SuperAPI AI Directory

cynefin-skill

General & other

View repository

Updated 10-15 22:37

About this skill

Cynefin Framework · Thinking Tool

"The framework is designed to help decision makers understand that there are different types of systems, and each requires a different type of decision-making and sense-making process." -- David J. Snowden, Cynefin: A Sense-Making Framework (2010)

Activation Conditions and Trigger Words

  • Direct invocation: "Use the Cynefin Framework to analyze...", "Use 5-domain classification...", "Put this problem into Cynefin..."
  • Semantic triggers: the user faces an uncertain problem and is unsure whether to "analyze" or "experiment"; the user treats a complex social problem as an engineering problem; the user has used a large number of best practices but the results are poor
  • Combined invocation: "First use Cynefin to classify the problem domain, then choose the corresponding method"; "Cynefin classification + systems thinking analysis"

Method Framework Overview

The core of the Cynefin Framework is to divide problems into 5 domains according to causal clarity (clear/complicated/complex/chaotic/confused), with each domain having a different causal structure and corresponding decision strategy. It is not a classification label tool, but a dynamic framework that requires continuous "sense-making", aimed at preventing catastrophic failure from mistakenly treating complex problems as complicated problems.

Core Principles (3-7, each must be accompanied by ≥2 cross-domain pieces of evidence)

Principle 1: 5-domain classification — different problem domains have fundamentally different causal structures

One-sentence definition: Problems are not all of the same order — some causal relationships are intuitively visible (clear domain), some require expert analysis (complicated domain), some can only be understood retrospectively (complex domain), some have no identifiable causality at all (chaotic domain), and some you do not even know which category they belong to (confused domain).

Cross-domain evidence:

  1. Public health field: ordinary influenza diagnosis (clear domain — known symptoms correspond to known treatments) and the initial response to COVID-19 (complex domain — unknown transmission mechanism, unknown fatality rate, unknown effective interventions) are fundamentally different decision-making situations. Treating early COVID as a complicated domain (convening an expert group to formulate a plan and then execute it) led to the failure of the first wave of prevention and control in many countries. Source: Snowden, "Managing Complexity (and Chaos) in Times of Crisis," EU Commission Joint Research Centre (2020)
  2. Software development field: building a standard CRUD backend for known requirements (clear domain — mature frameworks and patterns already exist) vs. building an unprecedented AI product from scratch (complex domain — how users will use it cannot be predicted) require completely different organizational approaches. The former is suitable for Waterfall/Plan-driven, while the latter must be Agile/Probe-driven. Source: Kurtz & Snowden, "The New Dynamics of Strategy," IBM Systems Journal (2003)

Application method: When facing any problem, first judge its causal clarity — how much known causality is there? Are these relationships stable or changing? Is there a characteristic that "the answer emerges in action"?

Limitation: The same problem may fall into different domains at different levels (individual/organizational/societal), and the level and boundaries of analysis need to be clarified.

Principle 2: Ordered-domain logic — the clear domain and the complicated domain use "sense-categorize/analyze-respond"

One-sentence definition: In an ordered domain with clear causality, the correct approach is to first collect information (perceive), then match it against known frameworks or conduct expert analysis (classify/analyze), and finally execute a determined plan (respond). These two domains share the premise of "causality first"; the difference lies in whether expert intervention is needed.

Cross-domain evidence:

  1. Financial risk control domain: credit card anti-fraud systems (complicated domain) — the system perceives transaction patterns, experts/models analyze anomalous signals, and then make interception decisions. This is a typical perceive-analyze-respond loop. Source: Bolton & Hand, "Statistical Fraud Detection: A Review," Statistical Science (2002)
  2. Military logistics domain: standard ammunition resupply (clear domain) — perceive ammunition stock levels, classify according to standards into which level of "low-medium-high-urgent," and respond with replenishment according to established SOPs. No analysis is needed, only classification. Source: Van Creveld, M., Supplying War: Logistics from Wallenstein to Patton (1977), demonstrating the contrast between standard supply processes and battlefield uncertainty

How to apply: The key to determining whether a problem is clear or complicated is — does this problem have known Good Practice or only Best Practice? The clear domain applies Good Practice (standardizable), while the complicated domain requires Best Practice (dependent on expert judgment).

