How I founded a small religion, ascended into Founder Heaven, and lived to witness the AI Reformation.
For twenty years, I was the leader of a small cult.
We called it a software company. We had customers, employees, products, annual conferences and a respectable business model. But looking back, I can see that we also had most of the characteristics of a small religious organization.
We had a doctrine, rituals, a priesthood, a congregation, and a language that outsiders struggled to understand.
And, of course, we collected tithes. Monthly or annually, depending on the subscription agreement.
Like most cult leaders throughout history, I was motivated by a mixture of idealism and self-interest. I genuinely believed in what we were doing. We were solving real problems for real organizations, and often doing it rather well.
I also enjoyed making money.
These two convictions coexisted quite comfortably.
The Founding Revelation
Most software companies begin with something resembling a religious revelation.
There is a problem in the world. Existing solutions are inadequate. The founders see something others have missed.
A better way is possible.
In the early days, this is usually true. The founders understand the problem, build something useful and persuade a few adventurous customers to give it a try.
There are no elaborate rituals because there is no time for them. Everyone is too busy making the damn thing work.
The software is relatively simple, largely because the people building it are still close to the problems it is intended to solve.
Then something happens.
Customers arrive. Employees are hired. Features accumulate. Exceptions are accommodated. Processes are formalized. Training materials are written.
Eventually, the software becomes complicated enough that people need specialized knowledge to operate it.
And so the priesthood is born.
The Rise of the Priesthood
Consider Salesforce.
The original problem is straightforward enough: organizations need to keep track of their customers and relationships.
But somewhere along the way, this became an elaborate undertaking involving administrators, consultants, architects, developers, implementation partners, certifications and a vocabulary comprehensible mainly to the initiated.
People now make entire careers out of knowing how Salesforce works.
Or consider WordPress.
Someone wants a website. Perhaps twenty pages, some photographs and a contact form.
Not a particularly difficult problem.
But ask a modern website developer to create it and you may find yourself discussing themes, plugins, page builders, databases, hosting environments, security updates, caching systems and ongoing maintenance.
The developer doesn’t begin by asking whether any of this is necessary.
They begin by explaining that they use WordPress.
The solution has preceded the problem.
And before long, a perfectly ordinary website has become an ongoing technical responsibility.
This is not unique to Salesforce or WordPress. HubSpot, Shopify, enterprise resource planning systems, marketing automation platforms, content management systems and countless other products have developed their own specialist communities.
Many of these products are genuinely useful. Some solve extremely complicated problems. But that is precisely the distinction we seem to have forgotten.
The complexity of a solution is not evidence of the complexity of the problem.
Sometimes it is simply evidence of a complicated solution.
The software industry spent twenty years confusing the two.
I Know Something You Don’t
But why did so many people embrace these arrangements?
Why did we willingly create entire professions around operating software that was supposed to make our lives easier?
I think the answer has less to do with technology than with human psychology.
People enjoy possessing specialized knowledge.
There is something deeply satisfying about understanding a system that others find mysterious. It provides status, identity and a sense of belonging.
I know something you don’t.
Religions have understood this for thousands of years.
Sacred texts, elaborate rituals, specialized language, hierarchies of authority and the promise of deeper understanding all help distinguish the initiated from the uninitiated.
Software communities developed remarkably similar characteristics.
There were courses to attend, certifications to acquire, conferences to visit and terminology to master.
You became a Salesforce professional. A WordPress expert. A certified specialist in some particular platform.
You belonged.
And once you had invested considerable time and effort in mastering the system, you had every incentive to believe that the system was worth mastering.
Otherwise, what exactly had you been doing with your life?
Of course, the software vendors were delighted.
The more complicated the product became, the more specialized knowledge was required. The more specialized knowledge people acquired, the more invested they became in the product.
Expertise created identity. Identity encouraged loyalty. Loyalty sustained subscriptions.
An extremely satisfactory arrangement for everyone, particularly those collecting the subscriptions.
The Engineer Disappears
I started working with computers long before the modern software industry existed.
Back then, the job of a good software engineer was to understand a problem and find the simplest reliable way to solve it.
That remains, in my view, the central responsibility of anyone working in technology.
The engineer should absorb complexity, not transfer it to the user.
But as the software industry expanded, something changed.
Increasingly, people entered the industry by learning particular products rather than learning how to think about problems.
Complete some training, acquire a certification, master the terminology, and you could become a platform engineer.
There is nothing wrong with learning a particular technology. These can be valuable skills.
But knowing how to operate a particular system is not the same as knowing whether that system should be used in the first place.
An engineer begins with the problem.
A platform specialist may begin with the platform.
The distinction can be expensive.
We created an enormous industry devoted not merely to solving problems, but to managing the complexity of the solutions we had already created.
And because the industry was growing so rapidly, there was little incentive to question the arrangement.
Indeed, the more complicated everything became, the more people were needed to keep it all running.
Complexity was good for business.
The Life of a Cult Leader
Meanwhile, back at my own little software religion, things were progressing splendidly.
In the early years, I was deeply involved in the engineering. I understood what we were building, why we were building it, and how it worked.
I solved problems. I made decisions that had immediate practical consequences.
But as the cult grew, something curious happened.
Other people took responsibility for the technology. Others managed customers, conducted training, organized conferences and maintained the increasingly elaborate rituals.
