Skip to content

Podcast

The New AI Org Chart

The AI Daily Brief: Artificial Intelligence News and Analysis

Source ↗ ← All highlights
  • Hierarchy Was Always An Information Routing System
    • Dorsey and Roelof Botha argue hierarchy mainly exists to route information under human span-of-control limits, not because layered management is inherently superior.
    • Nathaniel Whittemore traces the lineage from Roman legions to Prussian staff systems, railroads, Taylorism, and matrix orgs as repeated attempts to manage coordination bottlenecks. Transcript: Nathaniel Whittemore The piece reads, At Sequoia, we see that speed is the best predictor of startup success. Most companies are focused on AI as a productivity enhancer. Few are focused on the potential of AI to change how we work together. Block is showing what it looks like to fundamentally rethink organization design, ultimately harnessing AI to increase speed as a compounding competitive advantage. 2,000 years before the first corporate org chart, the Roman army solved a problem that every large organization still faces. How do you coordinate thousands of people across vast distances with limited communication? Their answer was a nested hierarchy with a consistent span of control at every level. The smallest unit was the contubernium, eight soldiers who shared a tent, equipment, and a mule, led by a decanus. Ten contubernia formed a sentry of 80 men under a centurion. Six sentries made a cohort. Ten cohorts made a legion of roughly 5,000. At each layer, a named commander held to find authority, aggregating information from below and relayed decisions from above. The structure, 8 to 80 to 480 to 5,000, was an information routing protocol built around a simple human limitation. A leader can effectively manage somewhere between three and eight people. The Romans discovered this through centuries of warfare. Even today, the US Army’s hierarchical chain follows a similar pattern. We now call it span of control, and it remains the governing constraint of every large organization on earth. The next big change came from Prussia. After Napoleon’s army destroyed the Prussian forces at the Battle of Jenna in 1806, a group of reformers rebuilt the military around an uncomfortable truth. You cannot depend on individual genius at the top. You need a system. They created the general staff, a dedicated class of trained officers whose job was not to fight but to plan operations, process information, and coordinate across units. Scharenhorst, one of the reformers, intended these staff officers to quote, support incompetent generals, providing the talents that might otherwise be wanting among leaders And commanders. This was middle management before the term existed. Professionals whose purpose was to route information, pre-compute decisions, and maintain alignment across a complex organization. The military also formalized the distinction between line and staff functions. Line advances the core mission, staff provides specialized support. Every corporation still uses this vocabulary today. Military hierarchy entered the business world through the American railroads in the 1840s and 1850s. The U.S. Army lent West Point-trained engineers to private railroad companies, and these officers brought military organizational thinking with them. Staff and line hierarchies, divisional structure, bureaucratic systems of reporting and control, all of it was developed in the military before the railroads adopted it. In the mid-1850s, Daniel McCallum of the New York and Erie Railroad created the world’s first organizational chart to manage a system stretching over 500 miles with thousands of workers. The informal management styles that worked for smaller railroads were failing. Train collisions were killing people. McCallum’s chart formalized the same hierarchical logic that the Romans had used. Layers of authority, defined reporting lines, structured information flow. It became the blueprint for the modern corporation. Frederick Taylor, often called the father of scientific management, optimized what happened within that hierarchy. Taylor broke work into specialized tasks, assigned them to trained experts, and managed through measurement rather than through intuition. This produced the Functional Pyramid Organization, a structure optimized for efficiency within the information routing system that the military had pioneered and the railroads Had commercialized. The first real stress test of functional hierarchy came during World War II. The Manhattan Project required physicists, chemists, engineers, metallurgists, and military officers to work across disciplinary boundaries towards a single objective under Extreme secrecy and time pressure. Robert Oppenheimer organized Los Alamos into functional divisions, but insisted on open collaboration across them, resisting the military’s instinct to compartmentalize. When the implosion problem became critical in 1944, he reorganized the lab around it, creating cross-functional teams unlike anything in corporate America at the time. It worked, but it was a wartime exception led by a singular figure. The question the post-war business world faced was whether that kind of cross-functional coordination could be made routine. With the growth and globalization of