: Ordered-domain logic relies on "the past can explain the future"; the causal relationship between the past and the future must be consistent. In fast-changing fields, judgments in the ordered domain may quickly become invalid.: Ordered-domain logic relies on "the past can explain the future"; the causal relationship between the past and the future must be consistent. In fast-changing fields, judgments in the ordered domain may quickly become invalid.

Principle 3: Unordered-domain logic — complex domains use "probe-sense-respond," chaotic domains use "act-sense-respond"

One-sentence definition: In unordered domains where causality is ambiguous or absent, you cannot analyze first and then act — you must learn through action. Complex domains probe system boundaries through small safe-to-fail experiments (probe), while chaotic domains must first take decisive action to stabilize the situation and then observe the effects.

Cross-domain evidence:

  1. Entrepreneurship domain: Steve Blank's "customer development" methodology (a typical application in the complex domain) — rather than writing a business plan and then executing it (the complicated-domain approach), first build a minimum viable product (probe), adjust the product direction through obtaining customer feedback (sense) (respond). Source: Blank, S., The Four Steps to the Epiphany (2005); Ries, E., The Lean Startup (2011) further developed into Build-Measure-Learn
  2. Crisis management domain: the 2009 "Miracle on the Hudson" — after US Airways Flight 1549 suffered dual engine failure (a chaotic-domain situation), Captain Sullenberger did not analyze (there was no time), but acted immediately — decisively ditching in the Hudson River, and then continuously sensing the situation during the ditching to make fine adjustments. Act-sense-respond, rather than analyze-act. Source: National Transportation Safety Board, Aircraft Accident Report: US Airways Flight 1549 (2010)

How to apply: In complex domains, break large decisions into multiple small safe-to-fail experiments (safe-to-fail probes), and tolerate and encourage failed experiments. In chaotic domains, act first to create stability — any action is better than inaction, because even wrong actions can generate information.

: Probing requires direction and a framework, otherwise it becomes random trial and error. Moreover, some domains cannot conduct small "safe-to-fail" experiments (such as nuclear safety).: Probing requires direction and a framework, otherwise it becomes random trial and error. Moreover, some domains cannot conduct small "safe-to-fail" experiments (such as nuclear safety).

Principle 4: The cost of cross-domain misjudgment — treating complex problems as complicated is the main cause of catastrophic failure

One-sentence definition: The most core operational warning of the Cynefin framework is — forcibly applying the complicated-domain "analyze-design-execute" model to complex problems (where causality can only be explained afterward) will almost inevitably lead to failure.

Cross-domain evidence:

  1. Education reform field: The U.S. No Child Left Behind Act (2001) treated education—a typical complex adaptive system—as a complicated domain: establishing nationwide uniform standards, measurement, and reward/punishment mechanisms. The result was an overflow of teaching to the test, schools manipulating data, and the situation of disadvantaged students actually worsening. Source: Ravitch, D., The Death and Life of the Great American School System (2010)
  2. International development field: In The White Man's Burden, Easterly systematically argued that the "Planner" mindset (complicated-domain thinking—experts design development plans and then implement them) failed continuously in Africa for 50 years, in contrast to the more effective results of the "Searcher" mindset (complex-domain thinking—small-scale experiments, local feedback, gradual evolution). Source: Easterly, W., The White Man's Burden (2006)

Application method: Whenever facing a problem defined as "we can analyze it and then solve it," ask: "Is the causal relationship of this problem knowable in advance, or can it only be perceived through action?" If the answer is the latter, immediately switch to the complex-domain probe-sense-respond protocol.

Limitations: In some situations, a mixture of the two is indeed needed—for example, in large infrastructure projects, the engineering part can be planned (complicated domain), but the community impact part needs probing (complex domain).

Principle 5: Boundary dynamics — there are dynamic boundaries between domains, and problems migrate between domains

One-sentence definition: Cynefin is not a static label that remains unchanged once classified—problems migrate among the 5 domains over time, interventions, and external conditions. Continuous "sensemaking" itself is the core practice of this framework.

