Learning Library

← Back to Library

AI: The Fourth Way to Scale Expertise

Key Points

  • Historically, expertise could only be “scaled” by working longer hours, hiring less‑experienced staff, or raising prices—each method ultimately hits a hard limit and creates bottlenecks.
  • These three approaches fail because true expertise resides in the expert’s brain and can’t be duplicated or delegated without loss of depth or quality.
  • AI introduces a fourth, previously untapped avenue by handling the “translation layer” that converts raw expertise into reusable, scalable outputs.
  • By capturing, structuring, and automating expert knowledge, AI lets businesses extend the reach of a single expert far beyond the constraints of time, staffing, or price.
  • This shift transforms expertise from a non‑scalable asset into a repeatable product, unlocking growth opportunities for any knowledge‑intensive profession.

Full Transcript

# AI: The Fourth Way to Scale Expertise **Source:** [https://www.youtube.com/watch?v=L32th5fXPw8](https://www.youtube.com/watch?v=L32th5fXPw8) **Duration:** 00:09:25 ## Summary - Historically, expertise could only be “scaled” by working longer hours, hiring less‑experienced staff, or raising prices—each method ultimately hits a hard limit and creates bottlenecks. - These three approaches fail because true expertise resides in the expert’s brain and can’t be duplicated or delegated without loss of depth or quality. - AI introduces a fourth, previously untapped avenue by handling the “translation layer” that converts raw expertise into reusable, scalable outputs. - By capturing, structuring, and automating expert knowledge, AI lets businesses extend the reach of a single expert far beyond the constraints of time, staffing, or price. - This shift transforms expertise from a non‑scalable asset into a repeatable product, unlocking growth opportunities for any knowledge‑intensive profession. ## Sections - [00:00:00](https://www.youtube.com/watch?v=L32th5fXPw8&t=0s) **AI Unlocks Scalable Expertise** - The speaker argues that expertise has historically been unscalable, critiques three traditional methods of trying to expand it, and asserts that artificial intelligence now provides a fourth, truly scalable way to multiply expert knowledge. - [00:03:21](https://www.youtube.com/watch?v=L32th5fXPw8&t=201s) **Separating Expertise from Documentation with AI** - The speaker highlights how the universal bottleneck of turning rapid professional knowledge into slow, detailed documentation can be eliminated by using AI to convert brief voice memos into polished outputs, letting experts focus on their core work. - [00:06:47](https://www.youtube.com/watch?v=L32th5fXPw8&t=407s) **Structured Context Drives Scalable Documentation** - The speaker argues that precisely articulating requirements through templatized, structured context is the key to obtaining high‑quality, focused output and unlocking the ability to scale expertise beyond documentation bottlenecks. ## Full Transcript
0:00For thousands of years, there have been 0:01only three ways to scale your expertise. 0:04And AI just invented the fourth one. 0:06Nobody talks about it, but expertise is 0:08actually the one thing in business that 0:09doesn't scale. You can scale products by 0:11manufacturing more. You can scale 0:12content by publishing more. You can 0:14scale distribution, reach more people. 0:16But expertise only lives in your brain. 0:19Anyone who's hired will tell you you 0:22cannot magically scale expertise by 0:23hiring people. It lives in your brain. 0:25And traditionally, there have been only 0:27three ways to scale experts. And all of 0:29them suck. All of them are bad. And AI 0:32created a fourth way, and I'm going to 0:33get into it. Most people don't know 0:34about it yet. So, let me show you the 0:36problem. Option number one, this sucks. 0:38Work more hours. Say you're a lawyer. 0:40You're great at what you do. The client 0:42demand goes up. You work nights and 0:44weekends. You burn out. This doesn't 0:46scale. Or else you just pad the billing 0:48hours. Whatever it is. But the point is, 0:50you can't scale your hours infinitely. 0:52There is not infinite time in the day. 0:53Option two, hire people. I told you this 0:56sucked. It sucks. You hire associates. 0:58Maybe you hire junior lawyers. They 1:00don't scale expertise. They dilute 1:03expertise. In this analogy, the junior 1:06associate isn't you. The nurse that 1:08scales with the doctor isn't the same as 1:09the doctor. Although some nurses will 1:11tell you they know more than doctors, 1:13but the point is they don't have the 1:15years of pattern recognition that go 1:18with true expertise. Every piece of work 1:21that a junior person touches needs to be 1:23reviewed by the person with expertise. 