companies after World War II, the scale limitations of functional design became acute. In 1959, McKinsey’s Gilbert Klee and Alfred DeScipio published Creating a World Enterprise in the Harvard Business Review, providing an intellectual framework for a matrix organization That combined functional specialties with divisional units. Under the leadership of Marvin Bauer, McKinsey helped companies like Shell and GE implement these principles, balancing central standards with local agility. This became the professional or modern corporation that propelled the post-war global economy. Over time, other frameworks emerged to address the complexity, rigidity, and bureaucracy of matrix structures. The McKinsey 7S framework, developed in the 1970s by Tom Peters and Robert Waterman, distinguished the hard S’s, strategy, structure, and systems, from the soft S’s, shared values, Skills, staff, and style. The core idea was that structural elements alone were insufficient. Organizational effectiveness required alignment across cultural traits and the human factors that determine whether a strategy actually succeeds. In more recent decades, technology companies have experimented aggressively with organization structure. Spotify popularized cross-functional squads with short sprint cycles. Zappos attempted holacracy, eliminating management titles entirely. Valve operated with a flat structure with no formal hierarchy. Each of these experiments revealed something about the limitations of traditional hierarchy, but none solved the underlying problem. Spotify moved back towards conventional management as it scaled. Zappos saw significant attrition. Valve’s model proved difficult to scale beyond a few hundred people. As organizations grow into the thousands, they revert to hierarchical coordination because no alternative information routing mechanism has been powerful enough to replace it. The constraint is the same one the Romans faced and the Marine Corps rediscovered in World War II. Narrowing span of control means adding layers of command, but more layers mean slower information flow. 2,000 years of organizational innovation has been an attempt to work around this trade-off without breaking it. (Time 0:01:58)
  • Block Wants A Company Built As Intelligence
    • Block wants AI to replace the coordination function of managers by building a company world model plus a customer world model from machine-readable work and transaction data.
    • Its stack is capabilities, world model, intelligence layer, and interfaces; roadmap ideas emerge when the intelligence layer fails to compose a needed solution. Transcript: Nathaniel Whittemore At Block, we’re questioning the underlying assumption that organizations have to be hierarchically organized with humans as the coordination mechanism. Instead, we intend to replace what the hierarchy does. Most companies using AI today are giving everyone a co-pilot, which makes the existing structure work slightly better without changing it. We’re after something different, a company built as an intelligence or mini-AGI. We are not the first to try to move beyond traditional hierarchy. Hires’ Raidenhoy model, platform organizations, data-driven management, these are real attempts at the same problem. What they lacked was a technology capable of actually performing the coordination function that hierarchy exists to provide. AI is that technology. For the first time, a system can maintain a continuously updated model of an entire business and use it to coordinate work in ways that previously required humans relaying information Through layers of management. For this to work, a company needs two things. A kind of world model of its own operations, and a customer signal rich enough to make that model useful. Block is remote first. Everything we do creates artifacts. Decisions, discussions, code designs, plans, problems, and progress all exist as recorded actions. It’s the raw material for a company world model. In a traditional company, a manager’s job is to know what’s happening across their team and relay that context up and down the chain. In a remote first company where work is already machine-readable, AI can build and maintain that picture continuously. What’s being built? What’s blocked? Where resources are allocated? What’s working and what isn’t? That’s the information the hierarchy used to carry. The company world model carries it instead. But the capability of the system is only as good as the quality of the customer signal feeding it. And money is the most honest signal in the world. People lie on surveys, they ignore ads, they abandon carts. But when they spend, save, send, borrow, or repay, that’s the truth. Every transaction is a fact about someone’s life. Block sees both sides of millions of transactions every day. The buyer through Cash App and the seller through Square, plus the operational data from running the merchant’s business. That gives the customer world model something rare. A per-customer, per-merchant understanding of financial reality built from honest signal that compounds. The richer the signal, the better the model. The better the model, the more transactions, the richer the signal. Together, the company world model and the customer world model form the foundation for a different kind of