Have I lost my critical eye?

The advice is old and it is good: fall in love with the problem, not with your solution. I have given some version of it in workshops, in kickoffs, in rooms where a client arrived with the answer already written down and had to be walked back to the question. I believed it every time I said it. I still do.

Then I launched a venture of my own, and did the exact opposite.

I thought I had found the gem. That particular feeling, the one where the idea is so obviously right that examining it would only slow you down. And because I now had tools that could take me from idea to working thing in an afternoon, nothing slowed me down. I built. It was the most fun I have had working in years.

Somewhere in there I stopped being in love with the problem and started being in love with the product.

The tools did not cause this. They removed the distance.

I want to be careful here, because the easy version of this essay is the one where AI is the villain, and that version is wrong. Falling for your own solution is a very old human failure. Designers did it long before any of this. I have watched people do it with a foam model and a marker.

What changed is the distance between having an idea and holding the thing. That distance used to be measured in days, and the days were not empty. They were where the doubt lived.

When I started working this way (building fast, prompting, iterating, watching the thing appear), I felt like I had grown wings. Everything suddenly seemed possible. That feeling is real and I am not going to pretend otherwise. But the shortcut to the solution has become so short that you can arrive before you have finished asking whether you should go.

What the slow version actually was

I have been trying to describe what I lost, and it keeps coming back to something very small.

I used to draw. Not properly: a little sketch in the margin of a page, three seconds, immediately deleted. Useless as a drawing. But drawing it and killing it gave me the idea for something else. Then that one would go the same way. It was a conversation with myself, and I did not think of it as a process at all, because it did not look like work. It looked like fidgeting.

That loop was doing two jobs at once. It was generating ideas, and it was killing them. The generating part is the part everyone talks about, and the part the new tools do spectacularly well. The killing part came free with it, because every idea had to survive my own hand before it could survive anything else.

Now the generating is free and the killing is not. They have come apart.

The same thing happens with other people, and there it matters more. Put an idea in front of several stakeholders and something useful occurs: each of them gets a turn to push back: on the work, on the judgment behind it, on whether the question underneath is even the right one. Not politely agreeing. Actually interrogating it. That exchange is the most reliable critical instrument I know, and it is slow by nature. It needs a version to react to, time to form a reaction, and a reason to bother.

When you can produce a new version faster than anyone can form an objection to the last one, the instrument stops working. Nobody decided to switch it off.

What this costs, precisely

It is not that I have become worse at thinking. I do not believe that, and I would be suspicious of anyone who claimed it about themselves after a year.

It is that self-criticism used to be a by-product and is now a separate task. It used to arrive on its own, carried by the friction: by the sketch that failed, by the week between the idea and the prototype, by the colleague who asked the annoying question on Thursday. None of that friction is coming back, and I would not want most of it back.

So it has to be deliberate now. It has to be scheduled, which is an absurd thing to write about doubt. But the alternative is what happened to me: three months of genuine delight, building something well, without seriously re-examining whether it answered a real problem.

The step sideways cost me actual effort. Going back to the offer, back to the problem, asking the question I would have asked any client on day one: that took more discipline than building the whole thing had.

Where that leaves me

I am in the middle of that re-examination now, and I do not have the tidy ending.

I still believe in what I built. I might be wrong about that, and I am aware that “I still believe in it” is exactly what someone in the grip of this failure would say. That is the uncomfortable part. The same conviction that got me here is the one I am now using to judge whether to continue.

What I have committed to is the part I would have demanded of a client: if the honest answer is that it solves no real problem, it gets killed. Not softened, not repositioned. Killed. I have not reached that answer. I have also not ruled it out.

So the question in the title stays open. I do not think I have lost my critical eye. I think it used to come along for free, and now it does not. It took losing three months of it to notice.

The part that actually worries me

My own case is not the worrying part. What it implies is.

I had everything that was supposed to protect me. Years of asking the question. The habit of criticising my own work before anyone else could. Teams that taught me to be contradicted and to want it. None of it held. Which means the safeguard was never the expertise. It was the slowness. The expertise only worked because the slowness gave it somewhere to happen.

The slowness is not coming back.

Now multiply that by an entire profession accelerating at once, all of us shipping answers faster than we can examine the questions. The failure is not that we build bad things. It is an asymmetry: the value of what gets built can trend towards nothing while its effect on people’s lives does not trend towards nothing at all. A solution can be elegant, well made, beautifully argued in the deck, and still turn out to be an ecological write-off, an economic one, or the quiet author of a habit nobody wanted to acquire.

I am not going to list the canonical examples. Everyone has the same three, and reciting them mostly serves to reassure the reciter that they would not have done the same. I would have. I just did, at a smaller scale, with better intentions and a faster toolchain.

What follows from that is not caution, which is useless advice, but something more specific and more annoying: studying the consequences of what we build used to be a phase of the work, carried along by the pace. It now has to be a decision taken against the pace, against the momentum of your own project, at the moment when everything is working and nobody wants to stop.

If we do not take it deliberately, the thing we will get very good at is answering, faster and faster, problems that were never there.