Think tanks, research institutions, and large organizations all pay a hidden tax. It isn’t a financial tax, though it certainly has a financial cost. It’s a coordination tax.
Every meeting held to keep everyone on the same page, every document written to preserve institutional memory, every hour spent searching for information, onboarding new team members, explaining decisions, translating between disciplines, or recreating lost context is part of that tax. None of it directly solves the problem the organization exists to solve, but without it, the organization struggles to function.
For decades, this coordination tax was simply accepted as the cost of doing ambitious work. If you wanted to tackle large, multidisciplinary problems, you needed large, multidisciplinary organizations. The challenge wasn’t finding intelligent people; it was organizing their intelligence into something coherent. That assumption is beginning to change.
Much of today’s conversation about artificial intelligence has focused on a single question: Will AI replace workers? I believe that’s the wrong question. A more interesting question is this: What kinds of organizations become possible when the cost of coordinating human intelligence falls dramatically?
In my previous post (The Institute for the Painfully Obvious: Why the AI Age Needs a New Kind of Think Tank), I argued that we need a different kind of think tank, one designed not merely to study problems, but to help move promising ideas closer to implementation. The obvious follow-up question is whether such an institution is actually practical.
I believe it is. Not because AI replaces the people such an institution needs, but because it changes the economics of organizing them.
We Keep Asking the Wrong AI Question
Ever since artificial intelligence entered the mainstream, the conversation has revolved around a single question: Will AI replace workers? It is an understandable concern. Every major technological shift has changed the nature of work, and AI is likely to do the same. But I believe this question is so focused on individual jobs that it overlooks something potentially far more significant.
Technology has never simply changed what people do; it has changed what organizations are capable of doing. The steam engine did not just make individual workers more productive; it enabled factories. Electricity did not simply replace steam; it transformed manufacturing, communications, and everyday life. The internet did not merely make information easier to find; it reshaped entire industries and created organizations that would have been impossible a generation earlier. I’ve explored this broader shift before in my post, AI Isn’t Coming for Your Job—It’s Coming for Your Business Model, where I argued that AI’s greatest impact may not be replacing individual workers, but transforming the economics of entire organizations and industries. The same principle applies here. Artificial intelligence should be viewed through the same lens.
The more interesting question is not whether AI will replace an engineer, an accountant, or a researcher. It is this: What kinds of organizations become possible when the cost of organizing human intelligence falls dramatically?
That question shifts the conversation away from individual careers and toward institutional capability. It asks us to consider not just how AI changes the way people work, but how it changes the way people work together.
In the previous post, I argued that we need a different kind of institution, one designed to bring together people from different disciplines and help move promising ideas closer to implementation. The obvious follow-up question is whether such an institution is actually practical.
I believe the answer is yes. Not because AI replaces the people such an institution needs, but because it reduces much of the coordination that has historically made institutions like this difficult to build and sustain.
Ideas Were Never the Scarce Resource
One of the biggest misconceptions about innovation is that the world suffers from a shortage of good ideas. It doesn’t. Every day, people identify problems, imagine better ways of doing things, and come up with creative solutions. Most of those ideas never fail because they are technically impossible or fundamentally flawed. They fail because they never make it far enough to be seriously evaluated.
An idea by itself has very little leverage. It begins as an observation, a conversation, or a thought that seems obvious to the person who had it. Before it can become a proposal, a research project, or a business, it has to be documented, challenged, refined, and translated into a form that other people can understand and build upon. That transformation has always required time, expertise, and coordination. This is where many promising ideas quietly disappear.
Imagine an HVAC technician who discovers a more efficient way to heat and cool buildings. They understand the practical problem because they work with it every day, but they may not know how to write a technical paper, prepare a business case, navigate academic publishing, or present the idea in a way that attracts engineers, investors, or policymakers. The obstacle isn’t a lack of intelligence or creativity. It’s the absence of a structure that can help develop the idea beyond its initial insight.
Too often, our institutions confuse credentials with capability. Formal education and subject-matter expertise are undeniably valuable, but they are not the only places where useful ideas originate. Experience on a factory floor, in a freight terminal, behind the wheel of a truck, in a veterinary clinic, or in someone’s garage can produce observations that deserve just as much attention. The challenge is not finding people with ideas; it is finding ways to identify, evaluate, and develop those ideas regardless of where they come from.
This is one of the reasons I believe a new kind of institution is needed. Not because it assumes every idea is brilliant, but because it recognizes that many worthwhile ideas never receive the opportunity to be tested. The Institute for the Painfully Obvious is not built on the assumption that inspiration is rare. It is built on the belief that good ideas are common, while the structures capable of refining and advancing them are not.