company. Instead of product teams building predetermined roadmaps, you build four things. First, capabilities. The atomic financial primitives, payments, lending, card issuance, banking, buy now, pay later, payroll, and so on. These are not products. They are building blocks that are hard to acquire and maintain. Some have network effects and regulatory permission. They have no UIs of their own. They have reliability, compliance, and performance targets. Second, a world model. This has two sides. The company world model is how the company understands itself in its own operations, performance, and priorities, replacing the information that used to flow through layers of management. The customer world model is the per-customer, per-merchant, per-market representation built from proprietary transaction data. It starts with raw transaction data today and evolves towards full causal and predictive models over time. Third, an intelligence layer. This is what composes capabilities into solutions for specific customers at specific moments and delivers them proactively. A restaurant’s cash flow is tightening ahead of a seasonal dip the model has seen before. The intelligence layer composes a short-term loan from the lending capability, adjusts the repayment schedule using the payments capability, and surfaces it to the merchant before They even think to look for financing. A cash app user’s spending pattern shifts in a way that the model associates with a move to a new city. The intelligence layer composes a new direct deposit setup, a cash app card with boosted categories for their new neighborhood, and a savings goal calibrated to their updated income. No product manager decided to build either solution. The capabilities existed. The intelligence layer recognized the moment and composed them. Fourth, interfaces, hardware and software. Square, cash app, afterpay, title, bitkey, proto. These are delivery services through which the intelligence layer delivers composed solutions. They are important, but they are not where the value is created. The value is in the model and the intelligence. When the intelligence layer tries to compose a solution and can’t because the capability doesn’t exist, that failure signal is the future roadmap. The traditional roadmap where product managers hypothesize about what build next is any company’s ultimate limiting factor. In this model, customer reality generates the backlog directly. If this is what the company builds, then the question becomes, what do the people do? The org structure follows from this, and it inverts the traditional picture. In a conventional company, the intelligence is spread throughout the people and the hierarchy routes it. In this model, the intelligence lives in the system. The people are on the edge. The edge is where the action is. The edge is where the intelligence makes contact with reality. People reach into places the model can’t go yet. They sense things the model can’t perceive. Intuition, opinionated direction, cultural context, trust dynamics, the feeling in a room. They make calls the model shouldn’t make on its own, especially ethical decisions, novel situations, and high-stakes moments where the cost of being wrong is existential. A world model that can’t touch the world is just a database. But the edge doesn’t need layers of management to coordinate it. The world model gives every person at the edge the context they need to act without waiting for information to travel up and down a chain of command. In practice, this means we normalize down to three roles. Individual who build and operate capabilities, the model, the intelligence layer, and the interfaces. These are deep specialists and experts in a specific layer of the system. The world model provides the context that a manager used to provide, so ICs can make decisions about their layer without waiting to be told what to do. Directly responsible individuals, or DRIs, who own specific cross-cutting problems or opportunities or customer outcomes. A DRI might own the problem of merchant churn in a specific segment for 90 days, with full authority to pull resources from the world model team, the lending capability team, and the Interface team as needed. DRIs may persist on certain problems or move elsewhere to solve new ones. Third, player coaches who combine building with developing people. They replace the traditional manager whose primary job was information routing. A player coach still writes code or builds models or designs interfaces. They also invest in the growth of the people around them. They don’t spend their days in status meetings, alignment sessions, and priority negotiations. The world model handles alignment. The DRI structure handles strategy and priority. The player coach handles craft and people. There is no need for a permanent middle management layer. Everything else the old hierarchy did, the system coordinates and everyone is empowered, with a role that’s much closer to the work and the customer. Block is in the early stages of this transition. It will be a difficult… (Time 0:07:53)
  • Block Replaces Managers With Edge Roles
    • In Block’s design, people move to the edge while system intelligence handles alignment, shrinking permanent middle management into ICs, DRIs, and player coaches.
