Scribes were intellectuals once.
In ancient Egypt, in Mesopotamian courts, in medieval monasteries — being a scribe wasn't clerical work. It was power. Literacy was rare. Understanding a text, preserving it, interpreting it — these were acts of genuine intellect. The scribe's mechanical skill (writing) and cognitive skill (understanding what they wrote) were bundled together. Because so few people could do either, the bundle commanded enormous status.
Then literacy spread. The printing press arrived. And it didn't just automate the mechanical part — it unbundled the two skills. Suddenly you didn't need to understand philosophy to reproduce a philosophy text. The mechanical act of reproduction became trivial. The intellectual act of creating or interpreting the content became the only part that mattered.
The scribes didn't become publishers. Their grandchildren did.
We Are Still in a Scribe Economy
We tell this story about the past as if it's safely behind us. It's not. We are living in a scribe economy, and most of us are scribes.
Consider the software engineer. Coding started as something only a tiny elite could do — you had to understand the math, the hardware, the logic. The mechanical skill (writing code) and the intellectual skill (system design, problem-solving) were completely bundled. Being able to write a for loop was being smart.
But look at modern reality. Most day-to-day coding is pattern execution. Pulling from Stack Overflow, wiring up CRUD apps, writing boilerplate. The mechanical part dominated the job long before AI showed up. AI just made it obvious. The real cognitive work was always architecture, trade-offs, understanding what to build and why.
Lawyers — same unbundling. The profession began as rhetoric, argument, interpretation of law. Genuine intellectual combat. The modern practice is overwhelmingly document assembly, discovery review, contract redlining. The mechanical part ate the job from the inside.
Doctors. Diagnosis once required deep accumulated intuition. Now a huge portion of primary care is pattern matching against known protocols. The cognitive part — empathy, navigating uncertainty, complex multi-system judgment — is still deeply human. But most of a doctor's day isn't that.
Financial analysts building models in Excel. Consultants assembling slide decks. Accountants, radiologists, paralegals, copywriters, translators. The thread connecting all of them: these roles accumulated status and compensation based on the bundle, and the mechanical part was always the majority of billable hours.
This is an asymmetry problem.
These professions built an information asymmetry between themselves and the people they served — not through the cognitive work, but through the mechanical layer that made the cognitive work look inaccessible. The client couldn't draft their own contract, so they couldn't evaluate whether the lawyer's judgment was worth the fee or whether they were just paying for elaborate paperwork. The patient couldn't read their own scan, so they couldn't tell whether the radiologist's three minutes of attention deserved a five-figure bill. The mechanical layer was the asymmetry. AI doesn't attack cognitive work. It collapses the mechanical layer that hid how little cognitive work most of the job actually required.
Most prestigious knowledge work is a scribe job with better marketing.
How Every Profession Industrialized Itself
So how did we get here? Why did every profession wrap itself in layers of mechanical process?
Because cognitive work is hard and doesn't scale.
A brilliant lawyer can only hold so many complex cases in her head. But a law firm can bill thousands of hours of document review. The mechanical layer is what turned individual expertise into a business. You can hire associates, train them on the mechanical parts, bill their time, and build a firm. You can't easily replicate the senior partner's judgment — but you can replicate the process around it.
So every profession industrialized itself. It took the cognitive core and wrapped it in scaffolding — not because the scaffolding was the point, but because that's how you scale, hire, train, charge, and build institutions.
Medical schools don't primarily teach intuition. They teach protocols. Law schools teach case analysis frameworks. Coding bootcamps teach syntax and patterns. The mechanical layer became the credentialing system — the thing you could test, standardize, and gatekeep around.
And then something subtle happened. The mechanical layer became the identity. Lawyers started thinking of themselves as people who draft contracts, not people who exercise judgment. Coders started identifying with the act of writing code, not the act of solving problems. The means became the end.
Think about your own work. If you stripped away every task that could be automated — every template, every standard process, every routine deliverable — what would be left? How many hours of your week are genuinely cognitive? Not "cognitively demanding" in the sense of requiring concentration. Cognitive in the sense of requiring judgment that no system could replicate.
