How will Bureaucracy Evolve in the 21st Century?
Introduction
We will never again have 19th or 20th century reductionist bureaucracies emerge from 21st companies. Scaleups are what’s next and will soon spend to streamline execution and become profitable. But, what will be what’s after next? I am proposing a 10 year plan to research the next-generation bureaucracy, which realizes unprecedented opportunity from progressive information sharing technology. My proposal orients research around three progressive research milestones, each contributing to valuable intellectual property development: scaleups, information networks, and the final organizational design.
Unlike my other work on Medium, this paper presents musings and hypotheticals of the future derived from inductive analysis of trends. The theme throughout this article is that there are unknowns. There are people further in knowing these unknowns, but, nonetheless, these things are still yet unknown. My plan describes a research path to make known these unknowns and… monetize them.
Important context for the first milestone is included in another paper I wrote of the same name, Meet in the Middle: How Mature and Agile Organizations Share a Common Need.
Organizational Research Summary
Three progressive research milestones orient the path to this organizational structure.
- Meet in the Middle: How Mature and Agile Organizations Share a Common Need Is there a way to innovate without relying solely on grit+luck? Is there a method to consistent success in scaling and diversifying innovation? Could chaos in execution be reduced such that probability of success will increase?
- Information Highways: How Networks Define Organizational Success What is the future of organizational information sharing? If managers will not handoff information, what will they do?
- Networked: Reseaucracy, an Organizational Design to Replace Bureaucracy (From the French word réseau, meaning network.) The final what’s after next design for the 21st century, the fruit of a decade of intentional thought.
Organizational Research Plan
This stems from a logical conclusion: if scaleups scale, they lose their agility. No one yet offers an alternative, but my hypothesis is that technology in the next 10 years will create the necessary information highways to replace the wasteful information handoffs inherent to bureaucracies. New management roles will allow organizations to connect and scale differently.
Three progressive research milestones orient the path to this model: scaleups, information networks, and the final organizational design.
Meet in the Middle: How Mature and Agile Organizations Share a Common Need[1]
We will never again have 19th or 20th century reductionist bureaucracies emerge from 21st companies. Tech-enabled scaleups are what’s next. Scaleups are agile organizations growing 20% annually for at least three years. Tech scaleups are fundamentally more compatible with the complexity of the fast, information-rich operating environment of the 21st century. Whereas manufacturing and distribution of physical goods was the economic bedrock of the previous two centuries, this century will be marked by layering digital services (automated and unseen, and otherwise) over physical goods and services (e.g. the online-merge-offline economy discussed in AI Superpowers).
However, most tech scaleups are not profitable today, due to rapid cash burn directed at growth and innovation. Tesla is a rare example of a tenuously profitable scaleup. These companies enshrine chaos and entrepreneurship in core values, with many believing grit+luck (and great product) is the source of fame and fortune. Execution is disorganized and costly, and quality of life is horrendous leading to high turnover. But, investors are increasingly calling for profitability. Covid-19 prompted some scaleups, like Airbnb and Uber, to make changes leading to profitability due to disruption of prior business models in the pandemic.
Big organizations and startups share a common enemy: complexity. On a paradigm of stability vs. adaptability, both must swing inward toward a hybrid model required to compete in a fast changing, unpredictable environment. The first milestone demonstrates credibility for what’s next leadership.
In highly agile organizations, unnecessary chaos exists due to fragmentation of teams and lack of consistent understanding of tactical and operational objectives. Especially in startups, which are requiring increasing analytical skills, a major challenge is turning raw data into useful insights aligned to strategic objectives. A variety of solutions, including diversifying networks, establishing forums, and deploying an original framework can establish the missing interdependence and shared consciousness. These solutions improve execution, leading to greater consistency and profitability.
Scaleups pride themselves on their ingenuity, adaptability, and tenacity. But is there a way to innovate without relying solely on grit+luck? Is there a method to consistent success in scaling and diversifying innovation? Could chaos in execution be reduced such that profitability will increase?
In the appendix, I explain how scaleups might use the Team of Teams model to reduce chaos, increase success in scaling, and work toward profitability — without losing agility. This section also provides the operational framework D3SA “Diversify, Develop, Deliver, Support, Assess” to support the diversification of networks needed to achieve shared consciousness and interdependence.