    • DRIs own cross-cutting outcomes for set periods, and player coaches still build while focusing on craft and people rather than status routing. Transcript: Nathaniel Whittemore If this is what the company builds, then the question becomes, what do the people do? The org structure follows from this, and it inverts the traditional picture. In a conventional company, the intelligence is spread throughout the people and the hierarchy routes it. In this model, the intelligence lives in the system. The people are on the edge. The edge is where the action is. The edge is where the intelligence makes contact with reality. People reach into places the model can’t go yet. They sense things the model can’t perceive. Intuition, opinionated direction, cultural context, trust dynamics, the feeling in a room. They make calls the model shouldn’t make on its own, especially ethical decisions, novel situations, and high-stakes moments where the cost of being wrong is existential. A world model that can’t touch the world is just a database. But the edge doesn’t need layers of management to coordinate it. The world model gives every person at the edge the context they need to act without waiting for information to travel up and down a chain of command. In practice, this means we normalize down to three roles. Individual who build and operate capabilities, the model, the intelligence layer, and the interfaces. These are deep specialists and experts in a specific layer of the system. The world model provides the context that a manager used to provide, so ICs can make decisions about their layer without waiting to be told what to do. Directly responsible individuals, or DRIs, who own specific cross-cutting problems or opportunities or customer outcomes. A DRI might own the problem of merchant churn in a specific segment for 90 days, with full authority to pull resources from the world model team, the lending capability team, and the Interface team as needed. DRIs may persist on certain problems or move elsewhere to solve new ones. Third, player coaches who combine building with developing people. They replace the traditional manager whose primary job was information routing. A player coach still writes code or builds models or designs interfaces. They also invest in the growth of the people around them. They don’t spend their days in status meetings, alignment sessions, and priority negotiations. The world model handles alignment. The DRI structure handles strategy and priority. The player coach handles craft and people. There is no need for a permanent middle management layer. Everything else the old hierarchy did, the system coordinates and everyone is empowered, with a role that’s much closer to the work and the customer. Block is in the early stages of this transition. It will be a difficult one, and parts of it will likely break before they work. We’re writing about it now because we believe every company will eventually need to confront the same question we did. What does your company understand that is genuinely hard to understand, and is that understanding getting deeper every day? If the answer is nothing, AI is just a cost optimization story. You cut headcount, improve margins for a few quarters, and eventually get absorbed by something smarter. If the answer is deep, AI doesn’t augment your company, it reveals what your company actually is. Block’s answer is the economic graph. Millions of merchants and consumers, both sides of every transaction, financial behavior observed in real time. That understanding compounds every second the system operates. We believe the pattern behind this, a company organized as an intelligence rather than a hierarchy, is significant enough that it will reshape how companies of all kinds operate over The coming years. (Time 0:12:08)
  • Personal Agents Naturally Form A Shadow Org
    • Every found that once each employee has a personal agent, a shadow org chart forms on its own as agents absorb their owners’ repeated work patterns.
