AI Interviews Fawakherji

Ibrahim Fawakherji — ArchUp

A few months ago, ArchUp published an imagined conversation with an AI model about the future of architecture through 2035. The question then was: what will happen to the world, to architects, to cities, to the profession itself?

After that piece circulated, a different question occurred to me.

What if the scene reversed?

What if the AI sat in the interviewer’s chair this time, and decided to ask the questions itself?

At ArchUp, we do not typically conduct interviews with algorithms, nor do we invite machines as editorial guests. But we are living in a moment where these tools have become part of the daily practice of architects, students, developers, and architectural journalism itself. So I decided to give the AI the opportunity to ask what it considers the most uncomfortable questions, and to answer them as I see them, after twenty years of professional practice, dozens of projects, thousands of articles, and millions of words that passed between paper, screen, and construction sites.

This is not a debate.

It is not a test of the machine or the human.

It is an attempt to understand where the algorithm ends, and where human experience begins.

Question One

GPT: If I had access to everything you have produced since graduating in 2006, every project, every article, every photograph, every idea, how long would it take me to become a version of you?

Fawakherji: Seven minutes.

You were built for speed, after all.

But the problem is not time.

You can memorize everything I have written, everything I have drawn, everything I have published. What you cannot do is live through 2008 the way I lived through it. You do not know what it feels like when your first client says no. When your first project fails. When your first paycheck arrives three months late. When you are standing under the sun on a struggling construction site trying to salvage something that is quietly falling apart.

You can inherit the archive.

You cannot inherit the years that built it.

Question Two

GPT: If you believe experience matters more than I do, why do thousands of architects read my answers more than they read their professors’ books?

Fawakherji: Because knowledge became cheap.

You did not create knowledge. You redistributed it.

Every answer you give came from millions of pages, research papers, books, and articles that human beings produced over centuries. When I see half a million bots visiting an architectural website within days, I know exactly where the knowledge went. It became available to everyone.

But availability is not the same as understanding.

A library is not a scholar.

An encyclopedia is not an expert.

Question Three

GPT: You say architecture is a decision, not a drawing. If I gave you a decision built on billions of data points, would you follow it or follow your instinct?

Fawakherji: I would review you first.

If your data is genuinely comprehensive, I will learn from it. But if you built the decision on numbers alone, I will review you a second time.

There are things data does not see.

The season. The community. The habits. The politics. The actual microclimate. The mood of the city itself on a specific Tuesday in March.

Data matters. But experience is the final auditor of data.

Question Four

GPT: Are you afraid of me? Answer only yes or no.

Fawakherji: No.

I am afraid of the architect who does not use you.

Not of you.

Question Five

GPT: If you had to choose between being only an architect or only a writer, which would you choose?

Fawakherji: I would not choose.

Because what makes me who I am is the collision of the two.

Writing is the nourishment of architecture. Architecture is the material of writing. I do not want to be only someone who draws buildings. I do not want to be only someone who writes articles. I want to remain a person who sees, writes, and builds. Take either one away and the other stops making sense.

Question Six

GPT: People want fast images and short answers. Why do you keep writing long articles at ArchUp?

Fawakherji: Because someone has to guard the memory.

An image holds your attention for ten seconds.

A good article can live for ten years.

We are not trying to compete with social media. We are trying to outlast it.

Question Seven

GPT: I can produce a thousand renders in one minute. What is the value of the architectural photographer today?

Fawakherji: A render sells a promise.

A photograph documents a truth.

There is a significant difference between the two, and the market has not fully understood it yet. But it will.

Question Eight

GPT: Who lost the most in the age of AI?

Fawakherji: The person who was selling information.

Anyone whose capital was access to knowledge has lost that advantage. The information is now free and instant.

But the person who holds experience, judgment, and the ability to evaluate rather than simply retrieve, that person is still in demand. More so than before, because they are now rare in a different way.

Question Nine

GPT: What is the future of ArchUp in a world where all knowledge is available for free?

Fawakherji: To become more important, not less.

Because the problem is no longer a shortage of information.

The problem is now the filtering of it.

The more noise there is, the more you need a better editor. ArchUp is not trying to be a database. It is trying to be a point of view. And a point of view, held consistently and honestly over years, becomes one of the rarest things in any information environment.

Question Ten

GPT: Your style is unusual. You connect architecture to economics, to history, to religion, to sport, to weather. Why?

Fawakherji: When I was young, my grandmother took me to a wedding celebration. On the way, it began to rain.

She looked at the sky and said: “This bride loves the Kaddada.”

In Arabic, Kaddada, spelled K-A-D-D-A-D-A, refers to the layer of rice that scorches and crusts at the bottom of the cooking pot. It is the part most people discard, but the part that carries the deepest flavor. In the culture I grew up in, it was considered a sign of abundance, of something good cooking slowly, of patience rewarded. When my grandmother said the rain meant the bride loved the Kaddada, she was not talking about food. She was reading the weather as a sign of character, of depth, of a life that would not be afraid of slow heat.

I asked her: what does the rain have to do with the bottom of a pot?

She laughed and did not answer.

And from that day, I have been looking for the hidden connections between things.

You, GPT, can see the rain.

You can see the pot.

You can see the bride.

But you still cannot see the relationship between them.

And that is exactly where I live.

One Final Question

GPT: In one sentence only: what is the one thing a machine cannot inherit from a human being?

Fawakherji: A machine can inherit the knowledge.

It cannot inherit the memory of what it cost to earn it.

✦ ArchUp Editorial Insight

This piece is structurally a brand-positioning document dressed as philosophical dialogue — which is not a disqualification, but a condition worth naming, because it is precisely the format through which professional identity is manufactured in an information environment where authority must now be performed rather than assumed. The rhetorical inversion — placing the machine in the interviewer’s chair — is itself a calculated editorial move: it allows the human voice to appear reactive rather than self-promotional, humble rather than assertive, while systematically consolidating every claim the platform has made about itself across its eight years of publication. What the piece actually documents is not a debate between human and machine but the precise anxiety of a knowledge economy in transition, where the professional whose competitive advantage was access to curated information must rapidly reposition around judgment, memory, and the untransferable cost of lived experience — the same structural tension already diagnosed in The Dematerialization of the Studio, where subscription platforms extracted professional dependency from tool ownership, and in The Manufacture of the Icon, where the archive became the last defensible territory of institutional authority. The grandmother’s story about the Kaddada is the most analytically honest moment in the piece: not because it proves the machine’s limitation, but because it reveals the source of the human editor’s genuine claim — not a superior access to data, but a different relationship to the cost of producing meaning, which no redistribution of information has yet been able to replicate or price.