Information Highways: How Networks Define Organizational Success
I have started examining the importance of information networks in organizations. Bureaucracies, rooted in the French words for “office” and form of rule, efficiently manage critical information hand-offs between the tactical, operational, and strategic levels of an organization. The greater the size, the more offices and levels within each office are created. System first order controls, such as specialization, optimize the type of information consumed, adapted, and shared at each level, with tactical leaders focused on today and strategic leaders looking to the future.
Team of Teams describes four characteristics of bureaucratic info highways. This hybrid model emphasizes the importance of networks, authority, objectives, and unity in turning information into action to achieve action. Horizontal, vertical, and diagonal networks are critical to achieving objectives. The model has four characteristics:
- Shared consciousness — objectives alignment
- Interdependence — awareness and relationships with other teams
- Speed — the rate at which information is acted on
- Empowered execution — delegated decision making[2]
But, digital has created and enabled numerous new forms of information highways, such adaptive business process systems, many-to-many social networking, eventually natural language analysis queries, etc. Information handoffs in this complex world, are actually less efficient and become “waste.” The McChrystal Group has sought to reduce this waste by helping large bureaucracies become “hybrids,” implying information handoffs need to be not just vertical and horizontal but also diagonal and skip-level, coupled with shared understanding of objectives and the cultivated discernment to act on information. Further, the McChrystal Group understands the importance of relational connections and the importance, i.e. empowerment, of people in organizations. Information is objectively critical, how people subjectively discover, interpret, and act on information is just as critical. (I would add, the purpose of organizations is for the prospering of people and nations, and not ends in of themselves, which is a worldview I base in the Imago Dei.)
Two trends contribute to an unknown what’s after next we will start to see take shape in 5–15 years.
First, cloud, edge, and machine learning are disrupting traditional “linear” business process systems, such as SAP. As a former quasi-“super user” of SAP transactions at Dow, I can attest that SAP does (only) what it is programmed to do. To change what it does requires extensive development resources to scope out and implement a change, especially so that it does not create conflicts with these rest of the complicated system. Today, we have machine learning-enabled business process systems such as C3.ai, with companies like 3M joining the push. These systems are narrow in scope and not readily in use… yet. But implementation is much faster (6-month cycle for a full project according to C3’s website, versus the 8 years required for a total company-wide SAP overhaul for Dow).
Second, tech companies increasingly pursue horizontal and vertical integration in delivery of digital products and services. While a digital application is customizable and distributable to the entire world instantaneously, the backend hardware required to make this happen is huge, costly, and the matter of national disputes (even the State Department has recently created a “tech diplomacy” career track). Companies such as Apple, Amazon, Microsoft, and Google have started manufacturing processor chips optimized for artificial intelligence and cloud, what was once a profitable and powerful industry led by Intel and governed by Moore’s Law. Not only does Moore’s Law no longer represent the main goal (thermal transfer, for example, is more important than size), tech companies have become larger than the historical behemoth high tech chip suppliers. These four companies rank as #2–5 of the most valuable in the world, with the anomaly Saudi Aramco taking the top slot. This hardware/software combination of value will be as the Carnegie Steel Company of the 19th century, or the Standard Oil of the 20th century; steel and oil are still critical to the world’s economic infrastructure but the industries themselves are not valued as are fast growth, high value-add industries.
So, as we look to what’s next, we find an interesting observation about the application of a hybrid model on the bureaucratic structure: for big bureaucracies, like Dow, the “hybrid model” improves agility; for agile scaleups, like SpaceX, the model improves execution by aligning and unifying disjointed, fast moving teams. I believe information networks will be the basis of what’s after next needed to surpass today’s scale constraints on agility, and scale constraints at large. What is the future of information sharing in organizations? Why? A broader analysis of nepotism, machine learning ERP systems, social clubs, idea theory, social media, sharing economy, etc., along with organizational motivations for scale, may be supportive.
Networked: Reseaucracy, an Organizational Design to Replace Bureaucracy
Scaleups are compelled to scale. There are numerous benefits; competitive dynamics, power, risk reduction, economies of scale, etc. Bureaucracy remains a relatively recent human innovation as the only means by which to achieve scale as we know it. But even at some point, we see diminishing-to-negative marginal utility of scaling, as the ability to adapt to complexity is reduced. The U.S. intelligence community’s centralized bureaucratic leader lacks influence and authority over its 17 entities, yet while countries like China and Russia engage in cybercrimes with potential terrifying consequences. The flummoxed U.S. instead dukes out privacy politics while slowly building out capabilities. China’s President Xi was mentored by a man who said “enlightened autocracy” was the future, and not liberalization and empowerment, indicating the only way to scale further is for one man to rule with absolute power.
