How AI creates a fourth way to scale professional expertise beyond the traditional three methods
Nate B Jones presents a solo argument about why expertise has historically been impossible to scale, and how AI changes that.
Summary
Nate B Jones argues that expertise-based professionals — lawyers, doctors, architects, HVAC contractors, engineers — have always faced the same fundamental scaling problem: their knowledge is fast, but translating it into documentation is slow. He identifies three traditional scaling methods (working more hours, hiring staff, raising prices) and explains why all three fail to truly scale expertise. His central claim is that AI represents a genuine fourth method — not by replacing expertise, but by attacking the translation bottleneck between what an expert knows and what they need to produce. The key insight is that context-rich prompting allows documentation to compound alongside expertise, rather than remaining a fixed time cost regardless of how skilled the professional becomes.
Key Takeaways
FULL TRANSCRIPT
The Three Traditional Ways to Scale Expertise — and Why They All Fail
Nate B Jones: For thousands of years, there have been only three ways to scale your expertise. And AI just invented the fourth one. Nobody talks about it, but expertise is actually the one thing in business that doesn't scale. You can scale products by manufacturing more. You can scale content by publishing more. You can scale distribution, reach more people. But expertise only lives in your brain.
Anyone who's hired will tell you you cannot magically scale expertise by hiring people. It lives in your brain. And traditionally, there have been only three ways to scale experts — and all of them are bad.
AI created a fourth way, and I'm going to get into it. Most people don't know about it yet. So let me show you the problem.
Option number one: work more hours. Say you're a lawyer. You're great at what you do. Client demand goes up. You work nights and weekends. You burn out. This doesn't scale. Or else you just pad the billing hours. Whatever it is, the point is you can't scale your hours infinitely. There is not infinite time in the day.
Option two: hire people. You hire associates. Maybe you hire junior lawyers. They don't scale expertise — they dilute expertise. In this analogy, the junior associate isn't you. The nurse that scales with the doctor isn't the same as the doctor. Although some nurses will tell you they know more than doctors, but the point is they don't have the years of pattern recognition that go with true expertise. Every piece of work that a junior person touches needs to be reviewed by the person with expertise. This goes for lawyers. This goes for medicine. It goes for anybody with genuine expertise. When I have worked with very senior engineers — people who are at the principal level or above — it's the same thing. You end up trading your work for management work, and that is draining, and you're still the bottleneck.
Option number three: the only way we've found to scale that doesn't suck as bad as working more hours and hiring more people is raising your prices. You charge more. You raise your hourly rates. The lawyer raises the hourly rate to $600 an hour, $1,000 an hour — but there's a ceiling. Eventually you're too expensive for most clients, and you've traded volume for rate, and you're still limited to the amount of time that you have.
These are your options. This is why any expertise-based business hits a wall. This is true whether you're in the legal profession, whether you are a senior principal architect, whether you are a tradesman — whether you're in plumbing or installing HVAC systems. Your expertise hits a wall by scale. Your knowledge is the asset, and it's been trapped, and there's been only one of you.
The Real Bottleneck: The Translation Layer
Enter AI. Because the thing that people don't say out loud is the constraint has not really been your expertise. It's been the translation layer.
Let me show you what I mean. Say an HVAC contractor is diagnosing a failing system — that takes 20 minutes. She knows what's wrong. It's an undersized unit. It's leaking ductwork. Whatever it is, she has 15 years of experience. She can identify it right away. Writing the estimate that explains this to a homeowner takes longer. It has to be professionally formatted. It has to be translated into the appropriate language. It has to explain why this solution and not the cheaper one. It has to add photos. It has to be persuasive enough to win the job. Then it has to be delivered.
Her expertise did not take nearly as long as documenting her expertise. In this analogy, her expertise takes just a couple of minutes. Documenting her expertise takes much, much longer to get it all put together. That ratio has been the problem. That has been the bottleneck. That has been what doesn't scale well.
And this is true everywhere. A senior attorney might know the legal strategy in just a few minutes but take a long time to write the brief — and even with a paralegal, getting the intent down and getting it polished takes a long time. The doctor may know the diagnosis but completing the chart note takes a while. The architect knows the design solution; creating the presentation takes much, much longer. Your brain works fast. Documentation works slow.