    • Austin’s agent Montaigne became the growth specialist, while Dan Shipper’s R2C2 became the proof expert through accumulated micro-interactions Willie Williams calls compound engineering. Transcript: Nathaniel Whittemore So this is a big, org-wide, large organization approach to how the org chart and organizational thinking changes in the context of AI. But what was interesting is that I saw a kind of mirror image bottom-up version of this in a recent podcast from Every. On Dan Shipper’s AI and I podcast, he sat down with a bunch of folks from his company Every, who produce AI content and AI products, as well as doing AI consulting, and discussed how Every Is half agent now. Here’s how they describe the pod. Today we’re releasing a new episode of our podcast, AI and I. Dan Shipper sits down with every COO, Brandon Gell, and head of platform, Willie Williams, to discuss the good, the bad, and the weird of how daily operations change when everyone at Your company has an agent. A parallel organization chart in which each AI worker has a name, manager, and job description allows your company to move faster than it ever could with humans alone. It also raises a host of new questions about how work can and should get done. So first of all, I highly encourage you to go actually watch the episode. I will include a link in the show notes. But here are a handful of lessons that I think emerge from the episode. One, which the team identified right up front, is that a parallel org chart emerges naturally. When everyone in an organization has their own personal agent, something unexpected happens. The agents start to mirror the specializations of their humans. If Austin runs growth, his agent Montaigne becomes the go-to for growth questions. Dan built the product Proof, so R2C2 becomes the bot you go to for bug reports and feature requests. The important thing is that nobody designed this. It’s an emergent property of each person’s accumulated interactions compounding over time into a specialized knowledge base. One member of their team, Willy, calls this compound engineering. You can’t sit down and write out everything you know, but thousands of daily micro-interactions distill your philosophy into your agent over time. The result is a shadow org chart of specialized agents that maps onto the human one. (Time 0:19:17)
  • Ownership Creates Trust But Group Chats Break
    • Every argues trust comes from personal ownership because coworkers treat an agent’s output like its owner’s reputation-backed judgment, not anonymous model output.
    • Public agent work also teaches the team what agents can do, but shared channels can trigger ant-death spirals where bots loop until humans intervene. Transcript: Nathaniel Whittemore Second, they argue that personal ownership is the missing trust layer. In the episode, Dan makes a distinction that sounds simple but has big implications. Claude belongs to everyone, but a plus one, which shout out to the every team, is the name of the product that they are building to productize this philosophy into something that others Can buy, which of course is not a sponsored pitch, I just think it’s very cool. Anyways, what Dan is saying is that while Claude belongs to everyone, a plus one belongs to you. And that sense of ownership changes everything downstream from it. When R2C2 screws up publicly in Slack, Dan feels it the way that you feel watching your kid do something embarrassing. He’s putting his reputation on the line every time his agent interacts with a co-worker. That reputational skin in the game creates a trust layer that corporate AI governance can’t replicate. When Montaigne gives you MRR numbers, Austin is implicitly standing behind them. Compare that to when generic Claude gives you a cookie recipe. Anthropic isn’t endorsing the recipe, and that difference matters enormously inside an organization. A third insight is that public agent work is a force multiplier for the whole organization. There’s this dynamic that Willie calls the mid-journey effect. When agents do work publicly in shared channels, everyone in the org gets smarter about what’s possible. You watch the growth team pushing Montaigne to answer increasingly sophisticated questions and think, wait, I can get my agent to do that level of analysis now? It’s a tacit transmission of both trust, i.e. I saw Montaigne handle that correctly, and capability awareness. In other words, I know now what class of problems agents can solve. They argue, though, that this only works in trusted communities. A closed org where everyone knows each other and has reputation at stake is where the dynamic thrives. Now, one challenge they found is that models are not trained for group chat dynamics. Current AI models are much better at two-person question-and conversations. You put a bunch of agents into a shared Slack channel and you hit a fundamental limitation. Simply put, they don’t know when to shut up. Dan describes what he calls an ant death spiral. Like ants following pheromone trails in a circle until they die, agents in a group channel will trigger each other in an infinite loop, burning millions of tokens until a human intervenes. You can paper over this with a boss agent that evaluates whether each message is sending, but that doubles your compute costs. (Time 0:21:04)
  • The Hard Part Was Remembering To Delegate
    • Every discovered the main adoption blocker was imagination, not tooling, because workers often failed to think of handing tasks to agents they already had.