Whereas today’s tech behemoths make incremental innovations and act as second movers for fundamental shifts, tech scaleups are (high risk) experts on the latter. One wonders, as tech scaleups push increasing offline-merge-offline world, sharing economies, and other complex network-based approaches for products only possible through digital, how can organizations retain agility even in scale? If a tech scaleup can reach profitability without first being acquired for intellectual property and smothered by its corporate parent, how will competitive dynamics change among yesterday’s bureaucratic scions such as GE, Intel, and even Google, Apple, and others? Having interviewed at Amazon twice, and having friends there, I found their culture militaristic to a fault; it is Bezos’ personal aggression and relentlessness (literally, www.relentless.com) for perpetual (r)evolution that keeps Amazon competitive. It was the same with Steve Jobs; Tim Cook has made a valuable monstrosity out of Apple, but it no longer pursues first mover innovation, rather second (or tenth) mover optimization. Without Elon Musk personally upending his companies every few months, as I experienced, to avoid the stagnation and comfort which come with stability, how will his endeavors stay agile? Uber and the other sharing economy scaleups have an unprecedented labor model, but states like California insists on regulating gig workers as bureaucratic employees. Cryptocurrencies and other ideological decentralists, like open source efforts, have created a frame for the technology necessary to distributed coworking and resource sharing, but the ideology itself prevents their efforts from ever going anywhere.
Much as operating environment complexity is defined by a network of non-linear variables, hidden and visible, my hunch is that the organizational design which will build beyond the bureaucracy is a broad network of small bureaucracies. Hence, I have named the structure “reseaucracy” (pronounced ray-zock-rah-cee) from the French word réseau meaning network.
In a reseaucracy, the traditional role of managers as specialized information facilitators is largely eliminated. Some yet-unknown systems will provide adaptive and comprehensive information facilitation, even to the extent to which many decisions will not require human input, to a scale exceeding that when computers started replacing human decision makers 75 years ago. Therefore, much as a bureaucracy, there will be offices, but the offices will serve some yet-unknown function unrelated to information transfer. Managers will still supervise human workers, but there will be fewer levels of management. Organizations might span out in a circle (possibly multi-dimensional) from a center, rather than top-down from a single, fixed executive office. Thus, the reseaucracy will be more like a cooperative network than a formalized and strict structure.
As a hypothetical visual aid — like I said, it is impossible to know now given critical technologies and events remain in the future — I have drawn an example of a reseaucracy based on the purpose of the connection points to offices centered on integration (instead of information handoffs and specialization of time horizon). This example is founded in the trend for hardware and software companies to vertically integrate technologies to support horizontal integration of products and services at each vertical level. I envision the vertical integrators as functioning as the strategic center, as this seems to be the most significant adaptation the organization might make, to build/acquire a vertical unit. These leaders might form a governance council, with collaborative power sharing, possibly with a senior member functioning like a CEO. Middle management leads small, horizontally integrated product bureaus within each vertical. Middle managers have exceptional power to make product decisions, as they are fully networked to strategic leaders and other product teams. There are few levels of hierarchy within each bureau, as product teams are fully connected with each other such that there not need be a large office overseeing all work (e.g. like the “Chief Customer Experience Officer” at an airline, or a supply chain executive overseeing every single part of supply chain from logistics to procurement). All teams are supported by diagonally integrated units, which provide necessary services, such as human resources, even though such support is not the core competency of the organization.
Information handoffs in this age appear increasingly silly, as information itself is increasingly available. Max Weber, in his hope the bureaucracy would result in the uniform and equal treatment of all, was prophetic in some respects; slowly, hierarchical class-based methods have slowly given way from formal wear to flip flops in the office, from great deference toward managers to the rise of the individual. With information readily available, the idea managers should expect and require carefully summarized and briefed information from lower levels is an anachronistic and dangerous arrogance. In a complex world, the distance between CEOs and those who do the work must be negligible. A CEO will sit in an ideal position to view the full organizational network and work throughout to cultivate interdependence and shared consciousness. The CEO should have easy and collaborative access to middle management, which, in turn, will work directly with the front line.
Although the previous is just a hypothetical, the fruit of a decade of intentional thought will result in the final what’s after next design for the 21st century.