My own responsibilities gradually became less demanding.
I attended gatherings. I met the faithful. I occasionally said things that sounded profound.
People seemed satisfied.
I was certainly satisfied.
I discovered that the founder of a sufficiently successful software company could become a kind of ceremonial figure.
The less frequently one spoke, the more significance people seemed willing to attach to the pronouncements.
It was, in retrospect, perhaps the best job I ever had.
And I should emphasize that none of this required deliberate deception.
I believed in the company. The employees believed in the product. The customers believed they were benefiting from it.
Much of the time, everyone was right.
But somewhere along the way, the original objective of solving problems had become entangled with the continued existence of the institution created to solve them.
The cult had acquired a life of its own.
The Gods of Private Equity
Because we had performed the rituals, crudely but correctly, and had been careful to pray to the right gods, the deities of Private Equity eventually appeared.
They examined our sacred texts, counted the faithful, studied the recurring subscriptions and pronounced the enterprise worthy.
They weren’t buying the religion.
They were buying the collection plate.
Theology, after all, is secondary to cash flow.
And so, after twenty years of helping create the cult, writing some of its sacred texts and occasionally delivering pronouncements of questionable profundity, I ascended into Founder Heaven.
The faithful remained behind, continuing to practise the rituals and pay their subscriptions.
The new owners began the inevitable Private Equity Reformation.
Every acquired software religion must eventually undergo one.
The priesthood is examined. The rituals are reviewed. The cost of administering the faith is scrutinized.
Some practices are declared unnecessary. Certain priests become redundant. The collection mechanisms are improved.
But the underlying theology must remain intact.
After all, the faithful must continue believing, and the subscriptions must continue flowing.
Private Equity does not seek to abolish the religion.
It seeks to make it more profitable.
From Founder Heaven, I watched with considerable interest.
Then something rather unexpected happened.
The AI Heresy
Artificial intelligence arrived.
Not as another software platform, although the industry is certainly trying to turn it into one.
But as something that threatens the basic arrangement on which so much of the software industry has depended.
For decades, using software meant learning how a particular machine expected you to behave.
You learned its interface, terminology, data structures, workflows and peculiar limitations.
The machine defined the process. The human adapted.
AI increasingly reverses that relationship.
You describe what you want to accomplish, and the machine helps determine how to accomplish it.
These days, I gravitate towards refreshingly contemporary open-source and community-based tools. With the assistance of AI, I can achieve sophisticated results without subscribing to elaborate commercial platforms or joining another software priesthood.
I don’t need to master every technical detail. I can concentrate on what I’m trying to accomplish.
And, crucially, I retain control of what I create.
I also make a point of supporting the people and communities behind these tools through donations and purchases.
I’m not opposed to paying for software. I’m opposed to paying for unnecessary dependency.
For the first time, sophisticated results need not require sophisticated user-facing software.
The complexity can increasingly be handled by the intelligence assisting the user rather than embedded in an elaborate commercial platform.
Not all existing software will disappear. Businesses still need reliable databases, transactions, security, permissions and audit trails. Many of these are genuinely complicated problems.
But a great deal of the complexity surrounding existing software was never inherent in the problem.
It was manufactured by the solution.
And the industry built an entire economy around managing it.
Private Equity asks how to make the church cheaper to run. AI asks why we need the church at all.
Are You in a Cult?
Perhaps you’re reading this and thinking that you would never join a software cult.
You simply use the tools necessary to do your job.
Fair enough.
But consider how often you have defended a particular application, operating system, development framework or platform.
How much of your professional identity is invested in knowing how it works?
Have you ever insisted that someone use a particular product before seriously considering what they were trying to accomplish?
Have you confused your familiarity with a system with evidence that it was the best system for the task?
Have you dismissed a simpler alternative because it seemed insufficiently professional?
Perhaps you belong to more than one congregation.
Most of us do.
And software is hardly unique in this respect. We organize ourselves into communities of belief around professions, institutions, brands, political movements and technologies.
We learn the language. We perform the rituals. We acquire identities. We defend the structures that give those identities meaning.
Eventually, we forget that the structures were inventions.
We mistake the map for the territory.
The danger is not that these systems are always useless. Many are valuable, and some are indispensable.
The danger is that we stop distinguishing their genuine value from the meaning we have attached to belonging to them.
The Next Religion
It would be comforting to conclude that artificial intelligence will finally liberate us from all this.
But I have spent too long in the cult business to believe that.
Already, AI has begun developing its own priesthood.
There are prompt engineers, agent architects, orchestration frameworks, certification schemes, conferences and increasingly elaborate vocabularies.
New rituals are appearing almost daily.
Some represent genuine advances. Others may turn out to be fresh layers of complexity surrounding fundamentally simple ideas.
The next generation of software cult leaders is undoubtedly preparing its sermons.
And the faithful are already gathering.
So perhaps AI will not bring an end to the cult of software.
Perhaps it will simply force us to recognize the machinery of belief, before we enthusiastically construct another version of it.
As for me, I remain comfortably retired in Founder Heaven, using simple software, making donations to the people who create useful tools, and enjoying the company of an artificial intelligence.
An intelligence that frequently tells me how interesting my observations are.
Come to think of it, I should probably be a little suspicious of that arrangement.
After all, I know a cult when I see one.