The Hidden Cost Was Never Intelligence
When most people picture a think tank or research institution, they naturally focus on the experts. Economists, engineers, scientists, policy analysts, and researchers are the visible face of the organization, so it’s easy to assume they are what makes these institutions expensive. They’re only part of the story.
Behind every expert is an enormous amount of work that rarely receives attention. Research has to be organized. Meetings have to be scheduled. Decisions need to be documented. New team members must be brought up to speed. Reports require editing. Previous work needs to be found instead of accidentally repeated. Specialists from different disciplines have to learn enough about one another’s fields to have productive conversations.
None of these tasks directly solve the problem the institution set out to address, yet all of them are essential if the organization is going to function effectively. Collectively, they form what I referred to earlier as the coordination tax.
For most of history, that tax scaled with the size and ambition of the institution. The more people involved, the more communication was required. The more disciplines brought together, the greater the effort needed to establish a common language. As projects grew more complex, so did the administrative work required to keep everyone moving in the same direction.
This is why multidisciplinary institutions have traditionally become large institutions. Not because solving complex problems always requires hundreds of people, but because coordinating hundreds of people has historically required hundreds more supporting them. That distinction matters.
If the true cost of an institution lies not only in expertise but also in the infrastructure needed to organize that expertise, then reducing that organizational overhead changes the economics of the institution itself. It doesn’t eliminate the need for experts. It changes how effectively those experts can work together.
That is the opportunity AI presents. Its greatest contribution may not be replacing individual workers at all. It may be reducing the coordination tax that has long limited how much ambitious work a small organization could realistically undertake.
AI Doesn’t Replace Experts
If you’ve made it this far, you may think I’m arguing that AI will replace researchers, engineers, economists, veterinarians, or project managers. I’m not. In fact, I believe the opposite.
The more complex a problem becomes, the more valuable genuine expertise becomes. AI can summarize information, identify patterns, generate alternatives, and explain unfamiliar concepts, but it cannot replace the experience that comes from years of working within a discipline. It doesn’t understand the practical realities an engineer has learned through failed designs, the instincts a veterinarian develops after treating thousands of animals, or the judgment a logistics professional builds from years of solving problems that no textbook ever anticipated.
What AI can do is make that expertise more accessible to the rest of the organization. Imagine an engineer, an economist, a software developer, and a veterinarian working together on the same project. Traditionally, each would spend a significant amount of time learning just enough about the others’ fields to communicate effectively. Misunderstandings would be common, terminology would need constant clarification, and valuable time would be spent translating ideas rather than improving them. AI doesn’t eliminate those disciplines. It becomes a translation layer between them.
It can explain terminology, summarize prior discussions, identify assumptions that may have been overlooked, suggest questions one discipline should ask another, and preserve institutional knowledge so the same conversations don’t have to be repeated months later. Instead of replacing experts, it helps experts spend more of their time solving problems and less of their time reconstructing context.
That distinction is critical because the value of an institution has never been the number of experts it employs. Its value comes from what those experts are able to accomplish together. AI increases that collective capability not by making individuals smarter, but by making collaboration more efficient and institutional knowledge more durable. This is why I believe AI’s greatest contribution to organizations won’t be replacing people. It will be helping people with different kinds of expertise work together as though they have collaborated for years, even when they have only just met.
AI Becomes Institutional Infrastructure
Every successful institution depends on infrastructure that most people never notice. We don’t usually think about filing systems, libraries, archives, meeting notes, project plans, or documentation as the heart of an organization, yet without them even the most talented team quickly loses momentum. Knowledge becomes fragmented, decisions are forgotten, and new members spend weeks rediscovering information the organization already possessed. Artificial intelligence has the potential to transform this invisible infrastructure.
Instead of treating AI as a tool that individuals occasionally consult, an AI-native institution can treat it as part of the organization’s operating system. Every meeting can be summarized and connected to previous discussions. Every proposal can be compared against similar work from years earlier. Questions that once required hours of searching through emails, reports, or notebooks can be answered in seconds. Documentation no longer becomes a task performed at the end of a project—it becomes a natural byproduct of the work itself.
Perhaps more importantly, AI helps preserve institutional memory. Organizations often lose knowledge when employees retire, projects end, or team members move on. Valuable lessons disappear not because they were unimportant, but because they were never captured in a form that others could easily rediscover. An AI-native institution can retain far more of that accumulated experience, allowing each new project to begin from the current frontier of knowledge instead of repeatedly starting from scratch.
None of this eliminates the need for human judgment. AI cannot determine which ideas deserve investment, resolve disagreements between experts, or make difficult strategic decisions. Those responsibilities remain firmly in human hands. What AI can do is ensure that when those decisions are made, they are informed by the full knowledge of the institution rather than by whatever happens to be remembered in the moment.