It is possible that we will always have bureaucracies largely as they exist today. It is possible, then, that reseaucracy will be the basis of industry design, an intentional pursuit to bridge organizations with products and services coordinating, assisting, and deploying resources to help organizations adapt to complexity. For my MBA dissertation, I designed a 3rd party “sustainable, scalable financial mechanism” working among religiously motivated investors to coordinate, assist, and deploy financial capital toward world class investments without getting in the way of individual impact (a critical criteria of the religious motivation) as would happen with a centralized bureaucracy. This paper is accessible here.
APPENDIX
Chaos-Reduction Framework for Scaleups[3]
In my experience, startup culture revolves around grit and luck. Throw yourself at problems and pray your solution works. It is awfully tempting to write one more SQL query to look at the operational data just one more way, to spend a few more hours checking charts, and firing off emails trying to band-aid an overwhelming cascade of problems. Just as you are about to leave the office late at night, an email comes in from another software team and you exclaim, “What? They’re implementing a UX change tomorrow? Why is this the first I’m hearing about it?” Startups enshrine chaos to their detriment, which leads to friction in the workplace, slow or failed scaling, and delayed profitability.
The Task Force and SpaceX case studies show networks, authority, and unity in translating information to objective-aligned insights are just as important as the pace by which information turns into action. Startups typically have exceptional speed (the rate at which information is acted on) and empowered execution (delegated decision making) but lack interdependence (awareness and relationships with other teams) and shared consciousness (objectives alignment). At the core of interdependence and shared consciousness are information networks and unity.
This is rather a curiosity: “Information networks, you say? But startups have data! Tons of data!” Startups often have radical information sharing architectures. For example, at SpaceX, I could query data from almost anywhere in the company — no passwords or other security. Products might look and feel radically different in just 6 months or less with the pace of this cycle.
The freedom of information sharing is a significant competitive advantage for startups, especially tech startups fighting to unseat a big, old, analog corporation. The awareness of this advantage is evident in how tech startups are designed, deploying scrappy data analysts to facilitate rapidly revolving cycles of software development, execution, and analysis. Increasingly, startups are looking for more and more analytical capability, as evidenced in publicly viewable hiring trends and descriptions of ideal candidates. The goal seems to churn big and bigger data into faster and faster insights.
The need for analytical skills is a necessity beyond basic competition dynamics. Technologies like machine learning, previously defined in this paper, are potentially at calculating complexity than humans and useful for scaling processes in a “online-merge-offline” world.[4] Online-merge-offline is the integration of digital into physical experiences. Additionally, many-to-many[5] business models are practically ubiquitous, with startups creating digital platforms to facilitate these complex networks[6], often with asymmetric information between sharing economy users[7], and earning soaring valuations as a result.[8] Data is essential for the functioning of digital products and the continual development of features accommodating millions, if not billions, of users with different needs. This aspect is also an existential threat to the U.S.’s Silicon Valley, which has tended to roll out idealistic, simple, standardized products to global markets, to then be undercut by locally developed or Chinese tech products modified to specific experiences and aesthetics.[9]
Because of the unpredictability of the operating environment and general lack of available expertise, startups poach scrappy entrepreneur-types from big companies. There is a trend for new, functionless analyst/operator roles, which seem to have replaced all prior operations titles. This trend seems to have developed since 2017, when I went to SpaceX (I did not see these roles then). These roles, shown here, are the Strategy Manager, Strategy & Operations Manager, and Operations Manager. I have also seen Chief of Staff positions with tech companies, under either “Chief of Staff,” “Chief of Staff/Program Manager,” or “Strategy & Operations Manager.”[10]
These roles no longer just require analytical skills at the level of Excel but also database query languages (e.g., SQL) and complex visualization languages (Python, R). Many startups straight up prefer data scientists (machine learning, etc.). This makes sense — hiring a tech startup operations professional without data analysis skills is like hiring an art major as a rocket scientist.
Figure 1 Trend toward Functionless, Analytics-based Roles
The ability for startups to turn data into action in enshrined in incentivization for fast, exponential growth at the expense of profitability. Venture capital funding rounds seed, series A, series B, etc. jump in the $10s of millions for each round.[11] With a fast cash burn, funding standards incentivize startups to push products into the market, grow fast, and reach the next funding stage. (Also, higher valuation = higher personal wealth.) A lean startup begins with a small team testing a homegrown product for market fit, and a later stage startup consists of hundreds or thousands of 20somethings trying anything and everything to develop and sell more product.
