Reflections on Making Decisions in the Age of AI

Introduction

Lately, I’ve found myself in a strange state.

I think a lot, but act very little. I keep feeling that I haven’t fully figured out the direction, so I keep researching and waiting. But with AI evolving so quickly, the number of choices keeps growing. I’m constantly searching for the best methodology—one that would allow me to make the optimal decision at every crossroads. Yet recently, I’ve started to wonder: does the best choice really exist?

The Misconception

I’ve gradually realized that I’ve been holding onto a hidden assumption:

As long as my cognition is sharp enough and I have enough information, I can make the best possible choice.

So I keep accumulating knowledge—reading history, studying decision-making frameworks, learning from others’ successes, and always worrying that I might be missing some critical piece of information. I thought this was rationality, but looking back, isn’t this just procrastination?

I’ve been going back and forth for too long between “staying in a job” and “pursuing independent development.” More than one family member or friend has told me: you should have a clear direction for your future.

To be honest, I don’t have a clear answer myself. I just always feel like I’m not ready yet, and I can always find reasons. My programming skills still feel insufficient; there are so many things I haven’t learned. My writing still needs improvement; I need more material to work with…

In the blink of an eye, three years have passed.

Reality Never Gives You Complete Information

When I think about it carefully, the issue isn’t really about having a clear direction.

The real issue is that I’ve been avoiding one thing: making the wrong choice. What if I pick the wrong direction? What if I waste my time? What if I miss out on opportunities?

When I was younger, I always felt like there was plenty of time. But as I’ve gotten older, I’ve become increasingly aware: time cost is the most expensive cost of all.

And so rationality has become my defense mechanism—if I study a little more, think a little more, it feels like I can reduce the risk.

But the cruelest truth about reality is that information is never complete. Especially in the fast-moving field of AI. Just as I start to understand one concept, something new emerges: before I’ve fully grasped machine learning, deep neural networks appear; before I’ve finished exploring Prompt engineering, Workflow and Agent start to take center stage. The rules keep changing, and so do the directions.

If you insist on waiting until everything is perfectly clear before you act—then chances are, you’ll never act at all.

The Limits of Rational Decision-Making

There was a time when I really enjoyed playing Minesweeper. Gradually uncovering areas, safely clearing blocks, there was a sense of control that felt relaxing and satisfying. But every time, in the end, there would always be a corner where I encountered a spot that couldn’t be resolved through calculation or deduction.

That feels so similar to where I am now.

Rational decision-making is important, but it cannot eliminate risk. Decision theory itself cannot give you the single correct answer. The choices we face in reality are not like multiple-choice questions on an exam—there is no standard set of A, B, C, or D. Gathering information and thinking deeply can only improve the quality of your judgment and help you rule out options that are clearly wrong.

But no matter how much you learn or what tools you use, one thing is unavoidable:

You must bear uncertainty.

So at the end of rational decision-making, there’s always one thing you have to do: place your bet.

Choose a direction, and accept that it may be right or it may be wrong.

Looking back, in so many areas, the reason I’ve hesitated comes down to just one thing: I’m afraid to take on the risk that comes with making a choice. But if I keep hesitating, leaving uncertainty to time, then by the time opportunities and directions become clear, what I see will most likely be their aftermath.

Rationality is certainly important. But the endpoint of rationality is never certainty. It’s being willing to place your bet even in the face of uncertainty.

So, for some decisions, if I don’t act now, the cost will only keep growing. For many things, taking action immediately is, in the long run, actually the lowest-cost option.