Leadership Intelligence

Governing AI: Why Governance Literacy Matters More Than Technical Expertise

Tue, 07 Apr 2026 00:00:00 GMT

Governing AI: Why Governance Literacy Matters More Than Technical Expertise

On 1 January 2026, Manage My Health notified the Privacy Commissioner of a cyber incident affecting patient records. On 1 May 2026, IPP 3A comes into force, requiring organisations that collect personal information indirectly to notify those individuals. On 3 August 2026, the Biometrics Processing Privacy Code grace period expires.

Three regulatory events in eight months. None of them require your board to understand machine learning. All of them require your board to govern properly.

That's the distinction most boards are missing. The gap isn't technical literacy — it's governance literacy.

The stakes are moving faster than the agenda

New Zealand boards aren't being asked to become data scientists. They're being asked to do what they've always done — manage risk, ensure compliance, steward organisational value. The difference now is that the risks move faster, the compliance landscape shifts under your feet, and the consequences scale at algorithmic speed.

The upside scales too. International research suggests nearly 60% of executives found that investing in responsible AI practices improved both return on investment and innovation performance. But fewer than half have formalised AI governance frameworks. That gap between opportunity and oversight is where the damage happens — and where the value is lost.

Consider IPP 3A alone. If your organisation collects personal information from a third party — through data-sharing agreements, AI-powered research tools, referral processes, or even a CRM that enriches contact records — you'll need to demonstrate that the individual knows about the collection, its purpose, and who's receiving it. That's not a technical problem. That's a governance problem, and it needs board-level attention before May.

What boards consistently get wrong

Most AI governance failures follow predictable patterns.

Delegating it to IT. Many boards treat AI governance as a technical function — hand it to the data team, tick the box. But when an algorithm denies a loan on biased data, or a generative AI tool leaks commercially sensitive information, regulators and shareholders don't call the IT manager. They call the Chair. The board owns the risk. Execution can be delegated. Responsibility cannot.

Waiting for regulation to arrive. New Zealand has signalled a light-touch approach to AI regulation and will not introduce AI-specific legislation. The country relies on technology-neutral frameworks — privacy, consumer protection, human rights — updated as needed. Boards waiting for prescriptive rules will be waiting indefinitely. The regulatory pressure comes through Privacy Commissioner guidance, court decisions, and stakeholder expectations, not a single new Act.

Treating it as a procurement decision. Many boards approach AI the way they approach any technology investment — issue a tender, select a vendor, implement. But AI governance sits at the intersection of strategy, ethics, compliance, and culture. It touches multiple committees and functions. Buying a tool is not the same as governing its use.

The board owns the risk. Execution can be delegated. Responsibility cannot.

Ignoring what's already published. New Zealand released its first national AI Strategy in mid-2025, accompanied by Responsible AI Guidance for Businesses, adopting the OECD's AI Principles — inclusive growth, human rights, transparency, robustness, accountability. The Privacy Commissioner published guidance on AI and the Information Privacy Principles back in 2023, including specific consideration of te ao Māori perspectives on privacy. Any board not actively engaging with these frameworks is operating without a map.

What effective governance actually looks like

Three characteristics separate boards that are governing AI from boards that are being governed by it.

Clear ownership. There is no ambiguity about who owns AI governance at the executive level. A designated committee — whether technology, risk, ethics, or a standing subcommittee — reports directly to the board. Accountability is personal. If something goes wrong, there is no confusion about who answers for it.

Continuous oversight. Many boards treat AI like capital expenditure — approve once, assume it's handled. But models drift, data quality degrades, regulations shift, and competitive positions change. Effective boards receive regular, forward-looking updates on AI initiatives: performance, emerging risks, escalation triggers, and changes in the regulatory or competitive landscape. Point-in-time approval is not governance.

Integration, not isolation. AI governance is not a separate track bolted on to the existing agenda. It is woven into how the board thinks about strategy, risk, ethics, compliance, and management performance. Questions about AI belong on the risk register, in strategy discussions, and in how management is evaluated.

The Privacy Commissioner has been explicit about what this looks like in practice: conduct Privacy Impact Assessments before deploying AI tools. Ensure training data is relevant, reliable, and ethically sourced. Test systems for accuracy and bias. Maintain ongoing monitoring and audit processes. This isn't a technical exercise — it's a governance exercise that requires board-level involvement.