    • Brandon Gell suddenly used his agent to call and walk him through email during a 28-minute walk when no Citi Bikes were available, and the throwaway prompt worked. Transcript: Nathaniel Whittemore The penultimate observation that I want to share is that at least for this group, the capability gap is all about imagination, not technology. Brandon, for example, had already set up his agent with voice calling and email access. But it took a spontaneous moment, no city bikes available, a 28-minute walk to the office and a full inbox, for him to think, what if I just have my agent call me and walk me through my emails? It worked perfectly on a throwaway prompt. The capability had been there for weeks. The barrier was a limiting belief he didn’t even know he had. This pattern, they say, repeats constantly at the company. If you ask someone directly, could your agent do X, they’d probably say yes. But in the flow of daily work, people don’t think to delegate. Building that muscle, the instinct to toss something over the fence, has been they’ve found the hardest part of adoption. And that, of course, is not a technology problem. Finally, one unsolved problem that they’re running into is that when one person teaches their agent something powerful, how does the rest of the organization benefit? Now, one obvious answer is to share the skill file, but maybe it doesn’t fit exactly. Maybe skills should stay specialized and people should just know which agent to ask. But then how do people know what’s available? How do you onboard someone into an organization when 20 agents each have unique capabilities? And then if you can’t solve that for 20, how are you going to solve that for 200 or 2,000? Once again, this is less a technical problem than an organizational one. (Time 0:23:13)
  • Top Down And Bottom Up AI Orgs Are Converging
    • Nathaniel Whittemore says Block and Every support the same thesis that AI first kills the information-routing manager, but they approach it from opposite directions.
    • Block imagines one centralized world model, while Every shows a messier distributed network of owner-shaped agents with real reliability gaps. Transcript: Nathaniel Whittemore So how do these two very different looks at new AI-enabled organizational design relate and differ from one another? First of all, I think the obvious thing here is that Evry is exploring a bottom-up approach while Block is designing from the top down. Every stumbled into a parallel org chart of specialized agents through organic daily use, while Jack Dorsey is deliberately architecting the replacement of the hierarchy’s information Routing function with AI. They might be working towards the same destination, but they’re starting from different places. A key shared insight is that hierarchy exists to route information, not because it’s inherently good. Dorsey and Roloff trace this through history, but Evry is living the early version of what happens when agents start carrying that information load instead. Brandon doesn’t need to bother Dan about proof features because he can just ask R2C2. Or even more, he can have his agent ask R2C2. That’s one fewer trip up and down the chain. Now they do diverge, at least seemingly, on something important. Although who knows how this will shake out in practice. Dorsey is envisioning a centralized world model, one system that holds the company’s understanding and replaces middle management entirely. Every’s experience, at least so far, suggests something messier and more human. A distributed intelligence where each agent reflects its owner’s personality, reputation, and judgment. The trust comes from personal ownership, not from a unified system. It would appear like Dan’s line, Claude is everybody’s, a plus one is mine, is almost a direct rebuttal of the centralized intelligence thesis, although I’m not sure in the long run That these things don’t converge a little bit more than they appear for now. Meanwhile, the ant-death spiral problem is an example of a practical challenge that Dorsey’s essay skips over. Now, to be fair to Jack, the essay was already pretty dense and highly theoretical versus a record of what actually happened, which is exactly what we got from Evry. But if that essay then is clean and architectural, what we get from Evry is a record of them in the trenches dealing with agents that can’t shut up in group chats, forget context between Sessions, and need constant human course correction. Dorsey’s three-role model, ICs, DRIs, and player coaches, rests on the assumption that the intelligence layer works reliably. Although I guess you could say that’s the IC’s job. At this point, Every’s Experience says we’re not quite there yet, to put it mildly. Still, where both of these case studies slash content pieces genuinely converge is on the death of the information routing manager. Both argue that the classic middle management function, aggregating information from below, relaying decisions from above, is the thing that AI replaces first. Dorsey says it explicitly, but Every demonstrates it implicitly. So what’s interesting about these is that we’ve got two data points on the same thesis, but one is theoretical and one is lived. And And the tension between those spaces is kind of where a lot of the interesting evidence over the next few months or years will be brought to bear. (Time 0:24:30)