A daily bite-size selection of top business content.

PM edition. Issue number 1248

Latest 10 stories. Click the button for more.

Term: AI harness

"A harness (often called an agent harness or agentic harness) is an external software framework that wraps around a Large Language Model (LLM) to make it functional, durable, and capable of taking actions in the real world." - AI harness

An AI harness is the external software framework that wraps around a Large Language Model (LLM) to extend its capabilities beyond text generation, enabling it to function as a persistent, tool-using agent capable of taking real-world actions. Without a harness, an LLM operates in isolation-processing a single prompt and generating a response with no memory of previous interactions and no ability to interact with external systems. The harness solves this fundamental limitation by providing the infrastructure necessary for autonomous, multi-step reasoning and execution.

Core Functions and Architecture

An AI harness performs several critical functions that transform a static language model into a dynamic agent. Memory management addresses one of the most significant constraints of raw LLMs: their fixed context windows and lack of persistent memory. Standard language models begin each session with no recollection of previous interactions, forcing them to operate without historical context. The harness implements memory systems-including persistent context logs, summaries, and external knowledge stores-that carry information across sessions, enabling the agent to learn from past experiences and maintain continuity across multiple interactions.

Tool execution and external action represents another essential function. Language models alone can only produce text; they cannot browse the web, execute code, query databases, or generate images. The harness monitors the model's output for special tool-call commands and executes those operations on the model's behalf. When a tool call is detected, the harness pauses text generation, executes the requested operation in the external environment (such as performing a web search or running code in a sandbox), and feeds the results back into the model's context. This mechanism effectively gives the model "hands and eyes," transforming textual intentions into tangible real-world actions.

Context management and orchestration ensure that information flows efficiently between the model and its environment. The harness determines what information is provided to the model at each step, managing the transient prompt whilst maintaining a persistent task log separate from the model's immediate context. This separation is crucial for long-running projects: even if an AI agent instance stops and a new one begins later with no memory in the raw LLM, the project itself retains memory through files and logs maintained by the harness.

Modular Design and Components

Contemporary harness architectures increasingly adopt modular designs that decompose agent functionality into interchangeable components. Research from ICML 2025 on "General Modular Harness for LLM Agents in Multi-Turn Gaming Environments" demonstrates this approach through three core modules: perception, which processes both low-resolution grid environments and visually complex images; memory, which stores recent trajectories and synthesises self-reflection signals enabling agents to critique past moves and adjust future plans; and reasoning, which integrates perceptual embeddings and memory traces to produce sequential decisions. This modular structure allows developers to toggle components on and off, systematically analysing each module's contribution to overall performance.

Performance Impact and Practical Benefits

The empirical benefits of harness implementation are substantial. Models operating within a harness achieve significantly higher task success rates compared to un-harnessed baselines. In gaming environments, an AI with a memory and perception harness wins more games than the same AI without one. In coding tasks, an AI with a harness that runs and debugs its own code completes programming tasks that a standalone LLM would fail due to runtime errors. The harness essentially compensates for the model's inherent weaknesses-lack of persistence, inability to access external knowledge, and propensity for errors-resulting in markedly improved real-world performance.

Perhaps most significantly, harnesses extend what an AI can accomplish without requiring model retraining. Want an LLM to handle images? Integrate a vision module or image captioning API into the harness. Need mathematical reasoning or complex logic? Add the appropriate tool or module. This extensibility makes harnesses economically valuable: two products using identical underlying LLMs can deliver vastly different user experiences based on the quality and sophistication of their respective harnesses.

Evolution and Strategic Importance

As AI capabilities have advanced, harness design has become increasingly critical to product success. The harness landscape is dynamic and evolving: popular agents like Manus have undergone five complete re-architectures since March 2024, and even Anthropic continuously refines Claude Code's agent harness as underlying models improve. This reflects a fundamental principle: as models become more capable, harnesses must be continually simplified, stripping away scaffolding and crutches that are no longer necessary.

The distinction between orchestration and harness is worth noting. Orchestration serves as the "brain" of an AI system-determining the overall workflow and decision logic-whilst the harness functions as the "hands and infrastructure," executing those decisions and managing the technical details. Both are critical for complex AI agents, and improvements in either dimension can dramatically enhance real-world performance.

Related Theorist: Allen Newell and Cognitive Architecture

Allen Newell (1927-1992) was an American cognitive scientist and computer scientist whose theoretical framework profoundly influences contemporary harness design. Newell's "Unified Theories of Cognition" (UTC), published in 1990, proposed that human cognition operates through integrated systems of perception, memory, and reasoning-three faculties that work in concert to enable intelligent behaviour. This theoretical foundation directly inspired the modular harness architectures now prevalent in AI research.

Newell's career spanned the emergence of cognitive science as a discipline. Working initially at the RAND Corporation and later at Carnegie Mellon University, he collaborated with Herbert Simon to develop the "Physical Symbol System Hypothesis," which posited that physical symbol systems (such as computers) could exhibit intelligent behaviour through the manipulation of symbols according to rules. This work earned Newell and Simon the Turing Award in 1975, recognising their foundational contributions to artificial intelligence.

Newell's UTC represented his mature synthesis of decades of research into human problem-solving, learning, and memory. Rather than treating perception, memory, and reasoning as separate cognitive modules, Newell argued they must be understood as deeply integrated systems operating within a unified cognitive architecture. This insight proved prescient: modern AI harnesses implement precisely this integration, with perception modules processing environmental information, memory modules storing and retrieving relevant context, and reasoning modules synthesising these inputs into coherent action sequences.

The connection between Newell's theoretical work and contemporary harness design is not merely coincidental. Researchers explicitly cite Newell's framework when justifying modular harness architectures, recognising that his cognitive science insights provide a principled foundation for engineering AI systems. In this sense, Newell's work from the 1980s and early 1990s anticipated the architectural requirements that AI engineers would discover empirically decades later when attempting to build capable, persistent, tool-using agents.

Quote: Clayton Christensen

"When I have my interview with God, our conversation will focus on the individuals whose self-esteem I was able to strengthen, whose faith I was able to reinforce, and whose discomfort I was able to assuage - a doer of good, regardless of what assignment I had. These are the metrics that matter in measuring my life." - Clayton Christensen - Author

Clayton M. Christensen, the renowned Harvard Business School professor and author, encapsulated a lifetime of reflection in this poignant reflection on true success. Drawn from his seminal book How Will You Measure Your Life?, published in 2012, the quote emerges from Christensen's classroom exercise where he challenged students to confront life's deepest questions: How can I ensure happiness in my career? How can I nurture enduring family relationships? And how can I avoid moral pitfalls that lead to downfall?^1,2,3

Christensen's Life and Intellectual Journey

Born in 1952 in Salt Lake City, Utah, Christensen rose from humble roots to become one of the most influential management thinkers of his generation. A devout member of The Church of Jesus Christ of Latter-day Saints, he infused his work with ethical considerations, often drawing parallels between business strategy and personal integrity. He earned a DBA from Harvard Business School in 1992, where he later became the Kim B. Clark Professor of Business Administration.^3,7

Christensen's breakthrough came with The Innovator's Dilemma (1997), which introduced the theory of disruptive innovation - the idea that established companies often fail by focusing on high-margin customers while upstarts target overlooked markets, eventually upending incumbents. This concept, praised by Steve Jobs as deeply influential, transformed how leaders view competition and change.² His ideas permeated industries, from technology to healthcare, earning him accolades like the Economist Innovation Award.

Tragedy struck in 2010 when Christensen was diagnosed with leukemia, prompting deeper introspection. Amid treatments, he expanded his final HBS class into How Will You Measure Your Life?, co-authored with James Allworth and Karen Dillon. The book applies rigorous business theories - like marginal cost analysis and resource allocation - to life's choices, warning against 'just this once' compromises that erode integrity over time.^3,7 Christensen passed away in 2020, but his emphasis on relationships over achievements endures.

Context of the Quote in 'How Will You Measure Your Life?'

The quote anchors the book's core thesis: conventional metrics like wealth or status pale against the impact on others' lives. Christensen recounted posing these questions to ambitious MBAs, urging them to invest deliberately in relationships, as career peaks fade but personal bonds provide lasting happiness.^1,4 He illustrated pitfalls through cases like Nick Leeson, whose minor ethical lapse at Barings Bank spiralled into fraud and ruin, underscoring that 100% adherence to principles is easier than 98%.³

In sections on career and relationships, Christensen advised balancing ambition with family time, using 'jobs to be done' theory: people 'hire' you for specific roles, like parents modelling values or partners providing support. At life's end, he argued, success lies in friends who console you, children embodying your values, and a resilient marriage - not accolades.^4,5

Leading Theorists on Life Priorities and Fulfilment

Christensen built on a lineage of thinkers prioritising inner metrics over external gains:

• Viktor Frankl, Holocaust survivor and author of Man's Search for Meaning (1946), posited that fulfilment stems from purpose and love, not pleasure - influencing Christensen's focus on meaningful impact.³
• Abraham Maslow's hierarchy of needs culminates in self-actualisation, where self-esteem and relationships foster peak experiences, aligning with Christensen's relational emphasis.⁴
• Martin Seligman, father of positive psychology, advocated measuring life via PERMA (Positive Emotion, Engagement, Relationships, Meaning, Accomplishment), reinforcing that relationships yield the highest wellbeing.²
• Daniel Kahneman, Nobel laureate, distinguished 'experiencing self' (daily highs) from 'remembering self' (enduring memories), cautioning that peak achievements matter less retrospectively than sustained bonds.³

These theorists converge on a truth Christensen championed: true leadership - in business or life - measures by upliftment of others, not personal ascent. His framework equips readers to audit priorities, ensuring actions align with eternal metrics of good.^1,7

Term: Loss function

"A loss function, also known as a cost function, is a mathematical function that quantifies the difference between a model's predicted output and the actual 'ground truth' value for a given input." - Loss function

A loss function is a mathematical function that quantifies the discrepancy between a model's predicted output and the actual ground truth value for a given input. Also referred to as an error function or cost function, it serves as the objective function that machine learning and artificial intelligence algorithms seek to optimize during training efforts.

Core Purpose and Function

The loss function operates as a feedback mechanism within machine learning systems. When a model makes a prediction, the loss function calculates a numerical value representing the prediction error-the gap between what the model predicted and what actually occurred. This error quantification is fundamental to the learning process. During training, algorithms such as backpropagation use the gradient of the loss function with respect to the model's parameters to iteratively adjust weights and biases, progressively reducing the loss and improving predictive accuracy.

The relationship between loss function and cost function warrants clarification: whilst these terms are often used interchangeably, a loss function technically applies to a single training example, whereas a cost function typically represents the average loss across an entire dataset or batch. Both, however, serve the same essential purpose of guiding model optimization.

Key Roles in Machine Learning

Loss functions fulfil several critical functions within machine learning systems:

• Performance measurement: Loss functions provide a quantitative metric to evaluate how well a model's predictions align with actual results, enabling objective assessment of model effectiveness.
• Optimization guidance: By calculating prediction error, loss functions direct the learning algorithm to adjust parameters iteratively, creating a clear path toward improved predictions.
• Bias-variance balance: Effective loss functions help balance model bias (oversimplification) and variance (overfitting), essential for generalisation to new, unseen data.
• Training signal: The gradient of the loss function provides the signal by which learning algorithms update model weights during backpropagation.

Common Loss Function Types

Different machine learning tasks require different loss functions. For regression problems involving continuous numerical predictions, Mean Squared Error (MSE) and Mean Absolute Error (MAE) are widely employed. The MAE formula is:

\text = \frac \sum_^ \left| y_i - \hat_i \right|

For classification tasks dealing with categorical data, Binary Cross-Entropy (also called Log Loss) is commonly used for binary classification problems. The formula is:

L(y, f(x)) = -[y \cdot \log(f(x)) + (1 - y) \cdot \log(1 - f(x))]

where y represents the true binary label (0 or 1) and f(x) is the predicted probability of the positive class.

For multi-class classification, Categorical Cross-Entropy extends this concept. Additionally, Hinge Loss is particularly useful in binary classification where clear separation between classes is desired:

L(y, f(x)) = \max(0, 1 - y \cdot f(x))

The Huber Loss function provides robustness to outliers by combining quadratic and linear components, switching between them based on a threshold parameter delta (?).

Related Strategy Theorist: Vladimir Vapnik

Vladimir Naumovich Vapnik (born 1935) stands as a foundational figure in the theoretical underpinnings of loss functions and machine learning optimisation. A Soviet and later American computer scientist, Vapnik's work on Statistical Learning Theory and Support Vector Machines (SVMs) fundamentally shaped how the machine learning community understands loss functions and their role in model generalisation.

Vapnik's most significant contribution to loss function theory came through his development of Support Vector Machines in the 1990s, where he introduced the concept of the hinge loss function-a loss function specifically designed to maximise the margin between classification boundaries. This represented a paradigm shift in thinking about loss functions: rather than simply minimising prediction error, Vapnik's approach emphasised confidence and margin, ensuring models were not merely correct but confidently correct by a specified distance.

Born in the Soviet Union, Vapnik studied mathematics at the University of Uzbekistan before joining the Institute of Control Sciences in Moscow, where he conducted groundbreaking research on learning theory. His theoretical framework, Vapnik-Chervonenkis (VC) theory, provided mathematical foundations for understanding how models generalise from training data to unseen examples-a concept intimately connected to loss function design and selection.

Vapnik's insight that different loss functions encode different assumptions about what constitutes "good" model behaviour proved revolutionary. His work demonstrated that the choice of loss function directly influences not just training efficiency but the model's ability to generalise. This principle remains central to modern machine learning: data scientists select loss functions strategically to encode domain knowledge and desired model properties, whether robustness to outliers, confidence in predictions, or balanced handling of imbalanced datasets.

Vapnik's career spanned decades of innovation, including his later work on transductive learning and learning using privileged information. His theoretical contributions earned him numerous accolades and established him as one of the most influential figures in machine learning science. His emphasis on understanding the mathematical foundations of learning-particularly through the lens of loss functions and generalisation bounds-continues to guide contemporary research in deep learning and artificial intelligence.

Practical Significance

The selection of an appropriate loss function significantly impacts model performance and training efficiency. Data scientists carefully consider different loss functions to achieve specific objectives: reducing sensitivity to outliers, better handling noisy data, minimising overfitting, or improving performance on imbalanced datasets. The loss function thus represents not merely a technical component but a strategic choice that encodes domain expertise and learning objectives into the machine learning system itself.

Quote: Clayton Christensen

"The only metrics that will truly matter to my life are the individuals whom I have been able to help, one by one, to become better people." - Clayton Christensen - Author

Clayton Christensen's assertion that personal impact-measured through the individuals we help develop-represents the truest metric of a life well-lived stands as a profound counterpoint to the achievement-obsessed culture that dominates modern professional life. This reflection emerges not from abstract philosophy but from decades of observing how talented, ambitious people construct meaning, and from Christensen's own wrestling with what constitutes genuine success.

The Context: A Harvard Professor's Reckoning

Christensen, the Thomas Bowers Professor of Business Administration at Harvard Business School and author of the seminal work The Innovator's Dilemma, developed this perspective through direct engagement with some of the world's most driven individuals: MBA students at one of the planet's most competitive institutions. Each year, he posed three deceptively simple questions to his students on the final day of class: How can I be sure I'll be happy in my career? How can I be sure my relationships with family become an enduring source of happiness? How can I be sure I'll stay out of jail?

These questions, which form the foundation of his 2012 book How Will You Measure Your Life? (co-authored with James Allworth and Karen Dillon), reveal Christensen's conviction that conventional metrics of success-wealth, title, achievement-systematically mislead us about what actually generates lasting fulfilment. The book, published by Harper Business, synthesises decades of academic research with personal narrative to argue that well-tested theories from business and psychology can illuminate the path to a meaningful life.

The Danger of Marginal Thinking

Central to Christensen's argument is his critique of how marginal-cost analysis-a cornerstone of business decision-making-infiltrates personal life with corrosive consequences. He illustrates this through the cautionary tale of Nick Leeson, the trader whose "just this once" decisions ultimately destroyed Barings Bank, a 233-year-old institution, and landed him in prison. Leeson's descent began with a single small error, hidden in a little-scrutinised trading account. Each subsequent deception seemed a marginal step, yet the cumulative effect was catastrophic.

Christensen argues that we unconsciously apply this same logic to our personal and moral lives. A voice whispers: "I know most people shouldn't do this, but in this particular extenuating circumstance, just this once, it's okay." The price appears alluringly low. Yet life, Christensen observes, presents an endless stream of extenuating circumstances. Once we justify crossing a boundary once, nothing prevents us from crossing it again. The boundary itself-our personal moral line-loses its power.

This insight directly connects to his central claim about measuring life through human development. If we measure success by quarterly results, promotions, or wealth accumulation, we unconsciously permit ourselves small moral compromises that seem justified by marginal analysis. But if we measure success by the individuals we've genuinely helped become better people, our decision-making framework shifts entirely. Helping someone develop requires consistency, integrity, and long-term commitment-qualities incompatible with marginal thinking.

The Theoretical Foundations

Christensen's perspective draws on several streams of organisational and psychological theory. His work on innovation theory-developed through The Innovator's Dilemma, which Steve Jobs described as "deeply influencing" Apple's strategy-emphasises how organisations often fail by optimising for present circumstances rather than building capabilities for future challenges. This same principle applies to personal development: we often optimise for immediate achievement rather than building the relational and moral capabilities that sustain meaning across decades.

The book also engages with motivation theory, particularly the distinction between intrinsic and extrinsic motivators. Research in psychology, notably the work of Edward Deci and Richard Ryan on self-determination theory, demonstrates that extrinsic rewards (money, status, recognition) provide temporary satisfaction but rarely generate enduring happiness. Intrinsic motivators-autonomy, mastery, and purpose-create deeper engagement and fulfilment. Christensen argues that helping others develop satisfies all three intrinsic motivators: you exercise agency in how you mentor, you develop mastery in your field, and you connect to a purpose beyond yourself.

Additionally, Christensen draws on research in positive psychology and life satisfaction studies. Longitudinal research, including the Harvard Study of Adult Development (which tracked individuals across decades), consistently demonstrates that the quality of relationships-not career achievement or wealth-predicts life satisfaction and longevity. Christensen synthesises this research with business theory to argue that the mechanism through which relationships generate happiness is precisely through the mutual development of the individuals involved.

The Concept of Being "Hired"

A distinctive element of Christensen's framework is his concept of being "hired" to do a job in someone's life. Rather than viewing relationships as passive connections, he suggests we should understand them as ongoing engagements where others, implicitly or explicitly, hire us to fulfil specific roles: mentor, example, confidant, supporter. This reframing transforms how we approach relationships. If your child has hired you to be an example of integrity, your daily choices take on different weight. If your colleague has hired you to help them develop their capabilities, your mentoring becomes a central measure of your professional contribution.

This concept echoes the work of Clayton Alderfer and other organisational psychologists who emphasise the importance of role clarity and psychological contracts in generating satisfaction. But Christensen extends it beyond the workplace into all human relationships, suggesting that clarity about what role we're playing-and commitment to excellence in that role-generates both happiness for ourselves and genuine development for others.

The Paradox of Achievement

Christensen acknowledges a subtle paradox: those with strong achievement drives-precisely the individuals most likely to attend Harvard Business School-face particular risk. Their ambition, which drives professional success, can simultaneously blind them to what generates lasting happiness. He recounts a personal moment when, as a young man, he faced a choice between attending an important basketball game (where his team needed him) and pursuing a business opportunity. He chose the game, reasoning that his team needed him. They won anyway without him. Yet he later recognised this decision as among the most important of his life-not because of the game's outcome, but because it established a boundary: relationships matter more than marginal professional gains.

This reflects research on what psychologists call the "arrival fallacy"-the discovery that achieving long-sought goals often fails to generate the anticipated happiness. Christensen argues this occurs because achievement-focused individuals have internalised the wrong metric. They measure success by what they accomplish, when they should measure it by who they've helped become.

Implications for Leadership and Mentorship

For leaders and managers, Christensen's framework suggests a radical reorientation of purpose. Rather than viewing your role primarily through the lens of organisational performance, financial results, or strategic objectives, you might ask: which individuals have I genuinely helped develop? Have I created conditions where they've grown in capability, confidence, and character? This doesn't negate the importance of business results-Christensen emphasises that career provides stability and resources to give to others. But it reorders priorities.

This perspective aligns with contemporary research on authentic leadership and servant leadership, which emphasises that leaders generate the greatest impact-both organisational and personal-when they prioritise the development of those they lead. Research by scholars like James Kouzes and Barry Posner demonstrates that leaders remembered as transformational are those who invested in developing others, not merely those who achieved impressive financial results.

The Long View

Christensen's metric requires patience and a long temporal horizon. You won't know if you've raised a good son or daughter until twenty years after the bulk of your parenting work. You won't know if you have true friends until they call to console you during genuine hardship. You won't know if you've built an enduring marriage until you've navigated the challenges that cause many relationships to fracture. This stands in sharp contrast to the quarterly earnings reports, annual performance reviews, and immediate feedback loops that dominate modern professional life.

Yet this long view, Christensen argues, is precisely what liberates us from marginal thinking. When you recognise that the true measure of your life will be assessed across decades, the temptation to compromise your principles "just this once" loses its power. The small decision to help someone develop, made consistently over years, compounds into a life of genuine impact. Conversely, the small decision to prioritise marginal professional gain over relational investment, repeated across years, compounds into a life of hollow achievement.

Christensen's insight ultimately suggests that the question "How will you measure your life?" is not merely philosophical but profoundly practical. It shapes daily decisions about where you invest your time, energy, and integrity. And those daily decisions, accumulated across a lifetime, determine not just your happiness but the legacy you leave: the individuals who became better people because you were present in their lives.

Term: AI scaffolding

"Scaffolding refers to the structured architecture and instructional techniques built around an AI model to enhance its reasoning, reliability, and capability." - AI scaffolding

AI scaffolding is the structured architecture and tooling built around a large language model (LLM) to enable it to perform complex, goal-driven tasks with enhanced reasoning, reliability, and capability.¹ Rather than relying on a single prompt or query, scaffolding places an LLM within a control loop that includes memory systems, external tools, decision logic, and feedback mechanisms, allowing the model to observe its environment, call APIs or code, update its context, and iterate until goals are achieved.¹

In essence, scaffolding bridges the critical gap between the capabilities of base models and production-ready systems. A standalone LLM lacks the architectural support needed to reliably complete multi-step tasks, interface with business systems, or adapt to domain-specific requirements.¹ Scaffolding augments the model's bare capabilities by providing access to tools, domain data, and structured workflows that guide and extend its behaviour.

Core Components of AI Scaffolding

Effective scaffolding operates through several interconnected layers:

• Planning and reasoning: Agents operate through defined reasoning and evaluation steps. Rather than acting immediately, scaffolding may prompt the model to plan or reflect before taking action, and to self-critique its outputs. Research demonstrates that allowing agents to plan and self-evaluate significantly improves problem-solving accuracy compared to action-only approaches.¹
• Tool integration: The LLM is wrapped in code that interprets its outputs as tool calls. When the model determines it needs external resources-such as a calculator, database query, API call, or web search-the scaffold safely executes that tool and returns results to the model for the next reasoning step.¹
• Memory systems: Scaffolding includes mechanisms for the agent to maintain and update context across multiple interactions, enabling it to build upon previous observations and decisions.¹
• Feedback and control: Robust agents include feedback loops and safeguards such as self-evaluation steps, human-in-the-loop checks, and policy enforcement. In enterprise settings, scaffolding adds logging, testing suites, and guardrails like content filters to ensure outputs remain controlled and auditable.¹

Types of AI Scaffolding Techniques

AI scaffolding encompasses several distinct approaches, which can be combined to enhance model performance:

• Tool access scaffolding: Granting models access to external tools such as code editors, web browsers, or specialised software significantly expands their problem-solving capabilities. For example, LLMs initially trained on finite datasets with fixed cut-off dates became substantially more capable when granted internet access.²
• Agent loop scaffolding: This technique automates multi-step task completion by placing AI models in a loop with access to their own observations and actions, enabling them to self-generate each prompt needed to finish complex tasks. Systems like AutoGPT exemplify this approach.²
• Multi-agent scaffolding: Multiple AI models collaborate on complex problems through dialogue, division of labour, or critique mechanisms. Research shows that extended networks of up to a thousand agents can coordinate to outperform individual models, with capability scaling predictably as networks grow larger.²
• Procedural scaffolding: This approach builds a structured process in which the model generates outputs, checks them, and revises them iteratively, enforcing process discipline rather than relying on raw prompts alone.³
• Semantic scaffolding: Using ontological frameworks and domain rules to validate outputs against formal relations, preventing deeper misunderstandings and moving AI closer to auditable, trustworthy reasoning.³

Practical Applications and Enterprise Use

Scaffolding is essential for operationalising LLMs in enterprise environments. Whether an agent is expected to generate structured outputs, interact with APIs, or solve problems through planning and iteration, its effectiveness depends on the scaffold that guides and extends its behaviour.¹ In sectors such as customer service, risk analysis, logistics, healthcare, and finance, scaffolding enables AI systems to maintain reliability and auditability in high-stakes contexts.³

A key advantage of scaffolding is that it improves accuracy whilst making AI reasoning more transparent. When a system reaches a conclusion, leaders can trace it back to formal relations in an ontology rather than relying solely on statistical inference, making the system trustworthy for critical applications.³

Scaffolding versus Model Scale

An important principle in modern AI development is that scaffolding often matters more than raw model scale. The future of AI-whether in homeland security, finance, healthcare, or other domains-will be defined not by the size of models but by the quality of the architectural frameworks surrounding them.³ Hybrid architectures that embed statistical models within well-designed scaffolded systems deliver superior performance and reliability compared to simply scaling larger models without structural support.

Key Theorist: Stuart Russell and the Alignment Research Tradition

The conceptual foundations of AI scaffolding are deeply rooted in the work of Stuart Russell, a leading figure in artificial intelligence safety and alignment research. Russell, the Volgenau Chair of Engineering at the University of California, Berkeley, and co-author of the seminal textbook Artificial Intelligence: A Modern Approach, has been instrumental in developing frameworks for ensuring AI systems remain controllable and aligned with human values as they become more capable.

Russell's contributions to scaffolding theory emerge from his broader research agenda on AI safety and the control problem. In the early 2000s, as machine learning systems began to demonstrate increasing autonomy, Russell recognised that simply building more powerful models without corresponding advances in control architecture would create dangerous misalignment between AI capabilities and human oversight. His work emphasised that the architecture surrounding an AI system-not merely the model itself-determines whether that system can be safely deployed in high-stakes environments.

One of Russell's most influential contributions to scaffolding concepts is his work on iterated amplification, developed in collaboration with researchers at OpenAI and other institutions. Iterated amplification is a form of scaffolding that uses multi-AI collaborations to solve increasingly complex problems whilst maintaining human oversight at each stage. In this approach, humans decompose complex tasks into simpler subtasks that AI systems solve, then humans review and synthesise these solutions. Over time, humans operate at progressively higher levels of abstraction whilst AI systems assume responsibility for more of the process. This iterative cycle improves model capabilities whilst preserving human auditability and control-a principle directly aligned with scaffolding's core objective.²

Russell's broader philosophical stance is that AI safety and capability enhancement are not opposing forces but complementary objectives. Scaffolding embodies this principle: by building structured architectures around models, developers simultaneously enhance capability (through tool access, planning, and feedback loops) and improve safety (through auditability, human-in-the-loop checks, and formal validation against domain rules). Russell's insistence that AI systems must remain interpretable and auditable has directly influenced how modern scaffolding frameworks incorporate semantic validation, ontological constraints, and transparent reasoning pathways.

Throughout his career, Russell has advocated for what he terms "beneficial AI"-systems designed from inception to be controllable, transparent, and aligned with human values. Scaffolding represents a practical instantiation of this vision. Rather than hoping that larger models will somehow become more trustworthy, Russell's framework suggests that intentional architectural design-the very essence of scaffolding-is the path to AI systems that are simultaneously more capable and more reliable.

Russell's influence extends beyond theoretical work. His research group at Berkeley has contributed to developing practical frameworks for AI governance, model evaluation, and safety testing that directly inform how organisations implement scaffolding in production environments. His emphasis on formal methods, constraint satisfaction, and human-AI collaboration has shaped industry standards for building enterprise-grade AI systems.

Quote: Clayton Christensen

"I had thought the destination was what was important, but it turned out it was the journey." - Clayton Christensen - Author

Clayton M. Christensen, the renowned Harvard Business School professor and author, encapsulated a profound shift in perspective with this reflection from his seminal work How Will You Measure Your Life? Published in 2010, the book draws on his business theories to offer timeless guidance on personal fulfilment, urging readers to prioritise meaningful processes over mere endpoints in life and career.^1,2

Who Was Clayton Christensen?

Born in 1952 in Salt Lake City, Utah, Christensen rose from humble beginnings to become one of the most influential thinkers in modern business. A devout member of The Church of Jesus Christ of Latter-day Saints, he integrated his faith with rigorous scholarship. He earned a BA from Brigham Young University, an MPhil from Oxford as a Rhodes Scholar, and both an MBA and DBA from Harvard Business School.

Christensen's breakthrough came with The Innovator's Dilemma (1997), introducing **disruptive innovation** - the theory that established companies often fail by focusing on high-end customers, allowing nimble entrants to dominate lower markets and eventually upscale.³ This framework reshaped industries like technology and healthcare. He authored over a dozen books, consulted for global firms, and taught at Harvard for decades until his death in January 2020 from complications of leukemia.

Despite professional acclaim, Christensen's later years emphasised personal integrity. He famously resisted 'just this once' compromises, a principle he credited for his life's direction: 'Resisting the temptation whose logic was 'In this extenuating circumstance, just this once, it's OK' has proven to be one of the most important decisions of my life.'^3,6

Context of the Quote in How Will You Measure Your Life?

The book stems from Christensen's 2010 Harvard MBA commencement address, expanded into chapters blending business strategy with life lessons. He warns against common traps: chasing rewards that scream loudest, neglecting family for career, or measuring success by wealth alone. Instead, he advocates allocating resources - time, energy, talent - towards aspirations.^4,5,6

This quote emerges in discussions of motivation and growth. Christensen reflects that true satisfaction arises not from arriving at goals, but from the daily pursuit of meaningful work, learning, and relationships. He writes: 'In order to really find happiness, you need to continue looking for opportunities that you believe are meaningful, in which you will be able to learn new things, to succeed, and be given more and more responsibility to shoulder.'^3,4 The journey, rich with motivators like progress and teamwork, forges character and joy.

Leading Theorists on Life Priorities and the Journey Metaphor

Christensen's insight echoes ancient and modern thinkers who elevate process over outcome.

• Aristotle (384-322 BC): In Nicomachean Ethics, he defined eudaimonia (flourishing) as a life of virtuous activity, not transient pleasures. Habits formed in daily practice, not endpoints, cultivate excellence.
• Lao Tzu (6th century BC): The Tao Te Ching states, 'A journey of a thousand miles begins with a single step.' Taoist philosophy prizes harmonious flow (wu wei) over forced achievement.
• Viktor Frankl (1905-1997): Holocaust survivor and Man's Search for Meaning author argued meaning emerges through attitude amid suffering. Logotherapy posits purpose in every moment's choices, prioritising inner journey.
• Mihaly Csikszentmihalyi (1934-2021): Pioneer of **flow theory** in Flow: The Psychology of Optimal Experience (1990). Peak experiences occur in immersive tasks matching skill and challenge - the essence of valuing journey.
• Daniel Kahneman (1934-2024): Nobel-winning psychologist distinguished 'experiencing self' (moment-to-moment) from 'remembering self' (end results). In Thinking, Fast and Slow, he showed people often overvalue peaks and endpoints, neglecting the journey's sum.

These theorists converge on Christensen's message: life's value lies in intentional, principle-driven paths. As he noted, 'The only metrics that will truly matter to my life are the individuals whom I have been able to help, one by one, to become better people.'^3,5

Enduring Relevance

In an era of hustle culture and metric-driven success, Christensen's words challenge us to recalibrate. His life exemplified this: battling illness while mentoring students, he measured legacy by impact, not accolades. This quote invites reflection - are we journeying with purpose, or merely racing to destinations that may disappoint?

Quote: Brian Moynihan - Bank of America CEO

"You can see upwards of $6 trillion in deposits flow off the liabilities of a banking system... into the stablecoin environment... they're either not going to be able to loan or they're going to have to get wholesale funding and that wholesale funding will come at a cost that will increase the cost of borrowing." - Brian Moynihan - Bank of America CEO

In the rapidly evolving landscape of digital finance, Brian Moynihan, CEO of Bank of America, issued a stark warning during the bank's Q4 2025 earnings call on 15 January 2026. He highlighted the potential for up to $6 trillion in deposits - roughly 30% to 35% of total US commercial bank deposits - to shift from traditional banking liabilities into the stablecoin ecosystem if regulators permit stablecoin issuers to pay interest.^1,2

Context of the Quote

Moynihan's comments arose amid intense legislative debates over stablecoin regulation in the United States. With US commercial bank deposits standing at $18.61 trillion in January 2026 and the stablecoin market capitalisation at just $315 billion, the scale of this projected outflow underscores a profound threat to the fractional reserve banking model.¹ Banks rely on low-cost customer deposits to fund loans to households and businesses, especially small and mid-sized enterprises. A mass migration to interest-bearing stablecoins would cripple lending capacity or force reliance on pricier wholesale funding, thereby elevating borrowing costs across the economy.^1,2

This concern echoes broader industry pushback. Executives from JPMorgan and Bank of America have criticised proposals allowing stablecoin yields or rewards, viewing them as direct competition. A US Senate bill aimed at formalising cryptocurrency regulation has stalled amid lobbying from the American Bankers Association, which seeks to prohibit interest on stablecoins. Meanwhile, the GENIUS Act, signed by President Donald Trump in July 2025, marked the first explicit crypto legislation, spurring financial institutions to enter the space while intensifying turf wars as crypto firms pursue banking charters.³

Who is Brian Moynihan?

Brian Moynihan has led Bank of America since January 2010, steering the institution through post-financial crisis recovery, digital transformation, and now the crypto challenge. A Harvard Law graduate with a prior stint at FleetBoston Financial, Moynihan expanded BofA's wealth management and consumer banking arms, growing assets to over $3 trillion. His tenure has emphasised regulatory compliance and innovation, yet he remains vocal on threats like stablecoins that could disrupt deposit stability.^1,2

Backstory on Leading Theorists in Stablecoins and Banking Disruption

The stablecoin phenomenon builds on foundational ideas from monetary theorists and crypto pioneers who envisioned programmable money challenging centralised banking.

• Satoshi Nakamoto: The pseudonymous creator of Bitcoin in 2008 laid the groundwork by introducing decentralised digital currency, free from central bank control. Bitcoin's volatility spurred stablecoins as a bridge to everyday use.¹
• Vitalik Buterin: Ethereum's co-founder (2015) enabled smart contracts, powering algorithmic stablecoins like DAI. Buterin's vision of decentralised finance (DeFi) posits stablecoins as superior stores of value with yields from on-chain protocols, bypassing banks.³
• Milton Friedman: The Nobel laureate's 1969 proposal for a computer-based money system with fixed supply prefigured stablecoins. Friedman argued such systems could curb inflation better than fiat, influencing modern dollar-pegged tokens like USDT and USDC.¹
• Hayek and Free Banking Theorists: Friedrich Hayek's Denationalisation of Money (1976) advocated competing private currencies, a concept realised in stablecoins issued by firms like Tether and Circle. This challenges the state's monopoly on money issuance.³
• Crypto Economists like Jeremy Allaire (Circle CEO): Allaire champions stablecoins as 'internet-native money' for payments and remittances, arguing they offer efficiency banks cannot match. His firm issues USDC, now integral to global transfers.^1,3

These thinkers collectively argue that stablecoins democratise finance, offering transparency, yield, and borderless access. Yet banking leaders like Moynihan counter that without safeguards, this shift risks systemic instability by eroding the deposit base that fuels economic growth.²

Implications for Finance

Moynihan's forecast spotlights a pivotal regulatory crossroads. Permitting interest on stablecoins could accelerate adoption, potentially reshaping payments, lending, and funding markets. Banks lobby for restrictions to preserve their model, while crypto advocates push for innovation. As frameworks like the GENIUS Act evolve, the battle over $6 trillion in deposits will define the interplay between traditional finance and blockchain.^1,3

Term: Right to Win

"The 'Right to Win' (RTW) is a company's unique, sustainable ability to succeed in a specific market by leveraging superior capabilities, products, and a differentiated 'way to play' that outperform competitors, giving them a better-than-even chance of creating value and growth." - Right to Win

A company's right to win is the recognition that it is better prepared than its competitors to attract and keep the customers it cares about, grounded in a sustainable competitive advantage that extends beyond short-term market positioning.¹ This concept represents more than simply having superior resources; it is the ability to engage in any competitive market with a better-than-even chance of success consistently over time.³ The right to win emerges when a company aligns three interlocking strategic elements: a differentiated way to play, a robust capabilities system, and product and service fit that work together coherently.¹

The Three Pillars of Right to Win

The foundation of a right to win rests on understanding what your company can do better than anyone else. Rather than pursuing growth indiscriminately across multiple areas, successful organisations focus on identifying three to six differentiating capabilities-the interconnected people, knowledge, systems, tools and processes that create distinctive value to customers.^1,5 These capabilities differ fundamentally from assets; whilst assets such as facilities, machinery, and supplier connections can be replicated by competitors, capabilities cannot.¹ The critical question becomes: "What do we do well to deliver value?"¹

A well-developed way to play represents a chosen position in a market, grounded in understanding your capabilities and where the market is heading.¹ This positioning must fulfil four essential criteria: there must be a market that values your approach; it must be differentiated from competitors' ways to play; it must remain relevant given expected industry changes; and it must be supported by your capabilities system, making it feasible.¹ Finally, the product and service fit ensures that offerings are directly aligned with the capabilities system, delivering superior returns to shareholders.¹

Coherence acts as the binding agent across these three elements.¹ Achieving alignment with one or even two elements proves insufficient; only when all three synchronise with one another and with the right market conditions can a company truly claim a sustainable right to win.¹

Building and Sustaining Competitive Advantage

The right to win is not inherited; it is earned through strategic alignment and disciplined execution.² This requires an in-depth understanding of the competitive landscape, customer expectations, and team capabilities.² A strategy that leverages unique assets or insights creates a competitive moat, making it challenging for competitors to catch up, though execution remains where many organisations falter.²

Innovation and adaptability prove essential to sustaining this advantage.² Organisations that continuously evolve, anticipate market shifts, and adapt their goods and services accordingly are more likely to maintain their competitive edge.² This does not mean chasing every new trend but rather maintaining a keen sense of which innovations align with core competencies and long-term vision.² Building a culture of excellence-attracting and nurturing top talent, fostering continuous improvement, and encouraging innovation-represents an often-overlooked yet significant asset in securing the right to win.²

Strategic Applications and Growth Pathways

Right-to-win strategies fall into four categories: customer-driven, capability-driven, value-chain-based, and those building on disruptive business models or technologies.⁴ The most utilised approach involves fulfilling unmet needs for existing customers that the core business does not currently address.⁴ However, the strategy delivering the biggest revenue gains involves leveraging core business capabilities-such as patents, technological know-how, or brand equity-to expand into adjacent and breakout businesses.⁴ Companies successfully utilising two or more right-to-win strategies to move into adjacent markets delivered 12 percentage points higher excess total shareholder return versus their subindustry peers.⁴

Assessing Your Right to Win

Organisations can evaluate their right to win through systematic analysis. This involves identifying the two most relevant competitors, determining three to six differentiating capabilities required for success, listing key assets and table-stakes activities, and rating performance across these dimensions.⁵ Differentiating capabilities should be specific and interconnected rather than merely listing functions or organisational units.⁵ For example, one of Apple's differentiating capabilities is "innovation around customer interfaces to create better communications and entertainment experiences."⁵ Assets, whilst less sustainable than capabilities, represent criteria important to the market and warrant inclusion in competitive assessment.⁵

Related Theorist: C.K. Prahalad and the Core Competence Framework

The concept of right to win draws significantly from the work of C.K. Prahalad (1941-2010), an influential Indian-American business theorist and consultant who fundamentally shaped modern strategic thinking through his development of the core competence framework. Prahalad's seminal 1990 Harvard Business Review article, co-authored with Gary Hamel, "The Core Competence of the Corporation," introduced the revolutionary idea that organisations should identify and leverage their unique, hard-to-imitate capabilities rather than pursuing diversification across unrelated business areas.¹

Born in Bangalore, India, Prahalad earned his undergraduate degree in physics and mathematics before pursuing business education. He spent much of his career at the University of Michigan's Ross School of Business, where he conducted extensive research on strategic management and organisational capability. His work challenged the prevailing strategic orthodoxy of the 1980s, which emphasised portfolio management and strategic business units. Instead, Prahalad argued that companies should view themselves as portfolios of core competencies-the collective learning and coordination of diverse production skills and technologies-rather than collections of discrete business units.

Prahalad's framework directly underpins the right to win concept. He demonstrated that sustainable competitive advantage emerges not from owning assets but from developing distinctive capabilities that competitors cannot easily replicate. His research showed that companies like Sony, Honda, and 3M succeeded not because they possessed superior resources but because they had cultivated unique organisational capabilities in areas such as miniaturisation, engine design, or innovation processes. These capabilities enabled them to enter adjacent markets and create new products that competitors struggled to match.

Beyond core competence theory, Prahalad later developed the concept of the "bottom of the pyramid," exploring how companies could create right-to-win strategies by serving low-income consumers in emerging markets through innovation and capability leverage. His work emphasised that strategic advantage comes from understanding what your organisation does distinctively well and then systematically building, protecting, and extending those capabilities across markets and customer segments.

Prahalad's intellectual legacy remains central to contemporary strategic management. His insistence that capabilities-not assets-form the foundation of competitive advantage directly informs how modern organisations approach the right to win. His framework provides the theoretical scaffolding that explains why companies with seemingly fewer resources can outperform better-capitalised competitors: they possess superior, integrated capabilities that create distinctive value. This insight transformed strategic planning from a financial exercise into a capabilities-centred discipline, making Prahalad's work indispensable to understanding the right to win in contemporary business strategy.

Quote: Clayton Christensen

"What's important is to get out there and try stuff until you learn where your talents, interests, and priorities begin to pay off. When you find out what really works for you, then it's time to flip from an emergent strategy to a deliberate one." - Clayton Christensen - Author

This profound advice from Clayton Christensen encapsulates a timeless principle for personal and professional growth: the value of experimentation followed by focused commitment. Drawn from his bestselling book How Will You Measure Your Life?, the quote urges individuals to embrace trial and error in discovering their true strengths before committing to a structured path. Christensen, a renowned Harvard Business School professor, applies business strategy concepts to life's big questions, advocating for an initial phase of exploration - termed 'emergent strategy' - before shifting to a 'deliberate strategy' once clarity emerges.^1,7

Who Was Clayton Christensen?

Clayton Magleby Christensen (1947-2020) was a Danish-American academic, author, and business consultant whose ideas reshaped management theory. Born in Salt Lake City, Utah, he earned a bachelor's degree in economics from Brigham Young University, an MBA from Harvard, and a DBA from Harvard Business School. Christensen joined the Harvard faculty in 1992, where he taught for nearly three decades, influencing generations of leaders.^1,5

His seminal work, The Innovator's Dilemma (1997), introduced the theory of disruptive innovation, explaining how established companies fail by focusing on sustaining innovations for current customers while overlooking simpler, cheaper alternatives that disrupt markets from below. This concept has been applied to industries from technology to healthcare, predicting successes like Netflix over Blockbuster. Christensen authored over a dozen books, including The Innovator's Solution and How Will You Measure Your Life? (2010, co-authored with James Allworth and Karen Dillon), which blends business insights with personal reflections drawn from his Mormon faith, family life, and battle with leukemia.^5,6,7

In How Will You Measure Your Life?, Christensen draws parallels between corporate pitfalls and personal missteps, warning against prioritising short-term gains over long-term fulfilment. The quoted passage appears in a chapter on career strategy, using emergent and deliberate strategies as metaphors for navigating life's uncertainties.⁷

Context of the Quote: Emergent vs Deliberate Strategy

Christensen distinguishes two strategic approaches, rooted in his research on successful companies. A deliberate strategy stems from conscious planning, data analysis, and long-term goals - ideal for stable, mature organisations like Procter & Gamble, which refines products based on market data.¹ It requires alignment across teams, where every member understands their role in the bigger picture. However, it risks blindness to peripheral opportunities, as rigid focus on the original plan can miss disruptions.^1,2