However, the fast-growth, no-profitability era is slowly coming to a halt. This seems in contrast to the 2000 dotcom bubble burst, when “easy” money suddenly evaporated and numerous startups went bust.[12] In my opinion, from discussions in the Yale School of Management class “The Future of Global Finance,” taught by Jeffrey Garten, the more recent era has been enabled by low interest rates, since the advent of quantitative easing en masse by central banks post-2008, and the supremacy of the U.S. dollar. But, shocking failures and scandals, such as WeWork[13] and the disruption of Masayoshi Son’s Vision Fund[14], has cooled the moods of venture capital.[15] The Covid-19 disruptive event forced organizations such as Uber and Airbnb to pursue improved profits, with Airbnb radically adjusting its target market[16] and Uber spinning off its JUMP and autonomous driving units, along with other unprofitable markets.[17] The ability for companies like Tesla to consistently (at least, relative to other startups) reach quarterly profits, even during the COVID-19 pandemic,[18] is a signal of change to the startup world.[19]
The open questions facing startups — even as they hire more and more analytical employees, especially from prestigious professions like management consulting and investment banking — is how to retain innovative spirits, speed of execution, the ability to quickly scale and diversify innovation, while reaching profitability?
The pursuit of unicorn employees who can accomplish all of these things is seen through public job postings. Qualifications usually require 2–4 years management consulting or investment banking experience and an MBA. But, startups prefer professionals with less than 8 years total experience.
This is quite strange given:
· A minority of talent on social media, such as LinkedIn, has this experience
· This combination (top tier MBA and 2–4 year’s experience in consulting/banking) is quite difficult to achieve, especially in less than 8 years
· If professionals have this combination, they should be quite expensive (salary requirements in the Bay Area/New York), but startups may not have the cash to pay for this; startups must provide big equity packages and a sense of purpose if paying below market
Also strange is that despite the seniority requirement (3–8 year’s experience typically), roles almost exclusively come with “Senior Manager” or “Manager” titles (and often report to Vice Presidents or Senior Directors). It is possible this nomenclature is intended to highlight the importance of data and encourage young titleholders to act with the authority and awareness of the historical manager seniority.
Startups probably pursue young professionals from prestigious industries like management consulting and investment banking for their training in logical methods and strategic communications — as well as a high threshold for burnout. This is a bit of a departure from the time of my hiring at SpaceX, which sought Fortune 100 types with creativity and change management potential. One recent, creative posting for a chief of staff role even said something like, “the ideal candidate will be a burned-out management consultant or investment banker looking to move to the sunny California shores.”
There is potential for a cautionary tale in these job postings. After all, data is inherently historical — data can help predict what, how, and when, but only one at a time, and is susceptible to unpredictable events arising from minutiae deviations amid a complex network of comprehensive variables (remember the Butterfly Effect?).[20] Should all employees at tech startups also have advanced data analysis skills, let alone MBAs and hyper-specific and rare experiences? Is the ability to consume and iterate on vast sums of data so critical to every startup role, or is this skills misuse resulting in missed execution from paralysis by analysis? Does it reduce the resiliency of startups as they constantly review data collected in the near-term past to predict the future?
I would submit the answer is, it is fine; go for it. Especially for tech startups. But be careful: if a startup is hiring an army of young analysts, be wary to ensure raw data is being efficiently processed into usable insights aligned to organizational objectives. Analysts, especially in fragmented startups, may lack the awareness and ability to bridge organizational lines. There is always chaos in new efforts, but my point is the extent of chaos in startups is unnecessary.
The alumni of the Task Force leverage lessons to help big, cumbersome organizations become more agile. This resiliency and adaptability model — speed, interdependence, empowered execution, and shared consciousness — is just as applicable to already agile startups, but in this case, the result is reduced chaos. Reducing chaos improves consistency of success in scaling and diversifying innovation, and in improving profitability. Said another way, the model improves execution without losing agility.
Interdependence and shared consciousness are the two attributes startups lack that contribute significantly to the chaos. These two attributes mean employees within startups lack the networked connectivity — despite the freedom of data — to other teams and an awareness of tactical, operational, and strategic objectives. We have already discussed several ways in which the Task Force and I, at SpaceX, improved interdependence and shared consciousness. But what might be an F3EAD equivalent for startups?
In one of my Yale classes,[21] I teamed up with eight other students, under the guidance of Chris Fussell, to deliver a consulting product to a U.S. military client. The client was facing a period of unpredictability in a worldwide shift in the dynamic national competition. The client needed to innovate new solutions to meet global competition. We developed an operational framework to guide diversification of external partnerships at all levels of the organization, similar to the F3EAD operational framework used by the Task Force to integrate intelligence and operations. The framework also works well with startups (or, more specifically, scaleups). I have included a modified version in Figure 11, applied to startup innovation.
