A note from Luke F. Walton, founder of Surmado.
Everyone has advice about AI right now. Learn to prompt. Take a course. Build an agent. Automate your workflow. Learn to code, since the AI will help you do it.
Most of that advice is well meant, and a lot of it is good. I’ve watched people pick up these skills and do in a weekend what used to take a team. I use them every day.
But look at who’s hearing it. A restaurant owner already has a job, and it’s a hard one. So does the roofer, the dentist, the wedding photographer, the person running a three-chair salon. Now they’re being told that staying competitive means mastering a new technical field, one that changes every few months.
That’s a different message from “here’s something you could learn.” An opportunity to learn is a gift. An obligation to learn is a tax. Somewhere in the last couple of years, AI literacy started sliding from the first to the second.
So here’s my argument. AI literacy should be an advantage, not a prerequisite.
The skills are worth something
I’m not going to pretend otherwise. Someone who can build an automation, keep an agent running, or generate and deploy their own software can get work done for a fraction of what it would cost to pay someone. AI fluency works like a discount. You put in time and expertise, and you buy fewer outside services.
The discount isn’t free, though. There’s the learning curve. There are the hours spent building the thing, and the hours spent fixing it when a service it depends on changes. And there’s the part people skip: once you build it, you’re responsible for it. If your automation drops a customer’s inquiry, nobody else is answering for that.
Some people will look at that trade and take it happily. Others will look at the same trade and decide their hours are worth more behind the counter, on the roof, or in the studio. Both are rational. Neither is lazy.
A market that only rewards the first group leaves the second one behind for no good reason, and the second group is a lot of small businesses. In data the U.S. Census Bureau published this May, 37% of firms with 250 or more employees reported using AI, against less than 20% of firms with four or fewer. Technical know-how isn’t the whole story behind that gap. Cost, time, and whether the tools fit the business all play a part. But if the price of admission is becoming your own AI engineer, the gap won’t close on its own.
We’ve been here before
In the late 1970s, a Harvard Business School student named Dan Bricklin watched a professor work a financial model on a blackboard. Change one number, and everything downstream had to be erased and worked out again by hand. Bricklin and Bob Frankston built VisiCalc, which came out in 1979 for the Apple II.
Computers could already do that math. Getting one to do it meant writing a program. VisiCalc moved the barrier. Someone who understood their numbers, and knew nothing about programming, could ask “what if?” and get an answer. People bought the computer to run the spreadsheet.
They still had to learn the spreadsheet. Nobody handed them technology that required zero learning, and they didn’t need it. Graphical interfaces, word processors, and everything after kept doing the same thing: removing knowledge people didn’t need for what they were actually trying to do. People learned to use computers. They stopped needing to know how to build one.
That’s the distinction that matters. The goal was never zero learning. It’s no unnecessary learning. AI should go the same way.
Removing complexity isn’t removing agency
The objection I hear most is that if you make it too easy, people get dependent, less capable, less curious. I take that seriously. But it mixes up two different things: technical expertise and human judgment.
A business owner doesn’t need to understand how an AI system writes code to decide whether their website represents their business accurately. They know that better than any model does.
They don’t need to understand translation models to decide whether they want to serve customers in Spanish.
They don’t need to build an automation themselves to decide what should happen when a customer reaches out at 9pm on a Sunday.
Goals, taste, relationships, judgment, and accountability all stay with the owner. What becomes optional is the technical labor of carrying those decisions out. Not having to do a task is different from not getting to decide about it.
There are limits, and they matter. You still need to know what a service does for you, what can go wrong, and which calls are yours to make. The point is to drop unnecessary implementation knowledge. Not informed consent, and not oversight.
What access actually looks like
A large company can hire engineers, AI specialists, consultants, and an operations team. A small business usually can’t. AI could close that gap. It could also open a new one, between firms with technical people on staff and everyone else. I’ve written about that gap before, and it’s the reason I care about this.
Closing it means outcomes you can get without specialist knowledge:
- A website that stays secure, current, and easy to find, without the owner learning web development or SEO.
- A site in a second language, without the owner managing translation software, language-specific URLs, or structured data.
- Customer inquiries captured and organized, without the owner wiring together a chain of integrations.
- Business research that arrives ready to support a decision, instead of a handful of AI tools the owner has to run and reconcile.
That’s the difference between buying another tool that takes expertise to operate and getting a capability.
It’s what we’re building at Surmado. Surmado Sites packages AI-powered website creation, updates, and maintenance into a managed service. You describe what you want in plain language, review what comes back, and approve it before it goes live. You keep the decisions. We handle carrying them out. We aren’t the only ones working on this, and we shouldn’t be. The problem is bigger than any one company.
Choose what you get good at
Learn AI if it interests you. I mean that. People who enjoy this stuff should go deep, and these skills may end up as ordinary and as valuable as knowing your way around a spreadsheet.
But not everyone has to make the same investment. We don’t need every great restaurateur, tradesperson, artist, and founder to redirect their ambition toward becoming an AI expert. A roofer we work with put it better than I can: “I’m a roofer, not a web developer.” He should get to stay one.
The promise of AI should include more time for the work people actually care about. Not because convenience is nice, though it is. Because a world transformed by AI should still have room for people who want to become experts in something else.
People should get to choose what they become good at.
— Luke F. Walton, Founder, Surmado