When the Dragons of Production Sleep

Sarah Gruneisen
Jun 9
9 min read

Today was XKE day at Xebia.

Every two weeks, normal work pauses for an afternoon and something unusual happens.

Anyone can propose a topic.

Anyone can attend.

There are no gatekeepers deciding whose ideas are worthy. No requirement to be the most senior person in the room. No expectation that the presenter is the world’s leading expert on the topic.

Just people sharing what they know, what they are learning, what they are exploring, and sometimes what they are still trying to figure out.

XKE stands for Xebia Knowledge Exchange.

At first glance, it sounds like a knowledge-sharing event.

By the end of today, I realized it is actually something else.

It is an exercise in humanity.

I attended three sessions.

One explored generative AI learning paths.
One explored what comes after the current AI wave and what it means for Xebia.
One explored how to evaluate and test increasingly autonomous AI agents.

Three completely different topics.

Different presenters.

Different audiences.

Different questions.

Yet as the afternoon unfolded, I kept noticing the same dragon 🐉 hiding underneath every conversation.

Not a technology dragon.

A human one.

For most of my career, software development was the bottleneck.

Organizations had more ideas than they could implement.

More opportunities than they could pursue.

More customer problems than they could solve.

Building was expensive.

Building was slow.

Building required scarce expertise.

Entire professions emerged to reduce the distance between an idea and reality. We created methodologies, frameworks, architectures, deployment pipelines, testing strategies, governance models, operating models, and delivery approaches. Much of modern technology has been one long attempt to shorten the journey between imagination and implementation.

Then the bottleneck started moving.

Infrastructure became easier.

Deployment became easier.

Cloud platforms removed barriers.

Open source put powerful tools into everyone’s hands.

And now AI is compressing the distance between an idea and a working solution even further.

As people shared examples throughout the sessions, I noticed the discussion repeatedly drifting toward the same topics.

💚 Trust.

💚 Governance.

💚 Quality.

💚 Leadership.

💚 Adoption.

💚 Decision-making.

💚 Motivation.

💚 Culture.

The conversation kept moving away from technology and toward people.

Because none of those are technology problems.

They are human problems.

The more I listened, the more I found myself thinking about Agile. Twenty years ago, organizations believed Agile would solve their problems. Today, organizations believe AI will solve their problems.

The similarity is striking.

Back then, many organizations adopted Scrum ceremonies without understanding why they existed.

Today, many organizations adopt AI tools without understanding what value they are trying to create.

Back then, organizations became obsessed with process.

Today, organizations are becoming obsessed with tools.

Back then, many eventually discovered that Scrum was never the problem.

🖤 The problem was trust.

🖤 The problem was leadership.

🖤 The problem was alignment.

🖤 The problem was culture.

🖤 The problem was humans.

I suspect AI is revealing exactly the same thing.

Only faster.

One moment from the second session has been circling in my head the last couple of hours.

People were talking about autonomous agents, AI-generated software, and increasingly automated organizations. The assumption underneath many of these conversations seemed obvious. If AI makes building faster, we will build more. But a question kept forming in my mind.

🔥 Why are we building this at all?

For decades we have asked:

Can we build it?

Then we asked:

Can we build it faster?

AI is forcing a new question:

Should we build it at all?

That question changes everything.

Because if AI continues reducing the cost of production, then the real challenge is no longer creating things.

The challenge becomes deciding which things deserve to exist.

And that led me down a rabbit hole that I have not yet escaped. …….

Many discussions about AI eventually arrive at the question:

“What are humans for?”

I think that is the wrong question.

Humans existed before jobs.

Humans existed before corporations.

Humans existed before software.

Humans existed before capitalism.

Humans are not a feature of the economy.

The more interesting question is:

🔥 What will humanity want when production is no longer scarce?

Then comes a possibility that feels almost impossible to discuss without sounding ridiculous.

What if money itself changes?

⬆️ thank you Alessandro for this mind spark this weekend.

Not next year.

Not necessarily in our lifetime.

But eventually.

History is filled with economic systems that once appeared permanent.

Barter.

Coins.

Feudal systems.

Mercantilism.

Capitalism.

Digital economies.

Cryptocurrencies.

Every generation assumes the system they inherited is simply how the world works.

Until it changes.

And it always changes.