A practical starting point

If your board is navigating this shift — or avoiding it because the ground feels unstable — I've built a starting point.

The AI Governance Scenario Checkup asks nine straightforward questions, takes about three minutes, and gives you a tiered read on where your governance posture sits — reactive, managed, or strategic. The output is a one-page summary you can bring straight into your next board agenda.

No jargon. No sales pitch. Just a clear signal of whether you're governing AI or being governed by it.

Take the checkup → (free, no email required)

The Question Worth Sitting With

If your board had to explain its AI governance posture to a regulator tomorrow — could it?

Not whether you've adopted the latest tool. Not whether your IT team is capable. Could you articulate your governance structure, demonstrate what risks have been identified and how they're managed, and show what the board is doing to ensure compliance with privacy, fair trading, and human rights principles?

If those answers don't come easily, you have a governance gap. And in an environment where IPP 3A takes effect on 1 May, the Biometrics Code grace period ends on 3 August, and the Privacy Commissioner is actively investigating major breaches — that gap is getting expensive.

The invitation is to start now. Before the regulator makes it mandatory. Before the breach happens. Before the market punishes you for inadequate oversight.

Governance isn't about slowing innovation. It's about making innovation sustainable, trustworthy, and yours to control.

What 10 Cups of Coffee Can Teach You About Data

Sat, 28 Mar 2026 00:00:00 GMT

What 10 Cups of Coffee Can Teach You About Data

Imagine ten people walk into a café. Each orders a coffee. Now imagine a data analyst walks in behind them. Suddenly those ten cups aren't just beverages — they're a dataset. And depending on what question you ask, you'll get ten different answers from exactly the same ten cups.

This is the quiet superpower of descriptive statistics: the data doesn't change, but your perspective on it does.

The Ten Questions You Could Ask

The Basics

1. Average caffeine per cup across the whole group Add up the caffeine in all ten cups and divide by ten. That gives you the mean caffeine per cup. It’s useful for estimating the overall level of caffeine in the group, but it can hide big differences between individuals — a triple-shot espresso and a decaf both get folded into the same average.

2. Average caffeine per person This only differs from the first question if people had different numbers of coffees. If each person has one cup, then average caffeine per person is the same as average caffeine per cup. If some people had two coffees and others had one, then the distinction matters — one is a property of the drinks, the other is a property of the drinkers. One is a product metric, the other is a consumption metric.

3. Average millilitres of milk per cup Now we're looking at a specific variable — milk. This is still a mean, but now we're looking at one ingredient rather than the drink as a whole. Note: if three people had no milk at all, do you include their zeros in the average? You should — but many people don't, which produces "average ml of milk from milk-based coffees only." That's a different question with a different answer.

4. Average cup size A classic mean calculation. But here's the thing: is a 480ml Venti Starbucks monster really "just a bit above average" compared to an 80ml espresso? Averages flatten variation. This is where the next two measures earn their keep.

5. Min/Max (Range) The smallest cup to the largest. Range tells you how spread out your data is — and sometimes that spread is the most interesting finding. A room where the smallest coffee is 80ml and the largest is 600ml tells a very different story than a room where every cup is between 200ml and 250ml.

The Patterns

6. Most frequently ordered type (Mode) If six out of ten people ordered a flat white, the flat white is your mode — the most common value in a categorical dataset. Mode is often overlooked, but it's the most useful measure when you're making decisions like "what should we put on special today?" Mean and median can't answer that for nominal categories like coffee type. Mode can.

7. Most common size AND type (Joint frequency pattern) Now you're cross-referencing two variables: size and type. This is where simple stats tip into pattern analysis — you're not just counting, you’re asking which combination occurred most often. "Most people ordered a medium flat white" is a richer insight than either the size or type alone.

8. Count with and without milk Binary categorisation — yes/no, milk/no milk. Simple count data. This might seem trivial, but in a real dataset it becomes the basis for segmentation: milk drinkers vs. black coffee drinkers might have completely different preferences on everything else too.

9. Count of venues purchased from Interesting shift here. Now you're not analysing the coffee — you're analysing the source. If all ten cups came from one café, that's a very different finding than if they came from ten different places. This is contextual information — data about the circumstances of the data, not the data itself. Easily overlooked, often crucial in how you interpret the results.

