There is a piece of advice handed to new writers so often it has almost become wallpaper: write what you know.
It sounds sensible because it is sensible. It is also incomplete.
If I had only written about what I already understood when I started building websites, working with AI, learning SEO, figuring out analytics, experimenting with distribution, and trying to understand what actually makes an online business work, I wouldn’t have gotten very far. I simply didn’t know enough.
That turned out not to be the problem I thought it was. The real mistake would have been believing I needed to know everything before I started.
I knew some things. I knew how to think through problems, explain what I understood, learn what I didn’t, and recognize when something smelled wrong. I had decades of experience making decisions, running into problems, changing course, dealing with people, and figuring things out when the instructions weren’t particularly helpful.
And, somewhat unexpectedly, I knew a lot about dogs.
That was enough to begin. Everything else could be learned when I needed it.
Start With Something Real
There is an enormous difference between starting with no knowledge and starting without complete knowledge, and most people aren’t starting from zero.
They have a profession, a hobby, a problem they have solved, a mistake they have made repeatedly enough to finally understand it, or years of experience that have become so familiar they no longer recognize it as expertise. The difficulty is that what you already know often doesn’t look particularly impressive to you because you’ve been carrying it around for years.
Someone who has spent fifteen years repairing houses may not think much about knowing which jobs homeowners routinely underestimate. Someone who has cared for an aging parent may not realize how much practical knowledge they have accumulated about appointments, paperwork, equipment, and daily routines. Someone who has lived with rescued German Shepherds for years may think they are simply talking about their dogs—until somebody else asks the question they can answer.
That is where useful work often starts. Not with a business model, a keyword spreadsheet, or a carefully selected niche dreamed up because somebody said the search volume looked promising, but with a subject you have actually lived with long enough to have something worth saying.
Then You Discover Everything You Don’t Know
The funny part begins once you decide to do something with that knowledge.
Suppose you want to put it on a website. Fine. Now you need a website, which leads to hosting, WordPress, themes, navigation, images, accessibility, page speed, search engines, analytics, email, backups, security, internal links, and approximately nine thousand settings whose importance is impossible to determine from their names.
You learn one thing and it exposes three more things you don’t understand. That can feel like evidence that you weren’t ready, but it isn’t. It is what learning looks like when it is attached to real work.
The project reveals the curriculum.
That has become one of the most useful things I have learned from building online. I don’t need to sit down and learn everything about SEO before I touch a website. I need to learn enough SEO to solve the problem in front of me. When the next problem appears, I learn that part.
The same thing happened with AI. At first, the obvious question was which tools could do what. Once I started actually using them, better questions appeared. How do I ask better questions? How do I verify an answer? Where does AI genuinely save time, and where does it make the work worse? When should I use it for research, editing, organization, ideation, or analysis? When should I shut it off and think for myself?
Those questions did not appear because I completed an AI curriculum. They appeared because I was trying to do real work.
The Work Tells You What to Learn Next
Traditional education often works in the opposite direction. You learn a body of material first and eventually get the opportunity to apply some of it.
Building something reverses the sequence. You apply what you know until you hit a wall, and then the wall tells you what to learn.
That is what makes this kind of learning so efficient. If I am trying to understand why traffic to a website is being classified incorrectly, analytics attribution suddenly matters. If I am publishing articles that nobody is finding, search intent matters. If people are seeing something on Facebook but aren’t reaching the website, I need to understand distribution and tracking. If AI keeps producing prose that sounds polished but strangely lifeless, I need better editorial standards.
None of those subjects needed to be mastered on day one. They became important when the work required them. Instead of collecting knowledge because it might become useful someday, I was learning something because there was already a problem sitting in front of me that needed solving.
Building While You Learn Is Not the Same as Pretending You Know
There is a dangerous interpretation of all this that is worth killing immediately.
Building while you learn does not mean bluffing your way through subjects you don’t understand. It does not mean asking AI for an answer, rewriting it confidently, and presenting yourself as an authority. It does not mean publishing first and worrying about whether you were right later.
There has to be a boundary between what you know and what you are still trying to understand, and that boundary matters even more now because AI can make ignorance look remarkably polished. A person can produce a convincing explanation of something they barely understand. The grammar may be perfect, the headings may be excellent, and the conclusion may arrive wearing a suit.
It can still be wrong.
Part of becoming competent is learning to distinguish between four very different statements: I know this. I think this. I need to verify this. I don’t know this yet.
Those are not interchangeable. Knowing the difference between them may be more valuable than memorizing another hundred facts.
Research Becomes Part of the Work
Once you accept that you do not need to know everything before you begin, research stops feeling like evidence of inadequacy. It becomes part of the process.
That changed something for me. There was a time when I probably would have thought that needing to research a subject meant I didn’t know enough to write about it. Now I see research as one of the things that allows lived experience to become genuinely useful to somebody else.