The Fourth Way: AI Attacks the Translation Bottleneck
The fourth way — the AI way — attacks that bottleneck. But we don't talk about it that way. So I want to talk about it here.
You need to separate your expertise in whatever domain you're in from documentation. Instead of viewing it as "I need to write up this whole thing," look at it as — here's an example from the HVAC contractor — a five-minute voice memo on the site. You capture the context walking through and looking at the HVAC unit. Then you tell the AI: turn this into a professional estimate, no jargon, please emphasize comfort and energy savings because that's what the client really emphasized to me. And then she goes on about her work. She drives to the next job site, and she can review the output on a mobile phone in the car, add any adjustments to pricing, upload a couple of photos. It takes just a few minutes, and she may multiply her ability to generate estimates by five as a result.
That's the breakthrough. That's the fourth way.
The Principles Behind Scaling Expertise with AI
Now let me get to the principles that underlie this, because not everybody here is going to be an HVAC contractor. Not everybody's a lawyer or a doctor. What are the principles that we can apply — whether we work in tech or not in tech — that help us think about expertise differently?
Principle number one: expertise compounds, but documentation doesn't. You get better at your craft every year — whether it's code, medicine, or law. You see patterns faster. You make decisions with more confidence. This is why a lot of business studies show expertise peaking much later in the career arc than people realize — into your 50s and beyond. But writing still takes the same amount of time. I still take as long to type as I took ten years ago, because I've maxed out my typing speed. And for voice, I'm talking as fast as I'm going to talk. AI helps your documentation compound with your expertise, because it can write for you.
Principle number two: quality control lives with you. The lawyer still reviews for legal accuracy. The doctor checks the diagnosis. You don't outsource your judgment. You outsource the translation.
Principle number three: the 80/20 threshold. AI will get you 80% of the way there much faster than anybody else — faster than a paralegal, faster than anyone I've seen. But the 20% that isn't done still requires your expertise. You want to get hands-on with your business in the right way. You need to set up the prompt and the context to make sure that works.
Principle number four: context is your multiplier. This is the secret. If you want expertise to compound with your documentation, if you want quality control, if you want to touch the right 20% of a document draft — which is what I'm advocating for here, regardless of your expertise, whether it's a code architecture review or an estimate for an HVAC system — context is the multiplier. The better you can articulate what you need, the better the output. If you just say "write an estimate," it's not going to be great. If you just say "write a generic NDA," it's not going to be great.
Your ability to articulate is your superpower. Specifically, your ability to articulate in templatized, structured context forms. You want to get to a place where you can say: this is what I need; this is the context you need for this task; this is the context you need about me; this is the context you need about the client, the patient, or whatever the reader — if it's technical documentation; and this is my expectation for this draft specifically. The more clearly you can articulate that, the more likely you are to have the right 20% to touch, and to be able to scale your expertise appropriately.
The Payoff: Optionality
And the volume that you get to creates more optionality for you. This is the payoff. When documentation was the bottleneck, you turned down work. You could only scale your time so far. By starting to attack this bottleneck, you unlock optionality — because your expertise is no longer bottlenecked. That is the 10x return that we get with AI that none of our older scaling levers delivered on expertise.
That is why I am convinced this is transformative, and I don't think we've talked about it enough. We talk about automation all the time. We don't talk about this idea of human expertise scaling with AI — and I think we should.
How to Start: A Practical Framework
So how do you actually do this? I would invite you to pick one repetitive task — something that requires your expertise, that you do every week, that takes you a couple of hours or more. Give the AI at least four things: your role, your audience, your goal, and your constraints. Those are all very important.
You'll get to a first draft, and your goal in this initial piece is just to see if the first draft is that correct 80%. If it's not, go back and refine what you're giving the AI until you get to the right 80% and you know where you need to add your expertise. Review for accuracy in your domain — that's the stuff you can verify, deliver value on, get hands-on with, maximize your human expertise, and ship.
Once you start to get into this habit, you're going to find lots of other two-plus-hour tasks that you can start to approach the same way.
So here's my challenge for you: what's the one thing this week where you spend hours translating your expertise? How can you stop being the bottleneck? How can you practice having AI lift that bottleneck for you?