The Point

10. Is one of these yours? And here's where it gets personal. All the statistics in the world don't tell you which cup to pick up. Statistics can describe the group, summarise patterns, and support judgment to inform your decisions, but they can’t make the decision for you. The moment you ask "which one is mine?" you've moved from analysis to action, and that's a step only a human can take.

Why This Matters Beyond Coffee

Every one of those ten questions used a different lens on identical underlying data. This is exactly why data literacy matters — not just knowing how to calculate a mean, but knowing which measure to reach for depending on what decision you're trying to make.

Get the question wrong, and the right answer is useless. Ask the right question, and even ten cups of coffee can tell you something worth knowing.

Want to go deeper? Check out my review of Everydata — a whole book built on this exact idea.

Drinking from the Hydrant?

Mon, 23 Mar 2026 00:00:00 GMT

In Groundhog Day, Bill Murray's character lives the same day on repeat, and spends most of it reacting, consuming, and going nowhere. It takes him thousands of cycles before he realises the point isn't to get through the day faster. It's to get through it better. Most of us are living our own version of this with information. We wake up, open the hydrant, and handle it the way a golden retriever handles a garden hose: mouth wide open, soaked through, retaining almost nothing, and then do it all again tomorrow. The AI era has turbocharged this dynamic. We now have tools that can summarise a book in thirty seconds, generate a briefing on any topic on demand, and surface answers before we've even finished forming the question. The result is a generation of leaders who are extraordinarily well-informed by the minute and deeply unwise by the decade.

The Illusion of Learning

Supposedly 90% of all data ever created was generated in just the last two to three years. That's the hydrant. And the goldfish meme captures how we cope with it. We scroll, we skim, we screenshot. We collect interesting fragments, a stat about AI adoption rates, a quote from some futurist, a framework someone posted on LinkedIn, and we mistake the collection for understanding. It feels like learning. It has the dopamine signature of learning. But intellectual trinkets are not the same as integrated knowledge, and knowing about something is not the same as knowing when to apply it, when to challenge it, or when to set it aside entirely.

T.S. Eliot sharpened the same thread in The Rock: “Where is the wisdom we have lost in knowledge? Where is the knowledge we have lost in information?” The hierarchy is clear — information is the raw material, knowledge is information organised by experience, and wisdom is knowledge tested by time and judgement.

Socrates would have recognised this instantly. In the Phaedrus, he argued that the written word would produce “the appearance of wisdom, not true wisdom” — students who had “heard many things but learned nothing.” Replace “written word” with “AI-generated summary” and the critique is startlingly current. Most of what AI delivers, and most of what we consume, never makes it past the first rung.

Efficiency Is Not Effectiveness

Peter Drucker saw this coming decades ago. His most quoted line may be the most ignored: “There is nothing so useless as doing efficiently that which should not be done at all.” His entire body of work returns again and again to the distinction between efficiency and effectiveness — between doing things right and doing the right things. Most AI productivity advice is squarely about the former.

Drucker also observed that “the most important thing in communication is hearing what isn’t said,” a skill that requires the kind of slow, reflective attention that speed-optimised workflows systematically eliminate. And his insistence that “knowledge has to be improved, challenged, and increased constantly, or it vanishes” is a direct rebuke to the prompt-and-forget cycle that passes for learning today.

W. Edwards Deming, the father of quality management, drew the line between quantity and quality even more starkly. He insisted that “in God we trust; all others must bring data,” but he also insisted that “the most important things cannot be measured.” The tension between these two statements is exactly the tension most organisations ignore: the obsession with quantifiable output at the expense of qualitative and relational depth.

As Alvin Toffler observed, “The illiterate of the 21st century will not be those who cannot read and write, but those who cannot learn, unlearn, and relearn.” The issue is no longer access to information, it is the capacity to metabolise it.

The Missing Infrastructure

What’s largely missing from the conversation about AI productivity is any serious attention to knowledge management: the deliberate practice of capturing, structuring, and connecting what you learn so that it compounds over time rather than evaporating. A second brain, a working journal, a note system that reflects how you actually think rather than how a template designer imagined you might.

Tiago Forte calls this "building a second brain" — an external system that captures what you learn so your biological brain is freed to do what it actually does best: synthesise, connect, and create. Without one, every insight you encounter is a Post-it note in a hurricane. Sönke Ahrens makes the same case through the lens of the Zettelkasten method pioneered by the sociologist Niklas Luhmann, who published over 70 books and 400 articles across disciplines — not because he was a faster reader, but because he had a system that turned every note into a node in a growing web of connected thought. The lesson isn't about the method. It's that the infrastructure of thinking matters as much as the thinking itself.

