Don't Lose Yourself in the Chat Window
I spend hours every day talking to AI. Claude, ChatGPT, Cursor - they're my coworkers now. I've written about how much more productive this makes me.
But I've also noticed something uncomfortable: it's easy to stop thinking for yourself.
The drift
It happens subtly. You ask a question. You get an answer. The answer sounds smart, well-reasoned, confident. So you accept it. You move on.
Then you do it again. And again.
At some point, you're not really making decisions anymore. You're ratifying AI's decisions. You've become the "approve" button in your own workflow.
I caught myself doing this recently. I was working through a product decision - nothing critical, just a feature priority question. I asked Claude. Claude gave a thoughtful answer with good reasoning. I nodded along. I was about to implement it when I realized: I hadn't actually thought about it myself. I'd just... accepted.
That's the drift. And it's worth paying attention to.
Why this happens
Part of it is that AI often is smarter than us - at least in specific ways. It has more information. It can hold more context. It doesn't get tired or emotional or distracted.
When AI gives you an answer that's better than what you would have come up with, it's rational to defer. That's the whole point of using these tools.
But "better at retrieving and synthesizing information" isn't the same as "better at knowing what matters to you." AI doesn't know your values, your constraints, your history - except what you tell it. And you probably don't tell it everything.
AI gives you confident answers based on incomplete context, and you accept them because they sound right. That's the trap.
The context problem
The quality of AI's guidance depends entirely on the context you provide.
If you ask "should I take this job?" without explaining your financial situation, your family constraints, your career goals, and your risk tolerance - you'll get generic advice that sounds wise but might be completely wrong for you.
AI doesn't know what you didn't tell it. It can't read your mind. It fills in the gaps with assumptions, and those assumptions might not match your reality.
I've seen this in my own work. Early in a project, when I haven't loaded enough context, AI will suggest approaches that seem reasonable but contradict decisions I made weeks ago. It's not wrong - it just doesn't know.
This is part of why I've been thinking about context so much lately. The better your AI understands your situation, the better its guidance. But even perfect context doesn't replace your judgment.
Staying in the game
A few things that help me:
Challenge the answer. When AI suggests something, ask it to argue the other side. "What would be the case against this?" or "What am I missing?" This keeps you engaged instead of just nodding along.
Notice when you're ratifying vs. deciding. There's a difference between "AI helped me think through this and I agree" and "AI said this so I'll do it." The first is collaboration. The second is abdication.
Remember what you bring. AI doesn't have your taste, your experience, your relationships, your values. It doesn't know what you've tried before and why it didn't work. It doesn't know what matters to you. You do.
Don't outsource your identity. It's one thing to delegate dinner choices. It's another to let AI shape how you think about your career, your relationships, your creative work. The more consequential the domain, the more important it is to stay engaged.
The balance
I'm not saying don't use AI. I use it constantly. It's the most powerful tool I've ever had.
But powerful tools require attention. The more capable AI gets, the easier it is to let it drive. And the easier it is to drift - to lose track of your own judgment, your own perspective, your own voice.
The goal isn't to resist AI. It's to stay in the conversation. To remain the author of your own decisions, even when AI is helping you write them.
Because the answers are only as good as the context you give. And no one knows your context better than you.