If AI eventually creates abundance at a scale we have never experienced before, we may discover that the things we currently pay for are no longer the things we value most.

After all, money was never the goal.

Money was a mechanism.

A way of allocating scarce resources.

A way of coordinating human activity.

A way of exchanging value.

But what happens when scarcity moves?

What happens if energy becomes abundant?

What happens if production becomes abundant?

What happens if knowledge becomes abundant?

What exactly are we paying for?

Perhaps the future economy revolves around reputation. Perhaps it revolves around trust. Perhaps it revolves around access, community, attention, belonging, creativity, or experiences.

Or perhaps it becomes something none of us have imagined yet.

I don’t know.

What fascinates me is not the answer.

It is the question.

Because every time humanity removes one form of scarcity, value seems to migrate somewhere else.

And if AI is fundamentally a scarcity-destroying technology, then value will migrate again.

The question is where.

The more interesting question is:

🔥 What will humanity want when production is no longer scarce?

Because almost every economic system we have ever built assumes scarcity.

Scarcity of food.

Scarcity of transportation.

Scarcity of information.

Scarcity of expertise.

Scarcity of labor.

Scarcity of software.

Historically, value emerged from scarcity.

If only a handful of people could do something, those people became valuable.

The blacksmith.

The architect.

The doctor.

The engineer.

The consultant.

The artist.

The programmer.

The leader.

Yet AI is steadily attacking scarcity itself.

Not perfectly.

Not tomorrow.

But directionally.

Imagine a future where software is nearly free.

Knowledge is nearly free.

Translation is nearly free.

Design is nearly free.

Education is nearly free.

Legal drafting is nearly free.

What remains scarce?

Because scarcity is where value lives.

One possibility is that lived experience becomes scarce.

Not information.

Experience.

AI may know what abandonment is.

But it was never removed from its mother.

AI may know what grief is.

But it never buried a brother.

AI may know what trauma is.

But it never spent years rebuilding itself.

AI may know what leadership is.

But it has never sat across from a struggling engineer wondering whether to push harder or show more compassion.

AI may know what motherhood is.

But it has never stayed awake through the night with a sick child.

Knowledge and experience are not the same thing.

Perhaps future value comes increasingly from lives that have actually been lived.

Another possibility is that trust becomes the new currency.

Today money often buys expertise.

Tomorrow expertise may become abundant.

Trust will not.

Who do you trust when the AI says one thing and your instincts say another?

Who takes responsibility when a decision goes wrong?

Who stands behind the recommendation?

Who tells the truth when it is inconvenient?

Trust may become more valuable than knowledge.

But trust requires something we rarely discuss.

Before we trust others, we must trust ourselves.

And that is where another dragon may be waiting.

What happens when people stop trusting themselves?

One future is easy to imagine.

Organizations become overconfident.

They believe AI can replace architects, engineers, product managers, leaders, and consultants.

They hand the keys to the dragon and hope for the best.

Most people can already see the risk in that future.

The more subtle danger is the opposite.

People gradually stop exercising judgment.

Not because they are lazy.

Not because they are incapable.

But because the AI is often right enough.

Just right enough to become dangerous.

Just right enough to make us doubt ourselves.

Just right enough to weaken muscles we spent decades building.

❤️‍🔥 Curiosity.

❤️‍🔥 Discernment.

❤️‍🔥 Critical thinking.

❤️‍🔥 Intuition.

❤️‍🔥 Wisdom.

❤️‍🔥 Judgment.

One presenter showed a fascinating model where judges evaluated AI systems, and other systems evaluated those judges.

At one point I found myself thinking:

Who judges the judge?

Someone even asked this question out loud (did they read my mind?!?!). The room laughed.

But underneath the joke is a profound question.

Where does authority ultimately live?

In the model?

In the framework?

In the system?

Or in humans?

Because if we stop trusting human judgment entirely, we eventually arrive at a strange place.

The machine becomes responsible for deciding whether the machine is correct.

That feels like a dragon story that rarely ends well.

This is also where I think many organizations are beginning to feel pain, even if they cannot yet name it.

The pain sounds like this:

🖤 Everyone is experimenting, but nobody knows what good looks like.

🖤 We bought tools, but we don’t see the value.

🖤 Some teams are moving fast while others are frozen.

🖤 Leadership wants results, but nobody knows how to measure success.

🖤 We have hundreds of ideas and no shared direction.