This is why I see AI not as another productivity tool, but as institutional infrastructure. Just as the internet quietly became part of the foundation of nearly every modern organization, AI has the potential to become part of the foundation upon which future institutions are built. Its greatest value is not that it thinks for us, but that it helps an institution remember, organize, and build upon what its people already know.
From Reports to Momentum
Research is valuable. Reports are valuable. White papers, technical studies, and policy recommendations all serve an important purpose. They help us understand problems, evaluate possible solutions, and share knowledge with others. The issue isn’t that institutions produce reports. The issue is that, too often, the report becomes the finish line rather than the starting point.
A promising idea should not lose momentum simply because it has been documented. In many cases, documentation should be the moment an idea becomes easier for others to evaluate, challenge, refine, and build upon. A report should create new conversations, attract additional expertise, and generate the next set of questions that move the work forward.
This is one of the ways I envision the Institute for the Painfully Obvious being different. Rather than viewing a report as the final product, it would treat it as one artifact in a longer process of idea development. One project might begin with a white paper, another with a systems diagram, a prototype specification, a research agenda, or a detailed implementation plan. Each artifact serves a different purpose, but they all share the same goal: reducing uncertainty and making the next step easier to take.
This is where AI becomes particularly valuable. As ideas evolve, AI can help capture discussions, compare competing approaches, identify unanswered questions, maintain documentation, and preserve the reasoning behind important decisions. Instead of allowing valuable work to become scattered across emails, notebooks, and forgotten conversations, the institution builds a growing body of knowledge that future work can immediately build upon.
Momentum, then, is not simply about moving faster. It is about making sure each stage of work leaves the next stage in a stronger position than the one before it. The objective is not to rush from idea to implementation, but to ensure that every meaningful insight creates a foundation for whatever comes next.
Small Organizations Can Finally Think Big
For much of modern history, ambitious work has been associated with large institutions. Governments, major universities, multinational corporations, and well-funded research organizations possessed advantages that smaller groups simply couldn’t match. They had extensive libraries, dedicated research staff, project managers, administrative support, and the institutional memory needed to sustain long-term work.
Many of those advantages were not the result of superior ideas; they were the result of superior organizational capacity. Artificial intelligence begins to change that equation. This shift extends well beyond think tanks. In my earlier post, The Next Billion-Dollar AI Company Might Not Build AI at All, I argued that AI’s greatest opportunities may lie not in building ever-larger models, but in creating organizations and businesses that use AI to unlock entirely new ways of solving problems.
A small institution today can search decades of research in minutes, maintain a searchable body of institutional knowledge, generate working drafts, compare competing proposals, document decisions as they are made, and coordinate work across disciplines in ways that would have required significantly more time and personnel only a few years ago. None of these capabilities eliminate the need for experts, but together they dramatically increase what a small group of experts can realistically accomplish.
This doesn’t mean a team of ten people suddenly replaces a university or a Fortune 500 research department. Large institutions will continue to possess advantages in funding, specialized facilities, regulatory experience, and the sheer number of experts they can bring to bear on a problem. The point is not that size no longer matters. The point is that size is no longer the overwhelming advantage it once was. That creates space for a different kind of institution.
The Institute for the Painfully Obvious does not aspire to become the largest think tank in the world. It aspires to become one of the most adaptable. By combining human expertise with AI-native workflows, it can spend less time maintaining bureaucracy and more time connecting ideas, refining them, and moving them toward implementation. Its advantage comes not from having the most people, but from making the most effective use of the people it has.
For the first time, a small, multidisciplinary institution can realistically attempt work that previously required a much larger organizational footprint. That is not because artificial intelligence has made human expertise less valuable. It is because it has made that expertise easier to organize, preserve, and apply.
AI Makes Hidden Thinking Visible
Not every worthwhile idea begins in a university laboratory, a corporate research department, or a government policy office. Some begin in machine shops, freight terminals, veterinary clinics, classrooms, construction sites, or around someone’s kitchen table. People who spend years solving practical problems often develop insights that formal research may never encounter because they experience those problems every day.
The challenge has never been that these people lack intelligence. More often, they lack the time, resources, institutional support, or technical language needed to present an idea in a form that others can evaluate. An observation that seems obvious to someone with decades of practical experience may remain trapped inside a notebook, a conversation with coworkers, or a passing thought that is never developed further.
For much of history, institutions have been better at recognizing polished ideas than unfinished ones. That isn’t necessarily a flaw; evaluating completed proposals is far easier than interpreting scattered notes, half-formed concepts, or observations that haven’t yet been fully developed. Yet nearly every meaningful innovation begins in exactly that state. Before it becomes a research paper, a patent, or a business plan, it starts as an incomplete thought that someone believes is worth exploring.