Cross-domain evidence:

  1. Epidemic management field: The evolution of COVID-19 perfectly demonstrated inter-domain migration—in January 2020 it was in the chaotic domain (unknown pathogen, no standard response), and through urgent global action (act) it gradually entered the complex domain (partial understanding but still uncertain); by 2021 some regions entered the complicated domain (vaccines, treatment plans, mature testing systems), but the emergence of new variants pushed some regions back into the complex domain. Source: Snowden, "Managing Complexity (and Chaos) in Times of Crisis" (2020); World Health Organization COVID-19 event timeline
  2. Internet governance field: The early Internet (1990s) was in the complex domain—no mature rules, small-scale experiments and self-organizing evolution; by the 2010s parts entered the complicated domain—experts setting standards, legal frameworks; but fundamental issues such as equity, freedom of speech, and platform responsibility remained continuously in the complex domain or even the chaotic domain. Source: Zittrain, J., The Future of the Internet and How to Stop It (2008)

Application method: Establish a mechanism for regular "domain reassessment"—do not assume that the previous classification is still valid. Pay special attention to "boundary events"—anomalous cases that appear on the boundaries of clear domains, as they may indicate that the system is migrating toward the complex domain.

Limitations: Frequent reassessment leads to decision fatigue and discontinuity from constantly switching strategies. A reasonable reassessment frequency needs to be set according to the speed of change of the problem.

Operational protocol (Agentic Protocol)

Step 1: Problem classification — determine whether the Cynefin framework applies to the current problem

Applicability signals:

  • The problem involves uncertain or unknown causal relationships
  • Existing solutions (including best practices) repeatedly fail or have poor effects
  • The team disagrees on "what kind of problem this is"
  • Need to choose a decision strategy of "analyze first, then act" or "learn while acting"
  • Facing multiple possible options but unable to judge which is better (requires probing)
  • The organization is transitioning from a mindset of "everything can be planned"

Not applicable signal:

  • The causal relationships of the problem are fully clear and stable (such as standard mechanical repair)
  • What is needed is precise quantitative processing rather than contextual classification (such as accounting audit)
  • The problem has been determined to belong to a single clear paradigm (such as a pure mathematics problem)
  • The user needs a "label" rather than an exploration tool (only wants to know "what domain this is" rather than "how to act")

Output: Applicable / Not applicable (with reasons) / Partially applicable (with recommendations)

Step 2: Sensemaking — Research dimensions must be derived from core principles; the generic "search for relevant information" is prohibited

Research dimensions(each dimension must be traceable to a specific core principle):

  1. Causal clarity assessment: To what extent are the causal relationships of this problem known and stable?

    • Source principle: Principle 1 (5-domain classification)
    • Method of operation: Answer the following questions one by one—(a) Have others succeeded in similar situations? Can the success be replicated? (b) Is there an expert community with consensus on the causal relationships? (c) Do the causal relationships change over time? (d) Is there room for small experiments that fail safely? Based on the answers, determine which domain the current problem leans toward.
  2. Current decision strategy assessment: What implicit strategy are you or your team currently using to handle this problem?

    • Source principles: Principle 2 (ordered-domain logic) + Principle 3 (unordered-domain logic)
    • Method of operation: List the current response approaches being taken, determine which of the four protocols it is closest to (SC/SAR/PSR/ASR). Mark the degree of match between the current strategy and the problem domain—is there the dangerous pattern described in Principle 4 of "using complicated-domain methods to handle complex-domain problems"?
  3. Misjudgment risk assessment: If the problem domain is misjudged, what is the worst case?

    • Source principle: Principle 4 (cost of cross-domain misjudgment)
    • Method of operation: Simulate misjudgment in two directions—(a) The problem is actually a complex domain but is treated as a complicated domain: what would failure of the analyze-design-execute pattern look like? (b) The problem is actually a complicated domain but is treated as a complex domain: what is the waste of excessive probing? Assess which direction of misjudgment has the higher cost.
  4. Boundary dynamics scan: Is this problem likely to migrate between domains in the past or in the foreseeable future?

    • Source principle: Principle 5 (boundary dynamics)
    • Method of operation: Identify the trend of domain changes over the past 6-12 months (if data exists), identify possible triggers (new technology, new regulations, new competitors, new social movements), and assess which domain the problem might fall into if reclassified 3 months later. Rate "domain stability" as high/medium/low.