AI isn't taking away the hard part of anyone's job. It's taking away the part they organized their entire career and self-image around.
What's Left When the Wrapper Goes
If the mechanical layer isn't the value — and it isn't — then what's left?
The honest answer starts in an uncomfortable place: most people don't actually know what their cognitive core is. A lawyer who's spent fifteen years doing document review might not be a great legal thinker. She just never had to find out. The mechanical work was a place to hide — and not cruelly. The system rewarded it. Forty years of revenue depended on it being there.
Which means the real task isn't "learn to use AI tools." That's just learning new mechanical work. The real task is figuring out whether you have a cognitive core worth building around. And that's terrifying, because the answer might be no. Or at least, not yet.
The people who will thrive are the ones who were always frustrated by the mechanical parts. The lawyer who hated doc review but came alive in the courtroom. The developer who resented boilerplate but lit up during system design. The analyst who rushed through the model-building to get to the insight. Those people were already living in the cognitive layer. AI just frees them.
The people who will struggle are the ones who chose the profession because of the mechanical layer. Not as a character flaw — as a preference. Some people genuinely prefer structured, repeatable, clear work. We built an entire economy that honored that preference and paid it well. We're about to find out what happens when the economy stops honoring it.
Either way, the test is the same. Think about the last time you did something at work that felt effortless — not because it was easy, but because it was the thing you're actually for. That's the signal. Everything else was the scaffolding you built around it to justify a salary.
Cognitive work doesn't feel like work. It doesn't fill eight hours. It doesn't produce visible output every day. It looks like staring out a window, having a conversation, reading something unrelated and making a connection.
The productivity culture we built around mechanical work is going to resist this with everything it has. Volume metrics. Billable hours. Output dashboards. Performance reviews tied to visible activity. None of it survives contact with cognitive work as the actual unit of value.
What Replaces the Scribe
The scribe analogy keeps proving useful because it answers the question people are most afraid to ask: what happens to me?
The scribes didn't all starve. But the role they thought they had — keeper of the sacred mechanical act of writing — vanished. What replaced it wasn't a new kind of scribe. It was a new economy organized around what the printing press made possible. Authorship. Publishing. Editing. Translation as a craft. New cognitive layers that couldn't have existed when reproduction itself was the bottleneck.
The same shape will hold here. The cognitive layer doesn't get smaller — it gets visible. Roles built around mechanical execution disappear. Roles built around judgment, taste, persuasion, navigating ambiguity, and seeing what others miss become the entire economy.
What sat behind the mechanical layer was always more valuable. We just couldn't price it directly. The mechanical wrapper provided the billing mechanism — two hundred hours of discovery to deliver thirty seconds of legal judgment, ninety slides to deliver one strategic insight. AI strips the wrapper away.
That's not the end of professional knowledge work. It's the beginning of professional knowledge work that's actually about knowledge.
Where A8C Operates
This is the same shift A8C is built to operate inside.
Most software for service businesses got built on the same logic as the professions it served — the mechanical layer was the product. Schedules, attendance reports, document management, billing dashboards. Useful, but never the question the business actually runs on. The question the business runs on is cognitive: which customer is about to leave, which decision is about to go wrong, which signal is the one that matters this week.
That layer used to be too expensive to instrument. According to Stanford's 2025 AI Index, the inference cost of GPT-3.5-equivalent performance fell roughly 280× between late 2022 and late 2024 — a step-change that takes decision-grade judgment from a special-occasion expense to ambient overhead. The same unbundling reshaping professional work is reshaping software for the people the professions serve. Mechanical-layer products lose. Cognitive-layer products win.
The scribes didn't become publishers. Their grandchildren did. The question is whether you've been doing the job you thought you were doing. Strip away the mechanical layer and look at what's left. That's what was always real.
Find out now. Before the machine does it for you.
*Zain is the founder of A8C Ventures, a venture studio building software for service businesses where the gap between what's known and what's available is widest.