1:25This goes for lawyers. This goes for 1:27medicine. It goes for anybody with 1:29genuine expertise. When I have worked 1:30with senior senior senior engineers, 1:33people who are at the principal level or 1:34above, it's the same thing. And you end 1:36up trading your work for management work 1:39and that is draining and you're still 1:41the bottleneck. Option number three, the 1:44only way we've found to scale that 1:46doesn't suck as bad as working more 1:47hours and hiring more people is raising 1:49your prices. So great, you charge more. 1:52You go, you raise your hourly rates. The 1:53lawyer raises the hourly rates to 600 1:55bucks an hour, a,000 bucks an hour, but 1:57there's a ceiling. Eventually, you're 1:59too expensive for most clients, and 2:01you've traded volume for rate, and 2:02you're still limited to the amount of 2:04time that you have. These are your 2:05options. This is why any expertise-based 2:08business hits a wall. This is true 2:10whether you're in the legal profession 2:12or whether you are a senior principal 2:14architect or whether you are a uh 2:19tradesman, whether you're in plumbing or 2:21whether you're installing HVAC systems, 2:23your expertise hits a wall by scale. 2:27Your knowledge is the asset and it's 2:28been trapped and there's been only one 2:30of you. Enter AI. Because the thing that 2:33people don't say out loud is the 2:35constraint has not been really your 2:37expertise. It's been the translation 2:40layer. So, let me show you what I mean. 2:41Let's say an HVAC contractor is 2:44diagnosing a failing system that takes 2:4620 minutes. She knows what's wrong. It's 2:48an undersized unit. It's leaking 2:50duckworked. Whatever it is, she has 15 2:52years of experience. She can do it right 2:53away. Writing the estimate that explains 2:55this to a homeowner takes longer. It has 2:57to be professionally formatted. It has 2:59to be translated into the appropriate 3:00language. It has to be explaining why 3:03the solution and not the cheaper one. It 3:05has to add photos. It has to be 3:06persuasive enough to win the job. Then 3:08it has to be delivered. Her expertise 3:11did not take nearly as long as 3:14documenting her expertise. Right? In 3:16this analogy, her expertise takes just a 3:17couple of minutes, a few minutes. 3:19Documenting her expertise takes much, 3:21much longer to get it all put together. 3:23That ratio has been the problem. That 3:26has been the bottleneck. That has been 3:28what doesn't scale well. And this is 3:31true everywhere. A senior uh attorney 3:33might know the legal strategy in just a 3:35few minutes but take a long time to 3:36write the brief and even with a 3:37parallegal getting the intent down and 3:40getting it polished takes a long time. 3:42The doctor may know the diagnosis but 3:44completing the chart note takes a while. 3:46Right? Architect knows the design 3:47solution. Creating the presentation 3:49takes much much longer. Your brain works 3:51fast. Documentation works slow. The 3:54fourth way, the AI way attacks that 3:57bottleneck. But we don't talk about it 3:59that way. So I want to talk about it 4:00here. You need to separate your 4:02expertise in whatever domain you're in 4:04from documentation. And so instead of 4:07viewing it as I need to write up this 4:10whole thing, look at it as here's an 4:12example from the HVAC contractor, right? 4:13Five minute voice memo on the site. You 4:15capture the context walking through 4:17looking at the HVAC unit. Then tell the 4:20AI, turn this into a professional 4:22estimate, no jargon. Please emphasize 4:24comfort and energy savings because 4:26that's what the client really emphasized 4:28to me. And then she goes on about her 4:30work, right? She goes drives to the next 4:32job site and she can review the output 4:34on a mobile phone in the car, add in any 4:37adjustment she has to pricing, upload a 4:39couple of photos. It takes just a few 4:41minutes and she may multiply her ability 4:43to generate estimates by five as a 4:45result. That's the breakthrough. That's 4:47the fourth way. Now, let me get to the 4:49principles that underly this because not 4:51everybody here is going to be an HVAC 4:52contractor. Not everybody's a lawyer, a 4:54doctor. What are the principles that we 4:55can scale to whether we work in tech or 4:57not in tech that help us think about 4:59expertise differently? Number one, 5:01expertise compounds and documentation 5:05erodess or doesn't compound. So you get 5:07better at your craft. Whatever your 5:09craft might be, it might be code, it 5:10might be medicine, it might be legal, 5:12you get better at it every year. You see 5:14your patterns faster. You make decisions 5:16with more confidence. This is why a lot 5:18of business studies show expertise 5:19peaking much later in the career arc 5:22than people realize into your age 50s 5:23and beyond. But