I make intent explicit before every session
- I don't rely on AI to "remember" what matters
- I treat context as a first-class artifact
7. Domain Knowledge Is My Unfair Advantage
AI knows syntax and patterns. I know why the business logic exists, what failed in production three years ago, and which stakeholder will push back on which trade-off. That context gap is where human judgment becomes irreplaceable.
In practice:
- I invest in understanding the problem space, not just the solution space
- I bring the "why" that AI cannot infer
- I stay close to users and systems, not just prompts
8. AI Should Lift Others, Never Gatekeep
The promise of AI is democratization, lowering barriers, enabling more people to build, create, and contribute. I refuse to use AI in ways that exclude, discourage, or create new hierarchies.
In practice:
- I share what I learn about effective AI collaboration
- I use AI to mentor and explain, not to intimidate
- I advocate for access and education
Working With Others in the AI Age
These principles aren't just personal. They affect how we work together as teams, how we review each other's code, and how we mentor the next generation.
When I review AI-generated code from teammates, I ask the same question I ask myself: "Do you understand what this does and why?" Not to gatekeep, but because that's where bugs hide and where learning stops.
When someone on the team uses AI differently than I do, more heavily or not at all, I stay curious about their approach. There's no one right way to collaborate with AI. The right level of assistance varies by experience, context, and task.
For junior engineers who've never coded without AI, I create space for them to struggle productively. Not because suffering builds character, but because the neural pathways you build when you solve a problem yourself are different from the ones you build when you prompt your way there. Both matter. Both have value.
I'm learning to mentor in an age where "look it up" now means something very different. The skill isn't memorization anymore. It's judgment. It's knowing when the answer AI gave you is correct, when it's almost correct, and when it's confidently wrong.
We're all figuring this out together.
A Note on the Future
We're at an inflection point. The engineers who thrive won't be the ones who resist AI, nor the ones who surrender to it. They'll be the ones who learn to collaborate with it, thoughtfully, intentionally, with clear boundaries and unwavering ownership.
AI is the most well-read but context-blind consultant I've ever worked with. It has perfect recall but no judgment. It can pattern-match across millions of codebases but can't tell me which trade-off matters for my users.
Here's what keeps me up at night: If AI can do 80% of what I used to do, what's the 20% that matters?
I think it's this: The ability to hold competing priorities in my head and make a call. The willingness to say "this is good enough" or "this needs another day." The experience to know when a pattern applies and when it doesn't. The judgment to understand what the business actually needs versus what it asked for. The relationships that tell me who will be affected by this decision.
It's the irreducible human core, context, judgment, care, and the willingness to be accountable when things go wrong.
My job is to care. My job is to think. My job is to decide.
That's what this manifesto protects.
The Short Version
If you remember nothing else:
- Think first. AI accelerates execution, not strategy.
- Own everything. There's no "the AI did it" defense. Speed serves quality, never the reverse.
- Stay curious. Convenience is the enemy of mastery.
- Know when to disconnect. Some growth only happens in struggle.
- Welcome friction. If AI feels too easy, you're probably over-relying.
- Keep your voice. Authenticity can't be generated.
- Bring the context. AI knows patterns; you know why they matter here.
- Lift others. AI should open doors, not close them.
This manifesto is a living document. As AI evolves, so will my understanding. But these principles, ownership, curiosity, authenticity, accountability, those don't change. They're not about AI at all. They're about what it means to be a professional.