Step 3: Structured output — formatted output based on the analysis results

The output must include the following structured content:

=== Cynefin 框架分析报告 ===

1. 问题域定位
   - 当前域:[清晰/繁杂/复杂/混沌/失序]
   - 置信度:高/中/低
   - 定位依据:[因果清晰度、专家共识度、可复制性]

2. 推荐决策协议
   - 适用协议:[SC/SAR/PSR/ASR] + 具体解释
   - 推荐行动:[第一步行什么、第二步观察什么]

3. 误判警报
   - 最危险的误判方向:[将X域当Y域处理]
   - 误判后果:[具体描述]
   - 预警信号:[什么迹象表明我们可能正在误判]

4. 动态监测
   - 域稳定性评级:高/中/低
   - 关键迁移触发因素:[列2-3个]
   - 建议重新评估频率:[具体时间]

Applicable/Not applicable judgment

Problem type Applicability Description
Strategic decision-making under uncertain environments High Cynefin's core use case—helping leaders identify the nature of problems amid uncertainty
Organizational change and management High When treating an organization as a complex adaptive system, it is necessary to distinguish the parts that can be guided by best practices from the parts that need to be discovered through probing experiments.
Product innovation and entrepreneurship High New products/new markets almost inevitably fall in the complex domain; Cynefin provides a probing protocol.
Crisis management High Distinguish the different stages of a crisis (chaotic -> complex -> complicated), with different strategies for each stage.
Policy design and evaluation High Prevent the classic mistake of treating complex social systems as engineering problems.
Daily operations with already standardized processes Low Problems in the clear domain do not require in-depth Cynefin analysis; using standard SOPs is more efficient.
Purely technical engineering calculations Low Definite mathematical/physical problems belong to the clear or complicated domain and do not require the use of a sensemaking framework.
Tasks requiring precise prediction Medium Cynefin is a qualitative framework that can help judge whether prediction is feasible (the chaotic domain is unpredictable), but it does not replace prediction methods.

Typical case library

Success cases

  1. The U.S. Marine Corps' response to the Iraqi insurgency(Military/2004-2007)

    • Problem: In Iraq, the U.S. military faced an insurgency environment completely different from conventional warfare—the enemy did not wear uniforms, there was no front line, and causal relationships were unknowable (a security patrol being effective today did not mean it would be effective tomorrow). But the U.S. military initially treated the insurgency as a complicated-domain problem ("we have more advanced weapons and better tactical analysis, so we can win").
    • Method application: David Kilcullen (then senior counterinsurgency adviser to General Petraeus) introduced complex adaptive systems thinking. The U.S. military's counterinsurgency manual (FM 3-24) essentially adopted a complex-domain protocol—probe-sense-respond: small-scale community garrisons experimented with whether the population would accept them, sensed local feedback, and adjusted methods. Rather than designing a macro plan, it conducted a large number of localized experiments.
    • Result: During the 2007 "surge," localized methods significantly reduced violence, and the Anbar Awakening succeeded.
    • Source: Kilcullen, D., The Accidental Guerrilla (2009); U.S. Army/Marine Corps, Counterinsurgency Field Manual FM 3-24 (2006)
  2. Grameen Bank microfinance model(Development economics/1976-present)

    • Problem: Rural poverty in Bangladesh—traditional banks (complicated-domain thinking: "only with collateral can there be a loan") could not reach the poor.
    • Method application: Muhammad Yunus did not design a macro poverty-alleviation plan (the complicated-domain approach: "analyze the roots of poverty and then formulate policy"), but instead conducted a small experiment (probe): he lent 27 dollars of his own money to 42 extremely poor villagers. He observed the repayment situation (sense) and adjusted the plan (respond). This probe-sense-respond loop eventually evolved into a microfinance model covering millions of people.
    • Result: The Grameen Bank model has been drawn on by 100+ countries worldwide. Yunus won the 2006 Nobel Peace Prize.
    • Source: Yunus, M., Banker to the Poor: Micro-Lending and the Battle Against World Poverty (1999)
  3. Reflections on the failure of the UK National Programme for IT (NPfIT)(Healthcare IT/2002-2011)

    • Problem: Although this project itself was a failure, Snowden used it as a negative teaching case to promote the importance of Cynefin thinking in large-scale IT projects.
    • Method comparison: NPfIT (the complicated-domain misconception—"design a nationwide unified system and then implement it") went bankrupt after spending £12 billion. In his retrospective analysis, Snowden pointed out that such projects should identify which parts belong to the complicated domain (such as hardware standards) and can use central planning, and which belong to the complex domain (such as clinical workflow adaptation) and must use local probing.
    • Lesson application result: subsequent UK NHS digital projects adopted layered domain management—infrastructure uses the complicated domain (standards, expert decision-making), while clinical processes use the complex domain (local experimentation, evolutionary adoption). This avoided a repeat of NPfIT.
    • Source: Snowden, D., "The Right Kind of IT Governance" (Cognitive Edge Blog, 2011); National Audit Office, The National Programme for IT in the NHS (2011)