The D3SA operational information network framework might be pronounced as either “dee-three-es-eh” or “dee-zah.” The framework is useful to any manager who needs a scrappy way to coach analysts, program managers, and other tactical employees to increase their information networks. It is also useful to startup executives who realize a need to bridge large, siloed teams. Whatever the situation, the framework is a simple tool — among others provided — to resolve the chaos inherent to agile organizations.
The framework begins with understanding the organization and objective and ends with assessing if the understanding of the organization is sufficient to achieve the objective. Thus, the framework is meant to create a culture of ever-expanding information networks to reduce chaos in execution. This is the basis of interdependence and shared consciousness.
The middle sections — develop, deliver, and support — are more inherently obvious to scrappy, entrepreneurial team members and likely need less attention. Support is particularly important for startups, especially those staffed by younger professionals. Without clearly defined roles and processes, execution is a whole-of-team effort. Regardless of title and functional alignment, the offer to help is a critical responsibility across the startup.
At SpaceX, I might have applied this framework to the O&I. I initially met with 15 people involved in the redesign after taking responsibility for the program — I would certainly invite these to the first O&I. We would collectively coordinate a strategy for the redesign (such as developing the missing design release plan), and implement, with me supporting. According to the operating rhythm — for me, it was about four days — I would then assess if I have sufficient relationships to achieve the redesign objectives. Certainly, I would have heard of other engineering teams working on smaller, specialized portions of the shield, and would have invited them to the O&I. If I had used this framework, I might have shaved weeks of chaos off the initial intervention to restore relationships and overcome tribalism.
I once told an intern at Dow that he was not expected to execute, only recommend. He was a smart guy and was thinking about second and third order effects of what he might recommend and had been overwhelmed by the complication of execution. After hearing the advice, he visibly relaxed, eased into his work, and became an effective collaborator. He was one of the only interns hired for full time in a tight-budget year. Years later, the broader principle behind my advice still rings true: everything requires the effort of broad collaboration. In a big company, executing a tactical and clearly defined process might seem like an individual effort, but even this is not true. The D3SA cycle enshrines this critical cross-team participation.
Grit+luck (and a great product) may still be the critical ingredients to a startup. Startup employees will still work long hours. But with these tools, those hours will be more productive with less stress. Execution will improve, scaling and diversifying innovation will be a little more methodical and a little less guesswork, and profitability will be ever closer.
[1] Adapted from Meet in the Middle: How Mature and Agile Organizations Share a Common Need, written by Evan Crain for Professor Ted Wittenstein’s GLBL 592 Intelligence Espionage and American Foreign Policy, Fall 2020 at Yale University.
[2] Adapted from Meet in the Middle: How Stable and Agile Organizations Share a Common Need by Evan Crain
[3] Excerpt from Meet in the Middle: How Stable and Agile Organizations Share a Common Need by Evan Crain
[4] AI Superpowers, p118
[5] Team of Teams, 63
[6] a16z, “…Fastest-Growing Social Apps…” 12/7/2020
[7] Advances in Economics, Business and Management Research, volume 144, “Asymmetric Information of Sharing Economy”
[8] WSJ, “Sizzling Tech IPO Market Leaves Investors Befuddled,” 12/13/2020
[9] AI Superpowers, p33–40
[10] From LinkedIn
[11] Fundz.net. 12/12/2020
[12] Investopedia, “Dotcom Bubble,” 6/25/19
[13] WSJ “The Fall of WeWork,” 10/24/19
[14] WSJ, “Masayoshi Son Again Pulled SoftBank From the Brink,” 11/11/20
[15] Forrestor, “Predictions 2020,” 11/1/19
[16] WSJ, “How Airbnb Pulled Back From the Brink,” 10/12/20
[17] WSJ, “Uber Sells Self-Driving-Car Unit,” 12/7/20
[18] The New York Times, “Tesla Turns a Profit in a Pandemic-Squeezed Quarter,” 7/22/20
[19] WSJ, “Tech Startups Face New Investor Mandate: Profits Over Discounts,” 12/27/19
[20] Team of Teams, p72
[21] Yale University, Fall 2020, GLBL 842: Special Operations Forces: History, Context, and Future with Professor Chris Fussell