Cal Newport argues in Deep Work that the capacity for sustained, undistracted concentration is becoming both rarer and more valuable at exactly the same time — a "superpower" in an economy that rewards shallow responsiveness. The person who can sit with a hard problem for two hours without checking their phone is increasingly the person who produces work that actually matters. Yet almost nothing about how we adopt AI tools is designed to protect or cultivate that capacity. If anything, the opposite: AI makes it easier to produce mediocre output on autopilot, and harder to justify the discomfort of going deeper.

Greg McKeown frames it as a discipline problem in Essentialism: "If you don't prioritise your life, someone else will." The same applies to your learning. Without a deliberate system for deciding what to retain, connect, and act on, you default to whatever the algorithm serves next. Viktor Frankl, writing from the most extreme circumstances imaginable, arrived at a complementary insight: meaning is not found in the volume of experience but in the depth of engagement with it. A life crammed with information but empty of reflection is, in Frankl's terms, a life that has missed the point entirely.

The leaders and practitioners who will build genuine, durable capability in this era won’t be the ones who used AI to save the most hours. They’ll be the ones who used those hours to reflect, to synthesise, to grow together with others, and who built the systems and habits that turned individual learning into shared capability: more garden than weed field.

The Question Worth Sitting With

Here’s the quiet irony: the entire pitch of AI productivity is time-saving, yet almost no one asks the obvious follow-up: saved for what?

If the time you recover from a faster workflow goes straight back into more consumption, more scrolling, more prompt-and-forget cycles, you haven’t become more capable. You’ve just become busier at a higher RPM.

Seneca wrote nearly two thousand years ago: “It is not that we have a short time to live, but that we waste a great deal of it.” Nassim Nicholas Taleb reportedly said: “The three most harmful addictions are heroin, carbohydrates, and a monthly salary,” to which we might now add the fourth: the illusion of productivity through information consumption.

The quality of a life is not measured by how efficiently it runs, but by what it accumulates, contributes, and eventually understands. Drinking faster from the hydrant is not the same as being less thirsty.

If this resonates —if you're thinking about how to build systems for durable learning and leadership capability rather than just faster output (though faster output has its place)— I'd welcome the conversation. Reach out.

“Everydata: The Misinformation Hidden in the Little Data You Consume Every Day”

Tue, 24 Mar 2026 00:00:00 GMT

Everydata: The Data You Misread Every Day

We consume roughly 30 gigabytes of data daily — weather forecasts, stock reports, news headlines — yet most of us have no framework for interpreting it. Everydata by John H. Johnson and Mike Gluck makes the case that understanding "little data" matters far more than obsessing over big data.

The Book's Core Argument

The real risk isn't a lack of data — it's misreading the data we already have. Johnson and Gluck walk through a series of compelling case studies that expose how easily smart people get it wrong:

The Challenger Disaster — engineers assessed O-ring failure risk using an incomplete dataset, excluding flights with no failures. The data looked safer than it was.

Correlation ≠ Causation — pregnant women avoiding caffeine, or house prices rising near Starbucks locations, are classic examples of confusing association with cause.

Outliers distort averages — Bill Gates walking into a room doesn't make everyone richer; it just changes the mean. A $1B jury verdict driven by outlier data is a real-world consequence of the same error.

Self-reported data is unreliable — 55% of Americans believe they're smarter than average; 93% think they're better drivers. Both are statistically impossible.

Sampling matters — Minnesota recalled 3.7 million food products not because it had the most food poisoning, but because it was best at reporting it. The sample wasn't representative.

The Ladder of Inference

One of the most useful frameworks in the book is Chris Argyris' Ladder of Inference — a seven-rung model describing how we move from raw observable data to firm beliefs and actions, often leaping several rungs without noticing:

Observe reality and facts

Select data (subconsciously filtering what we notice)

Interpret meaning

Make assumptions

Draw conclusions

Adopt beliefs

Take action

The danger is the reflexive loop at the bottom: our existing beliefs shape which data we select next, reinforcing themselves. Good data literacy means consciously stepping down the ladder before acting.