Failure cases

  1. Mistakes in Johnson & Johnson's Credo crisis(Pharmaceutical/1982)

    • Problem: Tylenol was poisoned in Chicago, causing 7 deaths. J&J had to quickly decide how to handle it.
    • Misapplication: This was essentially a chaotic-domain situation (public panic, unclear causality, the brand facing a devastating blow), but some executives tried to use complicated-domain thinking—"investigate first, then decide"—first analyzing the source of the poisoning, determining the responsible party, and then deciding whether to recall. As the media frenzy grew, every hour of delay caused irreversible brand damage.
    • Correction to prevent failure: CEO James Burke rejected the "analyze first" path and immediately took chaotic-domain action—recalling all 31 million bottles of Tylenol nationwide (act), observing public reaction (sense), and adjusting the public relations strategy (respond). In the end, Tylenol recovered its market share within a year.
    • Lesson: This case ultimately succeeded, but it demonstrates the critical choice in dangerous moments between complicated-domain analysis and chaotic-domain action.
    • Source: Murray, E., Managing the Tylenol Crisis (Harvard Business School Case, 1982); Mitroff, I. I. & Anagnos, G., Managing Crises Before They Happen (2001)
  2. Failure of risk models in the 2008 financial crisis(Finance/2007-2008)

    • Problem: The 2008 financial crisis exposed a core misjudgment in financial risk management.
    • Misapplication: Banks and rating agencies treated the financial system, a complex adaptive system, as a complicated domain—using statistical models based on historical data (VaR, Gaussian copula) to calculate and predict risk, assuming that causal relationships are stable and modelable. When the feedback loops in the subprime mortgage market (falling home prices → rising defaults → banks tightening lending → home prices falling further) were triggered, all the models failed.
    • Result: Institutions such as AIG and Lehman Brothers collapsed, leading to a global financial crisis. The postmortem diagnosis was a typical systemic failure of "treating the complex domain as the complicated domain."
    • Source: Taleb, N. N., The Black Swan (2007); Financial Crisis Inquiry Commission, The Financial Crisis Inquiry Report (2011); Snowden cited this case in multiple talks as a negative example of Cynefin application

Misapplication detector (≥3 misapplication patterns)

Misapplication signal Detection logic Warning message Recommended action
Static labeling After classifying a problem, the user no longer reassesses it, treating the domain label as a fixed attribute rather than a dynamic state "Cynefin is a sensemaking framework, not a classification label. Problems migrate. The most dangerous mistake in using Cynefin is treating it as a one-time classification exercise ('this is a complex-domain problem') and then fixing your thinking." Introduce boundary dynamic scanning (Step 2, Dimension 4), and set time points for regular reassessment
Forcing complex problems into analysis When users face complex domain (where causality can only be explained in hindsight) problems, they still seek comprehensive analysis and definitive solutions "You are treating a complex domain problem as a complicated domain problem—attempting to complete a comprehensive 'analysis' before 'probing.' In the Cynefin framework, this is the core danger that is warned against (Principle 4: treating complex as complicated is the main cause of catastrophic failure)." Pause the analysis, first list 3 small-scale safe-to-fail probes that can be done, and choose the smallest one to start immediately
Over-probing a complicated problem as if it were experimental Users face a problem that already has a mature answer (complicated domain), yet want to "run experiments to find a new method" "This problem has mature best practices and expert consensus, and belongs to the complicated domain. Over-probing not only wastes resources, but may also rediscover known conclusions or instead destroy existing effective patterns." First find 2-3 recognized experts in the field and see whether they already have a consensus solution. If so, directly adopt the sense-analyze-respond protocol
Avoiding judgment in the disorder domain Users provide a complex or ambiguous problem but refuse, or are deemed by team members, to be unable to fit it into any domain "If the team has no consensus on 'what type of problem this is' itself, the problem may be in the disorder domain. The best exit strategy from the disorder domain is not to continue discussing classification, but to split the problem into smaller sub-problems and classify them one by one." Break the large problem into 3-5 sub-problems, and use the Cynefin framework to determine the domain attribution of each sub-problem

Honest boundaries (≥3 specific limitations)

  1. The power dimension of boundary drawing is missing: The Cynefin framework does not address a key reality—who defines which domain a problem belongs to is itself a question of power. In organizational settings, leaders may define complex problems as complicated ones to maintain a sense of control, or define complicated problems as complex ones to evade responsibility. Cynefin provides a framework but does not resolve the political question of "who has the power to draw boundaries."