Experience tells me what I have observed; research can tell me whether the evidence supports it. An opinion about why something works becomes more useful when I am willing to look for competing explanations. And when a tool behaves in a way I don’t understand, documentation can tell me whether I have found an intended feature, a temporary change, or simply misunderstood what I was looking at.
Experience gives the work texture, and research gives it support. They are stronger together than either is on its own.
Competence Comes From Doing the Work
People sometimes treat expertise as though it has to be granted. You take the course, receive the certificate, get the job title, and then you are officially allowed to know things.
That is one path. It is not the only one.
You can also become competent by repeatedly doing difficult things, paying attention to what happens, studying what you don’t understand, correcting mistakes, and getting better. There is nothing particularly glamorous about that process. Most of it looks like sitting at a computer wondering why something that worked yesterday has suddenly decided not to work today.
But competence accumulates. One solved problem becomes ten. Ten become fifty. Eventually somebody asks you a question and you realize you know the answer without looking it up, and then you remember that six months earlier you didn’t even understand the question.
The problem is that many people never give that process enough room to happen. They take another course, watch another video, download another guide, research another tool, and wait until the plan feels complete. It never does, because many of the questions you genuinely need answered do not appear until something is real.
Some needs only become visible once a website has real users. Repeat a workflow ten times and the clumsy parts stop hiding. Publishing tells you whether readers respond to what you thought mattered; sooner or later, somebody asks a question that shows you exactly where your knowledge is thin.
At some point preparation becomes avoidance wearing sensible shoes.
You have to build something.
AI Makes This Faster. It Doesn’t Remove the Requirement.
AI has made this kind of learning dramatically easier for me.
I can ask for an explanation of something I don’t understand, compare approaches, interrogate assumptions, turn documentation into plain English, and keep asking follow-up questions without worrying that I have exhausted somebody’s patience. I can use it to organize research, challenge an argument, expose gaps in my reasoning, or help me understand why a technical problem is happening.
But AI also makes it easier to skip the learning part entirely. There is a temptation to let the machine produce the answer and move on. That works right up until the answer is wrong, the circumstances change, or you have to make a judgment the machine cannot make for you. Then you discover whether you actually learned anything.
I keep coming back to the same principle: use AI to increase your ability to understand the work. Don’t use it to avoid understanding the work.
The First Version Is Allowed to Be a First Version
This can be difficult if you care about doing good work. You want the first version to be excellent. I certainly do.
But there is a difference between having standards and demanding that your current ability somehow include skills you have not developed yet.
My first websites were not as good as the websites I can build now. My early understanding of AI was not as sophisticated as it is now. The systems I use today did not arrive fully formed, and neither did my editorial process. They developed because the earlier versions eventually became inadequate.
That does not make the earlier work wasted. It made the later work possible. You build something, use it, notice what is weak, learn what you need, and rebuild. The fact that the second version is better than the first does not mean the first one should never have existed.
Your Knowledge Expands With the Thing You’re Building
This may be the part I find most interesting because I can see it clearly in the path my own work took.
Fluffy Shepherds began with something I already knew: years of living with dogs, especially rescued German Shepherds, and learning that the generic advice people repeat about them often does not explain the dog standing in front of you. Once I decided to put those experiences somewhere useful, I needed a website capable of carrying them.
Building the website created the next set of questions. Publishing an article was one thing; figuring out whether anybody could actually find it was another, which pulled me into SEO. Once traffic started appearing, I needed to understand where it came from and what visitors did when they arrived, so analytics stopped being an abstract subject and became something I had a reason to learn.
Then came distribution. If a Facebook post was being seen but the website traffic did not appear to reflect it, I had another problem to solve. That led to questions about attribution, tracking, and whether the systems I had built were actually telling me the truth.
Writing created its own problems. Working with AI made some things faster, but it also made it remarkably easy to produce prose that looked finished before the thinking was finished. That led me deeper into editorial standards, verification, workflow, research, and the question I keep coming back to now: what should the machine be doing, and what still requires human judgment?
I did not sit down at the beginning and design any of that as a curriculum. One real problem simply opened the door to the next one, and because every new subject grew out of work I was already doing, the knowledge began connecting instead of accumulating as a pile of unrelated information.
That is the part I could not have learned in advance.
Write What You Know
So yes, write what you know. Start with the dog, the business you ran, the job you did for twenty years, the garden you keep killing and rebuilding, or the thing you have spent enough time doing that you can tell somebody what usually goes wrong.
Start there, but don’t stop at the edge of what you already know.
Follow the questions. Learn what the work requires. Research the parts you cannot support from experience. Use better tools when they help. Correct yourself when the evidence changes. Build systems when repeated problems demand them, and keep producing something while all of this is happening.
Because you do not become ready and then begin. Most of the time, you begin.
And that is how you become ready.
