We used to believe coding created value. In the age of AI, that belief is breaking down.
When anyone can build software with natural language, the real question becomes: where does value live now?
There was a time when knowing how to code felt like holding fire. That’s how it felt when I wrote my first “Hello World” in HTML. That’s me below. No, it’s not, but that’s how I felt.
In a heavily digital world, coding feels like bending reality. You could automate systems, build products from nothing but a laptop and intent. The barrier to entry was high, which created scarcity. Scarcity, historically, is where value lives.
That's no longer the world we're in. We're in an era of abundant intelligence, where building software requires curiosity, a problem to solve, and the ability to write in natural language — in any language.
So the real question becomes: if everyone can build, where does value live?
Value Was Never in Code.
Let me say something uncomfortable: the value was never in knowing how to code.
Coding was leverage.
It was a gatekeeping skill that allowed you to execute.
But it was never the value itself.
If coding alone created wealth, every engineer would be wildly rich. That is clearly not how the world works. Almost no one we associate with great wealth earned it simply by writing code.
Coding gets you off the starting blocks. But it does not win the race.
The companies that reshaped industries did not win because they wrote better code.
Airbnb didn’t reach 800,000 listings because of its tech stack. Airbnb's Brian Chesky has never claimed that software architecture built Airbnb’s marketplace.
Uber didn’t scale to over 100 countries because it discovered a new coding pattern.
And Stripe didn’t end up processing $1.9 trillion in payments because its founders were seasoned engineers from IBM or Microsoft. When they started, Patrick Collison and John Collison were just 21 and 19.
They won because they saw real-world value mispriced. They identified the friction that others ignored. They understood human behaviour, incentives, and economics deeply—and then built on those insights.
They were miners. Coding was the shovel. But the goal was always the gold.
The tools will continue to evolve. The equipment will get faster, cheaper, and more powerful. But the destination remains the same.
Value has always belonged to those who know where to dig.
And that distinction matters more now than ever.
AI Has Removed the Gate, Not the Need for Judgment and Insight.
Artificial intelligence has dramatically lowered the barrier to execution. A domain expert no longer needs to wait for engineering bandwidth to test an idea. A doctor can prototype a triage assistant. A real estate operator can build affordability calculators or automate client intake. A teacher can experiment with adaptive learning tools.
The technical gate is weakening. But something else becomes more important when the gate disappears: judgment and insight.
AI reduces build time. It does not reduce thinking. It does not reduce problem selection.
It does not reduce the need for strategic clarity.
You still need to know what to build. You still need to understand who it is for. You still need to understand why someone would pay for it.
And this is where domain expertise becomes powerful. AI is extraordinarily capable at execution, synthesis, and pattern recognition. It lacks lived context unless you provide it.
A mortgage operator understands underwriting friction in ways a generic model cannot.
A hospital administrator understands workflow bottlenecks that are invisible from the outside. A supply chain manager knows where margins silently leak.
AI becomes the engine.
Domain knowledge becomes the steering wheel.
Is Coding Dead? No, It Just Moved Up the Stack
There is a temptation to say coding is obsolete. Dead. A relic of the pre-AI world. That is lazy thinking.
Coding is not dead. It has simply moved up the stack. Much of what once required syntax can now be done with natural language.
This is not a new pattern. Computing has always moved toward higher levels of abstraction. Tools become easier to use. More people can build. Progress continues.
What changes is not the need for systems, but how we interact with them.
Engineers who understand architecture, scalability, security, and system design will become more leveraged, not less. In a world powered by AI, we still need people to orchestrate models, design guardrails, manage costs, build hybrid systems, and ensure reliability.
The difference is this. Being a strong coder is no longer enough on its own.
Engineers must think like value creators. The same is true in reverse. Domain experts must learn to build, at least at a systems level.
The wall between builder and operator is collapsing.
Distribution Becomes the New Scarcity
When building becomes easier, distribution becomes harder.
In an era where almost anyone can ship a product, the advantage shifts to those who control attention and reach. Companies like Amazon, Apple, and Meta Platforms command enormous power not because they alone know how to build software, but because they control distribution at scale.
In the age of abundant intelligence, distribution itself can be automated and optimized. AI agents can generate content, test messaging, personalize outreach, analyze engagement, and continuously refine campaigns. You can now run experiments that once required entire marketing teams.
But automation does not guarantee value. Distribution amplifies whatever you point it at. If the product lacks depth or clarity, scaling it only accelerates irrelevance.
The most leveraged individuals will combine three things: domain expertise, AI systems, and distribution control. That stack creates a compounding advantage. It is how moats are built.
Knowing When Not to Use AI
This is where maturity shows.
Not everything needs AI.
AI systems, particularly large language models, are probabilistic. They can produce different outputs from the same input. They can hallucinate, misinterpret edge cases, and become expensive when deployed carelessly at scale.
If a problem requires exact calculations, regulatory precision, or zero tolerance for error, a deterministic system is often better. A simple rules engine may outperform an AI model. A traditional database query may be more reliable than an intelligent retrieval layer. A calculator will always be better at arithmetic than a language model predicting tokens.
AI shines when ambiguity exists. It excels at synthesizing unstructured data, summarizing complex inputs, drafting content, categorizing messy information, and augmenting human reasoning. It is exceptional at compressing cognitive work.
But using AI where simplicity would suffice is not innovation. It is hype-theatre.
The strongest systems today are hybrid systems. They combine deterministic cores with intelligent augmentation. They use AI where it creates leverage and avoid it where it introduces fragility or unnecessary cost.
That is the disciplined approach to building.
Respect the Limitations, But Watch the Curve.
AI models predict patterns. They do not truly understand. They lack lived experience and contextual grounding unless connected to real systems and data. They can hallucinate with confidence and still require supervision and feedback.
At the same time, capability is improving rapidly. Research from independent organizations such as Model Evaluation and Threat Research shows consistent improvements in reasoning benchmarks, multimodal understanding, task length, and efficiency.
The right posture is neither blind faith nor dismissal. It is disciplined curiosity.
What is needed now is constant and strategic experimentation. The models change quickly, and capabilities expand with each release.
Learn the tools while they are imperfect. Build with them as they evolve. Design systems that can improve as the models improve.
So, Where Is Value Going?
Value in the age of abundant intelligence will flow to those who can see clearly. Those who identify real problems, understand users deeply, orchestrate intelligent systems thoughtfully, and distribute solutions effectively.
It will not reward those who cling to scarcity narratives. It will not reward those who mistake tools for strategy.
AI does not eliminate value creation. It accelerates execution.
When intelligence becomes cheap, vision becomes expensive.
We are not entering an age where machines replace humans. We are entering an age of super-leveraged ordinary humans.
You do not need to be super rich to be super leveraged. You need to understand where value truly lives.
Leverage has always favoured those who see clearly.
Be one of them.