🖤 People are afraid of being left behind.

🖤 People are afraid of becoming irrelevant.

🖤 People don’t know which skills still matter.

🖤 People no longer trust their own expertise.

None of these are technology problems.

Every single one is a transformation problem.

This is why I think transformation strategy is becoming more important, not less. Transformation is not a technology strategy. It is a human adaptation strategy.

An AI strategy asks:

“What can AI do?”

A transformation strategy asks:

“What should humans do?”

An AI strategy asks:

“How do we automate?”

A transformation strategy asks:

“How do we adapt?”

An AI strategy asks:

“Which tools should we buy?”

A transformation strategy asks:

“How do we create value?”

An AI strategy asks:

“How can we move faster?”

A transformation strategy asks:

“How do we move in the right direction?”

That strategy is much larger than technology.

It includes a shared vision of value.

🔥 Decision-making models.

🔥 Trust boundaries between humans and AI.

🔥 Leadership evolution.

🔥 Operating model evolution.

🔥 New definitions of expertise.

🔥 New definitions of accountability.

And perhaps most importantly:

How to preserve human judgment while embracing machine intelligence.

Because I suspect the organizations that thrive will not be the ones that trust AI the most.

Nor the ones that trust AI the least.

They will be the ones that learn how to create a healthy partnership between the two.

As I thought about this, I realized something else…

XKE itself is a small example of that future.

In many organizations, knowledge is power.

People protect it.

Hoard it.

Build careers around possessing it.

Yet every two weeks at Xebia, people voluntarily give it away.

A software engineer can challenge a strategist.

A consultant can challenge an architect.

A newcomer can challenge someone with decades of experience.

Nobody asks what your title is before listening to your idea.

The best idea wins!!!

Or more accurately, the best ideas collide and become something none of us would have reached alone. Something better emerges!!!

The more I thought about it, the more I realized that XKE quietly embodies the values that appear on Xebia’s walls.

💜 People First.

💜 Sharing Knowledge.

💜 Customer Intimacy.

💜 Quality Without Compromise.

Today I watched all four become visible.

I saw people discussing motivation, trust, leadership, and adoption.

I saw people openly sharing knowledge without protecting territory.

I contributed to conversations about understanding customer pain rather than simply producing more content.

I saw an entire session dedicated to ensuring AI systems behave responsibly rather than blindly trusting outputs.

And suddenly another thought emerged….

If AI continues reducing the value of knowledge itself, then perhaps these values become even more important.

Because People First is not knowledge.

It is judgment.

Customer Intimacy is not knowledge.

It is empathy.

Quality Without Compromise is not knowledge.

It is discernment.

Sharing Knowledge is not really about possessing information.

It is about helping others make sense of it.

Then comes another possibility.

One that feels both beautiful and unsettling.

Perhaps meaning becomes the scarce resource.

Imagine a world where almost everything can be produced.

The question stops being “How?”

The question becomes “Why?”

Why build this?

Why pursue this?

Why dedicate your life to this?

The more abundance we create, the more purpose becomes the bottleneck.

I see hints of that already.

Many talented engineers are not leaving organizations because they lack money.

They are leaving because they lack meaning.

Perhaps the future forces us to confront that reality even more directly. And perhaps that is why the transformation conversation matters so much. Every industrial revolution looked technological on the surface.

Underneath, it was psychological.

People had to learn new identities.

New ways of working.

New ways of finding meaning.

The factories were not the hardest part.

The humans were.

The Internet was not the hardest part.

The humans were.

Agile was not the hardest part.

The humans were.

AI will not be the hardest part.

The humans will be.

By the time the final session ended, everyone made their way outside for the summer barbecue. Yummy!!!

The smell of grilled food drifted through the evening air. The sun hung low enough to soften the edges of the day. People laughed. Stories were exchanged. New conversations began. And something interesting happened.

Nobody talked about evaluation frameworks.

Nobody talked about token budgets.

Nobody talked about agent architectures.

At least not at the table I sat at.

People talked about their children.

Their travels.

Their dreams.

Their frustrations.

Their hopes.

The things that mattered before AI existed and will probably matter long after this wave passes.

Sitting there, surrounded by people exchanging ideas, stories, laughter, curiosity, and experiences, I found myself wondering if we have the question backwards.

Perhaps the future is not asking: