A fundamental change in the basic concepts and practices of a discipline, where the old framework is replaced by an entirely new way of seeing the world.
Model #0073Category: General Thinking & Meta-ModelsSource: Thomas KuhnDepth to apply:
In 1962, a physicist-turned-historian named Thomas Kuhn published a slim book that changed how we think about how thinking changes. The Structure of Scientific Revolutions introduced the term "paradigm shift" to describe something scientists had been experiencing for centuries but had never named: the sudden, wholesale replacement of one framework of understanding with another.
Kuhn's central observation was deceptively simple. Science doesn't progress in a smooth upward line. It lurches. For long stretches, researchers work within an accepted framework — what Kuhn called "normal science" — solving puzzles, refining measurements, extending the paradigm's reach. The paradigm defines which questions are legitimate, which methods are acceptable, and which answers count as knowledge. It is the water the fish swim in.
Then anomalies appear. Results that don't fit. Experiments that produce the wrong numbers. Observations that the framework can't explain without increasingly tortured amendments. For a while, these are dismissed, ignored, or patched over with ad hoc adjustments. Ptolemy's astronomy added epicycle upon epicycle to preserve the geocentric model — circles within circles, growing more elaborate as each new observation demanded another correction. The model worked, in the sense that it could retrodict planetary positions with reasonable accuracy. But the machinery had become grotesque.
At some point, the weight of accumulated anomalies triggers a crisis. The ad hoc patches become more complex than the underlying theory. Confidence in the old paradigm fractures — not among the public, but among the practitioners who live closest to the data and can no longer ignore the discrepancies. And then — not gradually, but abruptly — a new framework emerges that explains the anomalies, simplifies the machinery, and redefines what counts as a legitimate question. Copernicus replaced Earth-centered astronomy with a heliocentric model in 1543. It wasn't just a better answer to the old question. It was a different question entirely. The shift wasn't from "wrong epicycles" to "right epicycles." It was from "how do planets orbit Earth?" to "how do planets orbit the Sun?"
The pattern recurs with striking regularity. Newton's mechanics held for over two centuries until Einstein's 1905 papers on special relativity showed that Newtonian physics was an approximation — accurate at low velocities, fundamentally wrong at speeds approaching light. The geological consensus that continents were fixed gave way to plate tectonics in the 1960s after decades of dismissed evidence, including Alfred Wegener's 1912 proposal of continental drift, which was ridiculed for fifty years before seafloor spreading data made the old paradigm untenable. In biology, the central dogma of molecular biology — DNA makes RNA makes protein, one-directional flow — was disrupted by the discovery of reverse transcriptase in 1970 and later by the ENCODE project's revelation that vast stretches of "junk DNA" were functionally active.
The critical insight for founders and strategists: paradigm shifts don't announce themselves. They are always obvious in retrospect and nearly invisible in real time. The people most invested in the old paradigm — the experts, the incumbents, the tenured professors — are systematically the last to see the shift, because their expertise, status, and identity are built on the framework being replaced. Kuhn called this "incommensurability" — the old and new paradigms are so fundamentally different that practitioners of one literally cannot see the world the way practitioners of the other do. It's not disagreement. It's mutual incomprehension.
This is why Max Planck's dark observation rings true across domains: "A new scientific truth does not triumph by convincing its opponents and making them see the light, but rather because its opponents eventually die, and a new generation grows up that is familiar with it." Planck wasn't being cynical. He was describing the mechanics of paradigm change with empirical precision. A 2015 study by Azoulay, Fons-Rosen, and Graff Zivin published in the American Economic Review confirmed Planck's intuition quantitatively: when a prominent scientist dies, their subfield experiences a measurable influx of new researchers and new ideas. The old guard doesn't convert. It retires. And the field only moves forward once it has.
The model extends far beyond science. Every industry has its paradigm — the shared assumptions about how value is created, what customers want, and which business models work. When those assumptions break, the companies built on them break too. Kodak's paradigm was that people wanted physical photographs. Nokia's paradigm was that phones were for calling. Blockbuster's paradigm was that content required physical media. Each company was excellent within its paradigm. Each was destroyed by a shift it couldn't see — or saw too late, from within a framework that made the new reality look like noise rather than signal.
The structural lesson: paradigm shifts don't reward marginal improvement. They reward the willingness to abandon the old framework entirely and build for the new one before the new one has fully arrived. This is extraordinarily difficult in practice because the old paradigm is still generating revenue, still rewarding its practitioners, and still explaining most of the observed reality. The new paradigm looks fragile, incomplete, and unprofitable. Every incentive — financial, social, psychological — favours staying with the old framework until it's too late. The founders and scientists who navigate paradigm shifts successfully share one trait: they trust the structural logic of the shift over the surface-level evidence that the old paradigm is still working.
Section 2
How to See It
Paradigm shifts leave a distinctive trail of signals. The challenge is that during the shift, these signals are indistinguishable from the ordinary noise of incremental change — except to those who know the pattern.
Technology
You're seeing Paradigm Shift when incumbents describe a new technology as "a toy." When Steve Ballmer laughed at the iPhone in January 2007 — "$500? Fully subsidised? With a plan? That is the most expensive phone in the world, and it doesn't appeal to business customers because it doesn't have a keyboard" — he was evaluating a paradigm shift through the lens of the old paradigm. The iPhone wasn't a better phone. It was a pocket computer that happened to make calls. Ballmer's framework couldn't process that distinction because the category didn't exist within his paradigm. Every major technology shift produces this reaction from incumbents: mainframes dismissed minicomputers, minicomputers dismissed PCs, PCs dismissed smartphones.
Business
You're seeing Paradigm Shift when a company's core competency transforms from an asset into a liability. Kodak employed some of the world's best chemical engineers and held the fundamental patents on digital photography — Steve Sasson built the first digital camera at Kodak in 1975. But Kodak's expertise in film chemistry, its massive film manufacturing infrastructure, and its retail photo-finishing network all became anchors rather than engines once the paradigm shifted from chemical to digital imaging. The better you are at the old paradigm, the harder it is to abandon it. Kodak's excellence killed it.
Investing
You're seeing Paradigm Shift when traditional valuation metrics stop working and analysts declare the market "irrational." In 1999, value investors were baffled by Amazon trading at 100x revenue with no path to profitability under any traditional retail model. The old paradigm said a retailer's value was a function of same-store sales growth and operating margins. The new paradigm said platform economics, network effects, and infrastructure leverage created value through mechanisms the old framework couldn't measure. Not every "the market is crazy" observation signals a paradigm shift — sometimes the market really is crazy. But the pattern of experts declaring the world irrational is one of the earliest indicators that the underlying model of reality is changing.
Science
You're seeing Paradigm Shift when a young researcher's work is rejected by prestigious journals not because the data is wrong, but because the conclusions "don't make sense." Barry Marshall and Robin Warren proposed in 1982 that stomach ulcers were caused by the bacterium Helicobacter pylori, not by stress or spicy food. The gastroenterology establishment rejected the claim for over a decade — not because the evidence was weak, but because the prevailing paradigm held that bacteria couldn't survive in stomach acid. Marshall eventually drank a petri dish of H. pylori to prove the point. He developed gastritis within days. He and Warren won the Nobel Prize in 2005. The twenty-year delay between discovery and acceptance is the paradigm's immune response doing its job.
Section 3
How to Use It
Decision filter
"Is the anomaly I'm seeing a flaw in the data — or a flaw in the model? If the same anomaly keeps appearing across multiple independent observations, the model is probably wrong. And if the model is wrong, the entire competitive landscape built on that model is about to restructure."
As a founder
Your single most important strategic question is whether you're operating within an existing paradigm or at the edge of a new one. The tactics are completely different. Within an established paradigm, execution wins — build faster, sell harder, optimise relentlessly. At a paradigm boundary, the old execution playbook is not just unhelpful, it's actively misleading. It optimises for a world that's about to stop existing.
The founders who capture paradigm shifts don't necessarily see the future more clearly. They see the present more honestly. Reed Hastings didn't predict that streaming would work. He observed that bandwidth costs were declining at 30% per year and asked what entertainment distribution looks like when moving bits is cheaper than shipping plastic. The conclusion was obvious once you accepted the premise — but accepting the premise required abandoning the DVD-by-mail paradigm that was generating Netflix's revenue. Most founders can't make that leap because the old paradigm is still paying the bills.
As an investor
The highest-returning investments of any generation are concentrated in paradigm shifts — and the timing is brutal. Invest too early and you fund the pioneers who get arrows in their back. Invest too late and you pay the premium the market charges once the shift is consensus. The window is the period when smart money recognises the shift but the mainstream still thinks it's noise.
Peter Thiel's $500,000 angel investment in Facebook in 2004 was a paradigm-shift bet. The prevailing paradigm said social networking was a feature, not a platform — something MySpace-like that would be replicated by portals like Yahoo. Thiel saw that a real-identity social network with college-campus density was a different category entirely, one that would exhibit winner-take-all dynamics once it reached critical mass. The bet returned over $1 billion. The insight wasn't about Facebook's product. It was about recognising that the paradigm of anonymous, fragmented internet identity was ending.
As a decision-maker
When your organisation's metrics are all green but something feels structurally wrong, take that feeling seriously. Paradigm shifts frequently arrive disguised as minor competitive threats or temporary market disruptions. The old paradigm's metrics will look healthy right up until they don't — because the metrics themselves are products of the paradigm.
Blockbuster's same-store revenue was stable in 2005. Nokia's market share was dominant in 2006. BlackBerry's enterprise adoption was accelerating in 2007. Each company's dashboard showed a healthy business. Each was eighteen months from an existential crisis. The metrics weren't lying. They were measuring the wrong things — artefacts of a paradigm that was already collapsing underneath the numbers.
Common misapplication: The most dangerous misuse of this model is seeing paradigm shifts everywhere. Not every technology trend is a paradigm shift. Not every market disruption signals a fundamental restructuring. Blockchain was described as a paradigm shift for finance in 2017; by 2023, the most valuable applications were largely speculative tokens and incremental improvements to settlement infrastructure. 3D printing was described as a paradigm shift for manufacturing in 2013; a decade later, it remains a useful prototyping tool, not a replacement for injection moulding at scale.
The model demands discipline: a real paradigm shift replaces the entire explanatory framework, not just one product or process within it. If the existing framework can absorb the innovation without changing its fundamental assumptions, it's not a paradigm shift — it's progress within the current paradigm. The litmus test: does the innovation require you to abandon the old model's core questions, or does it simply provide a better answer to the same questions? Better answers are progress. Different questions are paradigm shifts. Faster horses versus automobiles. Better film versus digital sensors. More efficient call centres versus self-service software. The first item in each pair is paradigm refinement. The second is paradigm replacement.
Section 4
The Mechanism
Section 5
Founders & Leaders in Action
Paradigm shifts don't reward the smartest person in the room. They reward the person who recognises, earliest, that the room itself is about to be demolished. The founders below didn't just adapt to new paradigms — they built the infrastructure of the new reality while the old one was still standing.
What distinguishes these cases from ordinary strategic pivots is the scale of the framework being replaced. Each founder confronted not just a competitive challenge but a fundamental restructuring of how their industry defined value, measured success, and served customers. Their advantage wasn't superior information — in most cases, the relevant data was publicly available. Their advantage was the willingness to accept what the data implied and act on it before the implications became consensus.
On January 9, 2007, Jobs walked onto the Macworld stage and announced three products: a widescreen iPod with touch controls, a revolutionary mobile phone, and a breakthrough internet communicator. Then he revealed they were one device. The audience laughed. They didn't understand yet.
The mobile phone industry in 2007 operated within a clear paradigm: phones were communications devices with hardware keyboards, carrier-controlled software, and feature differentiation based on call quality and battery life. Nokia held 49.4% of global smartphone market share. BlackBerry dominated enterprise. Both were executing brilliantly within the existing paradigm.
Jobs wasn't competing within that paradigm. He was replacing it. The iPhone redefined the phone as a general-purpose computer with a cellular radio — a platform for software rather than a device for calls. The paradigm shift wasn't the touchscreen (LG and others had touchscreen phones). It was the mental model: that a phone's value was determined by its software ecosystem, not its hardware specifications. Within four years, Nokia's smartphone market share collapsed from 49.4% to 3%. BlackBerry went from 20% to irrelevance by 2013. Neither company made bad phones. They made excellent products for a paradigm that no longer existed.
The signal Jobs read was structural, not competitive: when a computer fits in your pocket, the question stops being "what should a phone do?" and becomes "what should a pocket computer do?" The two questions produce entirely different product architectures, business models, and ecosystems. Every mobile phone company was answering the first question with increasing sophistication. Jobs answered the second question with the App Store in 2008, which created a platform that third-party developers turned into the most valuable ecosystem in technology history. The paradigm shifted from hardware to platform — and every company that continued optimising hardware within the old framework was competing for a prize that had already moved. By 2024, the App Store ecosystem had generated over $1.1 trillion in developer earnings. The paradigm shift didn't just change which company won the phone market. It created an entirely new market — the app economy — that the old paradigm couldn't conceive of, let alone capture.
In 2006, NVIDIA was a graphics card company serving the gaming market, competing with ATI (later AMD) on frame rates and pixel shader performance. The paradigm said GPUs rendered graphics. That was the category, the competitive landscape, and the investor thesis.
Huang saw something the paradigm couldn't frame. GPUs weren't just graphics processors — they were massively parallel computation engines. A CPU processes instructions sequentially, one thread at a time (or a handful with modern cores). A GPU processes thousands of threads simultaneously. For certain workloads — matrix multiplication, neural network training, physics simulations — the architecture was orders of magnitude more efficient than any CPU.
In 2006, NVIDIA released CUDA, a parallel computing platform that allowed developers to write general-purpose software for GPUs. The market response was tepid. Gaming investors didn't care about scientific computing. Scientific computing investors didn't think of NVIDIA as relevant. The initiative consumed R&D resources with no immediate revenue justification. Wall Street analysts questioned the strategy.
Then deep learning happened. When AlexNet won the ImageNet competition in 2012 by a dramatic margin — trained on NVIDIA GPUs — the paradigm began to shift. By 2016, every major AI research lab was running on NVIDIA hardware. By 2023, NVIDIA's data centre revenue exceeded its gaming revenue for the first time. By 2024, the company's market capitalisation surpassed $3 trillion, briefly becoming the most valuable company on Earth.
Huang had positioned NVIDIA at the centre of a paradigm shift — from sequential to parallel computation — nearly a decade before the market recognised the shift was happening. The CUDA investment that looked like a distraction in 2006 became the moat that made NVIDIA's dominance structurally unassailable. Competitors who tried to enter the AI accelerator market after 2016 found that NVIDIA's advantage wasn't just hardware — it was the decade of software ecosystem development, developer tooling, and library support that CUDA had built. The paradigm shift rewarded the company that committed to the new framework years before it generated revenue.
Hastings understood paradigm shifts because he'd already navigated one. Netflix launched in 1997 as a DVD-by-mail service that disrupted Blockbuster's retail model by eliminating late fees and leveraging the postal system. By 2006, Netflix had 6.3 million subscribers and a profitable, growing business.
Then Hastings cannibalised it. In 2007, Netflix launched streaming — initially as a free add-on for DVD subscribers. The quality was poor, the library was thin, and the broadband penetration necessary for mainstream adoption was years away. Every financial metric argued against prioritising streaming over the proven DVD business.
Hastings was reading the paradigm, not the metrics. Bandwidth costs were declining at roughly 30% annually. Hard drive storage costs were declining even faster. The physics of content delivery — once measured in plastic discs, postage, and warehouse square footage — was converging toward bits, fibre, and server farms. The old paradigm (physical distribution of content) was being replaced by a new one (digital distribution of content), and the transition point was predictable from the cost curves even though the exact timing wasn't.
The 2011 Qwikster debacle — when Hastings tried to separate DVD and streaming into distinct services, triggering an 800,000-subscriber loss and a 77% stock price decline — illustrated the violent discomfort of straddling two paradigms simultaneously. But the strategic logic was correct: Netflix needed to force itself and its customers into the new paradigm before a competitor (Amazon, Hulu, eventually Disney) built the streaming infrastructure first. By 2023, Netflix had 260 million streaming subscribers globally. The DVD business that once seemed indispensable had been quietly shut down. Hastings didn't predict the paradigm shift. He measured the cost curves that made it inevitable, then built for the world those curves described rather than the world that currently existed. The lesson is structural: paradigm shifts driven by exponential cost curves (bandwidth, storage, compute) are the most predictable category because the enabling economics move on a timetable independent of market sentiment or competitive dynamics. The physics of fibre optics didn't care whether Blockbuster's board believed in streaming. The cost curves moved regardless.
Marc AndreessenCo-founder, Netscape; Co-founder, Andreessen Horowitz, 2009–present
Andreessen has built two careers around paradigm shifts. The first was Mosaic, the 1993 web browser that made the internet visual and navigable — transforming the paradigm of computing from local applications to networked services. The second, more relevant to this model, was the thesis he articulated in a 2011 Wall Street Journal essay: "Why Software Is Eating the World."
The essay's argument was a paradigm-shift claim disguised as a technology observation. Andreessen wasn't saying that software companies were growing. He was arguing that the fundamental paradigm of industry — what constitutes a company's core value — was shifting from physical assets and human labour to software and data. Amazon wasn't a retailer with a website; it was a software company that happened to sell physical goods. Netflix wasn't an entertainment company that used technology; it was a technology company that happened to distribute entertainment.
The investment thesis that followed — Andreessen Horowitz's portfolio strategy from 2009 onward — was built on identifying companies positioned at the boundary between the old paradigm (industry-specific expertise, physical infrastructure, human-intensive operations) and the new paradigm (software-defined operations, data-driven decision-making, marginal-cost-zero distribution). Investments in Airbnb, Lyft, GitHub, Coinbase, and Databricks all reflected the same structural bet: that software would replace physical infrastructure as the locus of competitive advantage in one industry after another.
The pattern Andreessen identified is still playing out. The companies that resist the paradigm — "we're a bank, not a technology company" — are systematically losing to companies that embrace it. The ones that internalise the shift — "we're a technology company that operates in banking" — capture the value that the old paradigm leaves on the table.
When Nadella became CEO of Microsoft in February 2014, the company was the canonical example of a paradigm-shift casualty still in denial. Microsoft's paradigm — that computing was organised around desktop operating systems and licensed software — had been the correct framework for two decades. Windows and Office generated the vast majority of the company's profit. Every organisational incentive, every career path, every product decision was structured to defend and extend the Windows franchise.
The problem was that the paradigm had already shifted. Cloud computing, mobile platforms, and open-source software had collectively moved the centre of gravity away from licensed desktop software. Amazon's AWS was capturing the enterprise infrastructure market. Google's Android and Apple's iOS had made mobile, not desktop, the primary computing platform. Linux was displacing Windows Server in data centres. Microsoft, under Steve Ballmer, had responded by trying to force the old paradigm onto the new reality — Windows Phone, Windows 8's tablet-desktop hybrid interface, the $7.2 billion Nokia acquisition. Each was an attempt to extend the Windows paradigm into territory where it didn't belong.
Nadella's insight was that Microsoft needed to abandon the Windows-centric paradigm entirely and reorganise around cloud computing and artificial intelligence. Within months of taking office, he declared Microsoft a "cloud-first, mobile-first" company — a statement that horrified the Windows division because it explicitly demoted their product from the centre of the strategy. He made Office available on iOS and Android — previously unthinkable because it undermined the case for buying a Windows device. He embraced Linux on Azure, open-sourced .NET, and acquired GitHub for $7.5 billion in 2018.
The results speak in the only language that matters. Microsoft's market capitalisation was approximately $300 billion when Nadella took over. By 2024, it exceeded $3 trillion. Azure became the second-largest cloud platform globally. The company's early investment in OpenAI positioned it at the centre of the generative AI paradigm. Nadella didn't predict the paradigm shift. He recognised that it had already happened, accepted the organisational pain of abandoning the old framework, and rebuilt the company for the world that actually existed rather than the world Microsoft wished still existed.
Section 6
Visual Explanation
Section 7
Connected Models
Paradigm shifts don't happen in a conceptual vacuum. They interact with models that explain why incumbents fail, how new frameworks emerge, and what happens to the competitive landscape when the ground shifts. Some models reinforce the pattern, some create productive tension, and some describe where the logic leads next.
Joseph Schumpeter's concept of creative destruction — the process by which new innovations render existing technologies, companies, and business models obsolete — is the economic expression of paradigm shift theory. Kuhn described how scientific frameworks replace each other. Schumpeter described how economic structures replace each other through the same mechanism: incumbents optimised for the old paradigm are destroyed by entrants built for the new one. The digital camera didn't improve film photography. It destroyed the entire value chain — film manufacturing, chemical processing, retail photo labs — and replaced it with a different economic structure entirely. Schumpeter saw this pattern in capitalism's deep structure: the most productive force in a market economy is not competition within a paradigm but the replacement of one paradigm by another. Creative destruction is what paradigm shifts look like when they hit a balance sheet.
First principles thinking is the cognitive tool most likely to identify a paradigm shift in progress. When Elon Musk decomposed rocket costs to raw materials rather than accepting the aerospace industry's inherited pricing, he was doing first principles reasoning — but the output was a paradigm-shift insight. The existing paradigm said rockets cost $65 million because that's what rockets cost. First principles revealed that 98% of the cost was convention, not physics. The two models reinforce each other: first principles provides the analytical method; paradigm shift describes the structural outcome when that method reveals the current framework is built on false assumptions. Most paradigm shifts begin with someone asking a first-principles question that the current paradigm has made invisible.
Tension
Section 8
One Key Quote
"The successive transition from one paradigm to another via revolution is the usual developmental pattern of mature science."
— Thomas Kuhn, The Structure of Scientific Revolutions (1962)
Section 9
Analyst's Take
Faster Than Normal — Editorial View
The paradigm shift model is simultaneously one of the most powerful frameworks in the mental models canon and one of the most abused. Its power lies in explaining a pattern that recurs across every domain — science, technology, business, culture — with structural consistency. Its abuse lies in the fact that every startup pitch deck since 2010 has claimed to represent a "paradigm shift," and 99% of them were describing incremental improvements within an existing paradigm.
The distinction matters more than most people realise. A genuine paradigm shift replaces the questions, not just the answers. The shift from landlines to mobile phones didn't produce "better landlines." It eliminated the assumption that communication is tied to a physical location. The shift from on-premise software to cloud computing didn't produce "better servers in your closet." It eliminated the assumption that companies need to own their infrastructure. When the questions change, the entire competitive landscape restructures — because every incumbent is optimised to answer the old questions, and optimisation for obsolete questions is a liability, not an asset.
The founders I find most impressive are the ones who can articulate precisely which paradigm they're betting against and what specific assumption within that paradigm they believe is false. Not "we're disrupting healthcare" — that's a slogan, not an insight. But "the assumption that diagnosis requires a physician in a room with a patient is false because computer vision on dermatological images now exceeds dermatologist accuracy" — that identifies the specific assumption being broken, the mechanism by which it breaks, and the new paradigm that emerges on the other side.
The corollary for investors: the returns from paradigm shifts are concentrated in the companies that define the new framework, not the companies that are merely early to adopt it. Being an early user of cloud computing didn't create extraordinary returns. Building AWS did.
The timing problem is the hardest part. Kuhn observed that paradigm shifts in science can take decades — plate tectonics took fifty years from Wegener's proposal to mainstream acceptance. In technology and business, the cycles are faster but still slower than founders want them to be. The graveyard of startups is filled with companies that were right about the paradigm shift and wrong about the timing. Webvan was right that grocery delivery would replace physical shopping. It was ten years too early. The bandwidth, logistics infrastructure, and consumer habits weren't ready. Being right about the direction of a paradigm shift and wrong about its speed is a common and expensive failure mode.
Section 10
Test Yourself
Paradigm shift is one of the most misidentified models in business — and the cost of misidentification runs in both directions. Call a genuine paradigm shift "just a trend" and you miss the restructuring. Call incremental progress "a paradigm shift" and you abandon a working framework for something that hasn't actually replaced it.
These scenarios test whether you can distinguish genuine paradigm shifts from incremental progress, hype cycles, and disruption that operates within the existing framework. The key question in each case: are the fundamental assumptions changing, or is the existing framework simply being improved?
Is this mental model at work here?
Scenario 1
A smartphone manufacturer releases a phone with a foldable screen, marketing it as 'the paradigm shift in mobile computing.' The phone runs the same operating system, uses the same apps, and connects to the same cellular networks. It costs $1,800 and sells 3 million units in the first year.
Scenario 2
In 2012, a startup called Moderna is founded on the thesis that messenger RNA can be programmed to instruct human cells to produce any protein — effectively turning the body into its own drug factory. Pharmaceutical incumbents are sceptical, noting that mRNA is fragile, has never been proven in a clinical setting, and that the company has no approved products. Eight years later, Moderna produces a COVID-19 vaccine in two days after receiving the SARS-CoV-2 genetic sequence.
Scenario 3
A retail bank launches a mobile banking app with check deposit, instant transfers, and budgeting tools. The CEO calls it 'a paradigm shift in banking.' Customer adoption reaches 4 million users in the first year, and the bank closes 200 physical branches.
Section 11
Top Resources
The intellectual foundation of paradigm shift theory is unusually concentrated — Kuhn's original work remains the essential text, and the best subsequent treatments apply his framework to specific domains rather than attempting to improve on it. Start with Kuhn for the theory, then move to Christensen and Thiel for the business applications.
The source text. Kuhn's argument that science progresses through revolutionary breaks rather than smooth accumulation remains one of the most influential books in the history of ideas. The chapters on normal science, anomaly, and crisis are directly applicable to business strategy. Dense but not long — under 200 pages. The 50th anniversary edition (2012) includes an introductory essay by Ian Hacking that contextualises the work's influence. Essential for anyone who uses the phrase "paradigm shift" and wants to know what it actually means.
Christensen translated Kuhn's framework into business strategy. His central question — why do well-managed companies fail when disruptive technologies emerge? — is a specific application of paradigm shift theory to competitive dynamics. The case studies (disk drives, steel minimills, mechanical excavators) demonstrate the pattern with empirical rigour. The model of sustaining versus disruptive innovation gives founders a vocabulary for distinguishing paradigm-level shifts from incremental improvements. Read alongside Kuhn for the complete picture.
Thiel's framework for building companies that create genuinely new categories — going from zero to one rather than one to n — is the startup operator's guide to paradigm shifts. His concept of "secrets" (important truths that most people haven't yet recognised) maps directly to the anomalies that precede a Kuhnian crisis. The chapter on competition and monopoly explains why paradigm-shift companies capture disproportionate value: they define a new category rather than competing within an existing one.
Andreessen's Wall Street Journal essay is the clearest articulation of a paradigm shift in progress, written by someone positioning capital to capture it. The thesis — that software companies are poised to displace incumbents across every industry — identified the specific mechanism (marginal-cost-zero distribution, data network effects, platform economics) by which the old paradigm was being replaced. Read it alongside the subsequent decade of evidence: the industries Andreessen identified as vulnerable (retail, media, financial services, healthcare) have restructured largely along the lines he described.
Isaacson's biography documents a founder who has built multiple companies at paradigm boundaries — SpaceX at the boundary between government-monopoly and commercial space launch, Tesla at the boundary between internal combustion and electric propulsion, Neuralink at the boundary between biological and augmented cognition. The granular detail on cost decompositions, engineering arguments, and institutional resistance shows what paradigm-shift execution actually looks like at the operational level. The SpaceX chapters are particularly valuable for understanding how a new paradigm displaces an entrenched one through sustained iteration rather than a single breakthrough.
Paradigm Shift — Normal science accumulates anomalies until a crisis triggers the adoption of a new framework that redefines the entire field.
[Circle of Competence](/mental-models/circle-of-competence)
Circle of competence says stay within domains you understand deeply. Paradigm shifts say the domain you understand deeply may be about to become irrelevant. The tension is sharp: the deeper your competence in the old paradigm, the more you've invested in its assumptions, and the harder it is to recognise when those assumptions are breaking. Kodak's chemical engineers had profound circle-of-competence expertise — in a discipline that digital imaging was rendering obsolete. The resolution isn't to abandon circle of competence but to distinguish between two types of knowledge. Paradigm-transferable knowledge — understanding customer behaviour, capital allocation discipline, organisational design — retains value across shifts. Paradigm-specific knowledge — film chemistry, hardware keyboard design, on-premise server management — becomes a liability the moment the paradigm shifts.
The test: ask whether your expertise explains why something works (transferable) or how the current system works (paradigm-specific). The first type survives the shift. The second type doesn't.
Clayton Christensen's Innovator's Dilemma describes why well-managed companies fail when disruptive technologies emerge — they listen to existing customers, invest in sustaining innovations, and rationally ignore the low-end disruption until it's too late. The tension with paradigm shift theory: Christensen's framework implies that disruption is a predictable, manageable process that companies can navigate with the right organisational design and strategic awareness. Kuhn's framework suggests something more radical — that paradigm shifts create incommensurability, meaning the old organisation literally cannot perceive the new reality because its entire cognitive infrastructure is built on the old one.
Christensen offers hope that incumbents can adapt through separate autonomous organisations. Kuhn suggests they usually can't because the problem isn't structural — it's perceptual. The historical record splits the difference: a few incumbents survive paradigm shifts (IBM, Microsoft, Apple), but they typically do so only through wrenching internal transformation led by an outsider perspective — Lou Gerstner at IBM, Satya Nadella at Microsoft, Steve Jobs returning to Apple. Most incumbents don't survive at all.
Recognising a paradigm shift is the first-order insight. The real value comes from tracing the second and third-order consequences. When the paradigm shifts from physical to digital distribution of content, the first-order effect is obvious: streaming replaces DVDs. The second-order effect is less obvious: content becomes a commodity and original production becomes the differentiator. The third-order effect: the companies that own both the distribution platform and the original content (Netflix, Disney+) capture disproportionate value. Every paradigm shift creates a cascade of restructuring that extends far beyond the initial displacement. Second-order thinking is the tool that maps that cascade.
Leads-to
[Zero to One Theory](/mental-models/zero-to-one-theory)
Peter Thiel's zero-to-one framework — that the most valuable companies create something genuinely new rather than copying what exists — is where paradigm shift theory leads in a startup context. If you accept that paradigm shifts create entirely new categories of value, the strategic implication is clear: build for the new paradigm, not within the old one. Going from one to n — incremental improvement — is competition within an existing paradigm. Going from zero to one — creating a new category — is building at the frontier of a paradigm shift.
Thiel's question — "what important truth do few people agree with you on?" — is essentially asking "what paradigm shift do you see that most people don't?" The zero-to-one insight is a paradigm-shift bet placed before the shift becomes consensus. Every monopoly-creating company that Thiel admires — Google, Facebook, SpaceX — was built on a paradigm-shift thesis that the market hadn't yet priced in. The companies that capture the most value aren't the ones that compete best within the old framework. They're the ones that define the new framework before anyone else realises the old one is ending.
The organisational immune response is real and ferocious. Kuhn documented it in science: the old guard actively resists the new paradigm through institutional mechanisms — peer review, funding decisions, hiring committees, tenure processes. The business equivalent is just as brutal. Middle management in an incumbent company has every incentive to defend the existing paradigm because their expertise, authority, and compensation are tied to it. A shift to cloud computing doesn't just threaten the on-premise software product — it threatens the entire organisational structure built around selling, deploying, and supporting on-premise installations. The people who run that structure will fight the shift long after the rational case for it is overwhelming. Satya Nadella's transformation of Microsoft from 2014 onward was as much a political achievement as a strategic one — he had to overcome the internal immune response of a Windows-centric organisation that had every rational self-interest in preserving the old paradigm.
There's one more dimension worth flagging. The best paradigm-shift operators are not ideologues about the new paradigm. They don't dismiss everything about the old framework. They understand precisely which elements of the old paradigm transfer to the new one and which don't. When Nadella rebuilt Microsoft around cloud computing, he preserved the enterprise relationships, the developer ecosystem knowledge, and the sales infrastructure that had been built over three decades. He discarded the Windows-centricity that was paradigm-specific. The founders who fail during paradigm shifts are often the ones who throw out the baby with the bathwater — abandoning transferable assets in their enthusiasm for the new framework.
My honest assessment: learn this model for pattern recognition, not for prediction. You cannot reliably predict when a paradigm shift will occur. You can learn to recognise the early signals — accumulating anomalies, incumbent dismissal, metrics that stop measuring the right things — and position yourself to move quickly when the shift becomes undeniable. The edge isn't seeing the future. It's seeing the present clearly enough to recognise when the ground has already moved.
Scenario 4
Between 2010 and 2020, the cost of sequencing a human genome drops from $50,000 to under $600. Researchers begin identifying genetic markers for disease risk, drug response, and treatment optimisation at a pace that overwhelms the existing clinical trial framework. Regulatory agencies struggle to evaluate therapies that are personalised to individual patients rather than tested on population averages.