Conversely, an emergent strategy arises organically from bottom-up initiatives, experiments, and adaptations - common in startups like early Walmart, which pivoted from small-town stores after unplanned successes. Christensen notes that over 90% of thriving new businesses succeed not through initial plans but by iterating on emergent learnings, retaining resources to pivot when needed.^1,5,6

The quote applies this duality to personal development: start with emergent exploration - trying diverse roles, hobbies, and pursuits - to uncover what aligns talents, interests, and priorities. Once viable paths emerge, switch to deliberate focus for sustained progress. This mirrors Honda's accidental US motorcycle success, where employees' side experiments trumped the formal plan.⁶

Leading Theorists on Emergent and Deliberate Strategy

Christensen built on foundational work by Henry Mintzberg, a Canadian management scholar. In his 1987 paper 'Crafting Strategy' and book Strategy Safari, Mintzberg challenged top-down planning, arguing strategies often emerge from patterns in daily actions rather than deliberate designs. He identified strategy as a 'continuous, diverse, and unruly process', blending deliberate intent with emergent flexibility - ideas Christensen explicitly referenced.²

• Henry Mintzberg: Pioneered the emergent strategy concept in the 1970s-80s, critiquing rigid corporate planning. His '10 Schools of Strategy' framework contrasts design (deliberate) with learning (emergent) schools.²
• Michael Porter: Christensen's contemporary at Harvard, Porter championed deliberate competitive strategy via frameworks like the Five Forces and value chain (1980s). While Porter focused on positioning for advantage, Christensen highlighted how such strategies falter against disruption.¹
• Robert Burgelman: Stanford professor whose research on 'intraorganisational ecology' influenced Christensen, showing how autonomous units drive emergent strategies within firms like Intel.⁵

These theorists collectively underscore strategy's dual nature: deliberate for execution, emergent for innovation. Christensen uniquely extended this to personal life, making abstract theory accessible for leadership, coaching, and self-management.^3,4

Christensen's insights remain vital for leaders balancing adaptability with purpose, reminding us that true success - in business or life - demands knowing when to explore and when to commit.

Quote: Jamie Dimon - JP Morgan Chase CEO

"I think the harder thing to measure has always been tech projects. That's been true my whole life. It's also been true my whole life, the tech is what changes everything, like everything." - Jamie Dimon - JP Morgan Chase CEO

Jamie Dimon's candid observation captures a fundamental tension at the heart of modern business strategy: the profound impact of technology juxtaposed against the persistent challenge of measuring its value. Delivered during JPMorgan Chase's 2026 Investor Day on 24 February, this remark came amid revelations of the bank's unprecedented $19.8 billion technology budget - a 10% increase from 2025, with significant allocations to artificial intelligence (AI) projects.^1,2,4 As CEO of the world's largest bank by market capitalisation, Dimon's perspective is shaped by decades of navigating technological shifts, from the rise of digital banking to the current AI boom.

Jamie Dimon's Career and Leadership at JPMorgan Chase

Born in 1956 in New York City to Greek immigrant parents, Jamie Dimon began his career in finance at American Express in the 1980s, rising rapidly under the mentorship of Sandy Weill. He co-led the merger that created Citigroup in 1998 but parted ways acrimoniously in 2000. Dimon then transformed Bank One from near-collapse into a powerhouse, earning a reputation as a crisis manager. In 2004, he became CEO of JPMorgan Chase following its acquisition of Bank One, a role he has held for over two decades.³

Under Dimon's stewardship, JPMorgan has become a technology leader in banking. The firm employs over 300,000 people, with tens of thousands in tech roles, and invests billions annually in innovation. Dimon has long championed tech as a competitive moat, famously urging investors to 'trust him' on spending despite vague ROI metrics. In 2026, this commitment manifests in a tech budget swelled by $2 billion, driven by AI for customer service, personalised insights, and developer tools, amid rising hardware costs from AI chip demand.^1,5 Dimon predicts JPMorgan will be a 'winner' in the AI race, leveraging its data assets and No. 1 ranking in AI maturity among banks.^1,3

Context of the Quote: JPMorgan's 2026 Strategic Framework

The quote emerged in a Q&A at the 24 February 2026 event, responding to analyst pressure on tech ROI. CFO Jeremy Barnum highlighted technology as a major expense driver, up $9 billion overall, with $1.2 billion in investments including AI. Dimon acknowledged time savings from tech as 'too vague' to measure precisely, echoing lifelong observations from mainframes to cloud computing.^1,2 This aligns with broader warnings: AI will revolutionise operations but displace jobs, necessitating societal preparation like retraining and phased adoption to avoid shocks, such as mass unemployment from autonomous trucks.⁴

JPMorgan is aggressively deploying AI - its large language model serves 150,000 users weekly - while planning 'huge redeployment' for affected staff. Executives like Marianne Lake stress paranoia in competition, quoting 'Only the paranoid survive'. Rivals like Bank of America ($14 billion tech spend) underscore the sector-wide arms race.¹

Leading Theorists on Technology Measurement and Impact

Dimon's views resonate with seminal thinkers on technology's intangible returns. Peter Drucker, the father of modern management, argued in The Practice of Management (1954) that knowledge workers' output defies traditional metrics, prefiguring tech's measurement woes. He coined 'knowledge economy', emphasising innovation's long-term value over short-term quantification.

Erik Brynjolfsson and Andrew McAfee, MIT economists, explore this in The Second Machine Age (2014), detailing how digital technologies yield 'non-rival' benefits - exponential productivity without proportional costs - hard to capture in GDP or ROI. Their 'bounty vs. spread' framework warns of uneven gains, mirroring Dimon's job displacement concerns.⁴

Clayton Christensen's The Innovator's Dilemma (1997) explains why incumbents struggle with disruptive tech: metrics favour sustaining innovations, blinding firms to transformative ones. JPMorgan's shift from infrastructure modernisation to AI-ready data exemplifies overcoming this.⁵

In AI specifically, Nick Bostrom's Superintelligence (2014) and Stuart Russell's Human Compatible (2019) address measurement beyond finance - aligning superintelligent systems with human values amid unpredictable impacts. Dimon's pragmatic focus on phased integration echoes calls for cautious deployment.⁴

These theorists underscore Dimon's point: technology's true worth lies in reshaping 'everything', demanding faith in leadership over precise yardsticks. JPMorgan's strategy embodies this, positioning the bank at the vanguard of finance's technological frontier.