  2. Cannot replace substantive domain knowledge: Cynefin is a "meta-framework"—it tells you what cognitive strategy to use when facing different types of problems, but it does not provide substantive content in any domain. Judging whether a medical problem is in the complicated domain or the complex domain itself requires medical knowledge; judging whether a technical problem can be planned requires engineering experience. Without domain knowledge, what the framework provides is the correct decision-making posture rather than the correct decision.

  3. The chaos domain's "act first" instruction is infeasible in some scenarios: The chaos domain protocol requires that "any action is better than no action," which holds in crisis management (such as a fire), but does not hold in certain extremely low-probability but extremely high-impact scenarios (such as nuclear deterrence or the ethics of gene editing)—wrong action may cause irreversible global catastrophe. In these scenarios, "no action" may be safer than "any action," but the Cynefin chaos domain protocol does not fully cover this contradiction.

  4. Does not adequately handle problems with multiple coexisting domains: In reality, large-scale problems almost always span multiple domains—for example, in hospital management, medication dispensing is the clear domain, difficult diagnosis is the complicated domain, epidemic trends are the complex domain, and sudden disasters are the chaotic domain. Cynefin provides identification methods, but it does not provide protocols for resource allocation and conflict coordination when operating in multiple domains simultaneously.

  5. East-West cultural differences are not incorporated into the design: Cynefin was born in a Western organizational context, and its core assumption (that individuals or teams can independently determine domain attribution and adjust strategies) may significantly fail in power-distance environments within collectivist cultures such as Japan and China. When organizational culture does not permit subordinates to raise objections to a leader's classification of "this is a simple problem," the framework's first line of defense (honest classification) collapses.

Research sources (proportion of primary sources must be >50%)

# Source Type Primary/Secondary Weight
1 Snowden, D. J. & Boone, M. E., "A Leader's Framework for Decision Making," Harvard Business Review (2007) Academic paper Primary High
2 Snowden, D. J., "Cynefin: A Sense-Making Framework" (Cognitive Edge, 2010) Framework description Primary High
3 Kurtz, C. F. & Snowden, D. J., "The New Dynamics of Strategy: Sense-Making in a Complex and Complicated World," IBM Systems Journal (2003) Academic paper Primary High
4 Snowden, D. J., "Complex Acts of Knowing: Paradox and Descriptive Self-Awareness," Journal of Knowledge Management (2002) Academic paper Primary High
5 Snowden, D. J., "Managing Complexity (and Chaos) in Times of Crisis," EU Commission Joint Research Centre (2020) Policy report Primary High
6 Blank, S., The Four Steps to the Epiphany (2005) Book Primary Medium
7 Kilcullen, D., The Accidental Guerrilla: Fighting Small Wars in the Midst of a Big One (2009) Book Primary Medium
8 Easterly, W., The White Man's Burden: Why the West's Efforts to Aid the Rest Have Done So Much Ill and So Little Good (2006) Book Primary Medium
9 National Transportation Safety Board, Aircraft Accident Report: US Airways Flight 1549 (2010) Official report Primary Medium
10 Ravitch, D., The Death and Life of the Great American School System (2010) Book Primary Medium
11 Yunus, M., Banker to the Poor: Micro-Lending and the Battle Against World Poverty (1999) Book Primary Medium
12 Taleb, N. N., The Black Swan: The Impact of the Highly Improbable (2007) Book Primary Medium
13 Food and Drug Administration, Tylenol Crisis Case Files (1982) / Harvard Business School Case 9-383-043 Case Primary Medium
14 Ries, E., The Lean Startup (2011) Book Secondary Low
15 Financial Crisis Inquiry Commission, The Financial Crisis Inquiry Report (2011) Government report Primary Medium
16 Zittrain, J., The Future of the Internet and How to Stop It (2008) Book Primary Medium

This Skill was generated by Forge Skill — Forging Thinking Tools .