writing still takes the 5:26same amount of time. AI makes 5:28documentation compound with your 5:30expertise. That's the breakthrough, 5:31right? Writing, I still take as long to 5:34type as I took 10 years ago because I've 5:36kind of maxed out my typing speed. And 5:38for voice, I'm talking as fast as I'm 5:40going to talk. AI helps your 5:41documentation compound with your 5:43expertise because it can write for you. 5:45Principle number two, quality control 5:47lives with you. So the lawyer still 5:49reviews for legal accuracy. The the 5:51doctor checks the diagnosis. You don't 5:54outsource your judgment. You outsource 5:56the translation. And I want to call out 5:58principle one and principle two are tied 6:00into the rest of this. If you're like, 6:02well, this is obvious. There's a trick 6:03here that we're going to get to that we 6:05don't talk about enough when it comes to 6:07removing expertise. So principle number 6:09three, this is the first trick. The 8020 6:11threshold. AI will get you 80% much much 6:15faster than anybody else. Right? Like 6:17it's faster than a parallegal, faster 6:19than a nurse taking notes. Not that they 6:20do. Faster than anybody I've seen, 6:23right? Like I've seen those walls of 6:25text appear. That is correct because the 6:2920% that isn't done, the mess, that 6:31still requires your expertise. You want 6:34your expertise. You want to get hands-on 6:35and fingertippy with your business in 6:37the right way. You need to set up the 6:40prompt and the context to make sure that 6:42works. And that brings me to principle 6:44number four. Context is your multiplier. 6:45This is the secret. If you want 6:47expertise to compound with your 6:49documentation, if you want quality 6:51control, if you want to touch the right 6:5320% of a document draft, which is what 6:55I'm advocating for here, regardless of 6:57your expertise, maybe it's a code 6:58architecture review, maybe it's the 7:00estimate for the HVAC machine, context 7:03is the multiplier. So the better you can 7:05articulate what you need, the better the 7:07output. If you just say write an 7:09estimate, it's not going to be great. If 7:11you just say write a generic NDA, it's 7:13not going to be great. Your ability to 7:15articulate is your superpower. And 7:17specifically, your ability to articulate 7:19in templatized structured context forms. 7:22And that's what I'm going to get into 7:23here. You want to get into a place where 7:26you can say, "This is what I need. This 7:29is the context you need for this task. 7:31This is the context you need about me. 7:33this is the context you need about the 7:34client or the patient or whatever the 7:36reader if it's technical documentation 7:38and this is my expectation for this 7:40draft specifically. The more clear you 7:43can be the more you are likely to have 7:46the right 20% to touch and be able to 7:48scale your expertise appropriately and 7:50that volume that you get to creates more 7:53optionality for you. This is the payoff 7:55when documentation was the bottleneck. 7:57You turn down work. You can only scale 7:59your time so far. We talked about that 8:01by starting to attack this bottleneck, 8:03you unlock optionality because your 8:05expertise is no longer bottlenecked. 8:07That is the 10x return that we get with 8:09AI that none of our older scaling levers 8:11delivered on expertise. That is why I am 8:14convinced this is transformative and I 8:16don't think we've talked about it 8:17enough. We talk about automation all the 8:18time. We don't talk about this idea of 8:20human expertise scaling with AI and I 8:22think we should. So, how do you actually 8:24do this? I would invite you to pick one 8:26repetitive task, something that requires 8:28your expertise that you do every week 8:30that takes you a couple hours or more. 8:31And give the AI at least four things. 8:34Your role, your audience, your goal, and 8:37your constraints. Those are all very, 8:38very important. You'll get to a first 8:40draft, and your goal in this initial 8:42piece is just to see if the first draft 8:44is that correct 80%. And if it's not, go 8:46back and refine the things you're giving 8:48the AI till you get to the right 80% and 8:50you know where you need to add 8:51expertise. So you review for accuracy in 8:54your domain and that's the stuff that 8:56you can verify, deliver value on, get 8:58fingertippy with the business, maximize 9:00your human expertise and ship. Once you 9:03start to get into this habit, you're 9:04going to find lots of other two plus 9:06hour tasks that you can start to 9:08automate this way. So here's my 9:09challenge for you. What's what's the one 9:12thing this week where you spend hours 9:14translating your expertise? How can you 9:17not be the bottleneck? How can you 9:19practice AI lifting that bottleneck for 9:22you?