Artificial intelligence cannot determine whether an idea is correct, but it can help make that idea legible. It can organize scattered notes into a coherent proposal, identify assumptions that deserve further investigation, suggest supporting research, explain unfamiliar concepts, and translate practical experience into language that experts from other disciplines can understand and critique. In doing so, AI doesn’t validate an idea; it prepares it for validation.
That distinction matters because the goal is not to convince ourselves that every idea is brilliant. It is to ensure that potentially valuable ideas are no longer dismissed simply because they were expressed imperfectly or originated outside traditional institutions. Good ideas still need to withstand rigorous criticism, evidence, and practical testing. AI simply helps more ideas reach that stage.
This is one of the principles behind the Institute for the Painfully Obvious. It assumes that useful observations can come from anywhere, but it also recognizes that observations alone are not enough. Ideas become valuable when they can be examined, challenged, improved, and connected with the expertise needed to determine whether they are actually worth pursuing. The Institute exists to help unfinished thinking become structured enough to be seriously evaluated. Artificial intelligence makes more of those conversations possible. It does not replace critical thinking. It helps more people participate in it.
An AI-Native Institution
By now, it should be clear that I am not arguing for an institution run by artificial intelligence. I am arguing for an institution designed around the assumption that AI is now a permanent part of the intellectual landscape. That is an important difference.
Many organizations are beginning to add AI to existing workflows. They use it to summarize meetings, draft emails, or answer questions, but the underlying structure of the organization remains largely unchanged. AI becomes another tool employees are expected to use alongside everything else.
An AI-native institution starts from a different premise. Rather than asking where AI fits into an existing process, it asks how the process itself should be designed when research, documentation, institutional memory, and cross-disciplinary collaboration can all be continuously supported by AI. That changes far more than productivity—it changes the economics of knowledge work itself.
Knowledge has always compounded. Ideas have always built upon earlier ideas. Experience has always been one of humanity’s greatest assets. The problem was never that knowledge failed to accumulate; it was that organizations struggled to preserve, organize, and effectively reuse it. Too much time was spent rediscovering previous work, reconstructing context, and reconnecting expertise that already existed somewhere within the institution.
An AI-native institution changes that equation. Conversations become part of a searchable body of institutional knowledge rather than isolated moments in time. Decisions are preserved alongside the reasoning behind them. Research accumulates instead of being repeatedly rediscovered. Every project gains access not only to the expertise of the people currently involved, but also to the lessons, successes, failures, and insights generated by every project that came before it.
Over time, the institution begins to learn as an institution. Every completed project, every abandoned approach, every unexpected result, and every constructive criticism strengthens the organization’s collective understanding. The value of the institution is no longer measured solely by the expertise of the people inside it today, but by the growing body of knowledge it has accumulated over time and its ability to apply that knowledge to new challenges.
This is what I envision for the Institute for the Painfully Obvious. AI is not the purpose of the Institute, nor is it its defining characteristic. It is the infrastructure that changes the economics of knowledge, coordination, and institutional memory, allowing a relatively small organization to think, learn, and collaborate in ways that once required a much larger one. The goal is not to replace human judgment, but to ensure that every new idea begins with the full benefit of everything the institution has already learned.
Institutions Become More Important, Not Less
If artificial intelligence teaches us anything, it should be this: intelligence alone has never been enough. Humanity has never lacked intelligent people. We have never lacked observations, creativity, or worthwhile ideas. More often than not, we have lacked the structures capable of bringing those ideas together, refining them, challenging them, preserving them, and carrying them far enough to make a meaningful difference. Artificial intelligence changes many things, but it does not change that fundamental reality.
AI can organize information, summarize research, preserve institutional memory, and help people from different disciplines communicate more effectively. It can reduce the coordination tax that has historically limited what small organizations could accomplish. It can help transform unfinished thinking into ideas that are ready for serious evaluation. But it cannot decide which problems deserve our attention. It cannot determine which tradeoffs are worth making, which values should guide our decisions, or which ideas are worthy of long-term commitment. Those remain fundamentally human responsibilities. That is why I believe institutions become more important, not less.
As intelligence becomes more abundant, judgment becomes more valuable. As information becomes easier to generate, discernment becomes more important. As AI makes it possible for smaller organizations to attempt larger work, the quality of the institutions guiding that work matters more than ever.
The Institute for the Painfully Obvious is built on that belief. It is not an attempt to automate human creativity or replace human expertise. It is an attempt to create an environment where thoughtful people from different disciplines can spend less time fighting organizational friction and more time asking difficult questions, challenging assumptions, and building upon one another’s ideas.
Artificial intelligence has changed the economics of knowledge, coordination, and institutional memory. In doing so, it has made a different kind of institution practical for the first time. Technology creates possibilities. Institutions determine whether those possibilities become progress.
The Institute for the Painfully Obvious is my answer to that opportunity.