Why This Still Matters

If anything, the problem is getting worse. Greater computing power and larger datasets will likely increase reliance on estimation techniques and statistical modelling — which means more opportunities to misread results, not fewer. The strategic advantage goes to people who understand not just what data says, but what it can't say.

A few practical questions worth asking whenever you encounter data:

Is this a mean or a median? (They tell very different stories.)

What's the sample size, and how was it collected?

Is the claimed relationship causal, or just correlated?

What data might be missing from this picture?

Am I reading an anecdote as if it were evidence?

Verdict

Everydata is an accessible, example-rich read that doesn't require a statistics background. It won't make you a data scientist — but it will make you a more sceptical consumer of the data you encounter every day. That's worth more than most people realise.

Equations Gave Us the Laws. AI May Be Writing New Ones — And We May Not Be Able to Read Them.

Wed, 25 Mar 2026 00:00:00 GMT

💡

Newton didn't know his calculus would one day train weapons systems. Einstein didn't intend E=mc² as a bomb blueprint. The most powerful mathematical ideas in history share one feature: their creators lost control of them. We appear to be doing it again, only faster, and with less excuse. This time, we can see it coming. Does that change anything?

A Pattern Across History

Consider the arc. Newton needed a century to reshape science. Hinton needed a decade to reshape everything else. The people involved shrank even as the impact exploded — from centuries of slow scholarly diffusion to three people triggering a trillion-dollar industry in five years.

Something fundamental changed around 1990: the internet collapsed the adoption lag, and it hasn’t recovered.

The Acceleration Timeline

The pattern across history is stark. Every generation, world-changing mathematical ideas diffuse faster. Newton’s law of gravitation (1687) took 50–100 years to reshape science. Maxwell’s equations (1865) took about 30 years to spawn the telegraph and radio industries. Shannon’s information theory (1948) took 20 years to transform global computing. Backpropagation (1986) took 25 years to reach dominance — and then AlexNet (2012) triggered a trillion-dollar industry in just five years.

The equations AI runs on — Bayes, Shannon, gradient descent — were each world-changing in their own right. AI doesn’t replace them. It weaponises them simultaneously, at scale.

The Shift Nobody Talks About Enough

Every equation in the classical list shares one property: it is legible. E=mc² fits on a coffee mug. Maxwell’s equations fill half a page. A reasonably educated person can, with effort, understand what they mean and why they’re true.

AI is categorically different. Modern language models contain billions of parameters. No single equation governs their behaviour. Their outputs are sometimes surprising to the people who built them. We are, for the first time in history, deploying tools we cannot fully read.

This is not a small thing. It represents a genuine epistemological rupture — a shift from understanding our tools to merely deploying them and managing the fallout.

The Error Problem

History also reminds us that smart people, working within established frameworks, can get it catastrophically wrong, and entrench the error for generations.

Geocentrism persisted for 1,400 years. Phlogiston theory lasted a century before Lavoisier dismantled it with a set of scales. Continental drift was mocked for 50 years because Wegener was a meteorologist, not a geologist. Barry Marshall had to drink bacteria to prove ulcers weren’t caused by stress, then wait for a Nobel Prize.

The mechanism is always similar: institutional inertia, credentialism, and the career costs of dissent keep bad ideas alive long past their use-by date. Max Planck observed it bleakly: “Science advances one funeral at a time.”

The uncomfortable question for our moment: what are we currently wrong about with AI, and how long will it take to find out?

The Risks Worth Naming

Speed without comprehension. We are scaling AI capability faster than our ability to understand, govern, or course-correct it. Every historical equation had a lag between discovery and consequence. That lag gave society time to adapt. We may have lost that buffer.

Deployment without legibility. We can’t audit AI decisions the way we can audit an equation. This matters enormously in medicine, law, finance, and anywhere consequential decisions are made.

Concentration of influence. Newton’s equations diffused slowly through academia. AI capabilities are diffusing through a handful of companies, with enormous leverage over global systems, at speed.

The Wegener risk. Valid critics of current AI assumptions — particularly around whether scaling alone leads to general intelligence — are sometimes dismissed on credentialist or financial grounds. History suggests we should be paying attention to the outliers.

The Opportunities Worth Naming

AI is discovering equations, not just using them. DeepMind’s AlphaGeometry and symbolic regression tools are beginning to find mathematical relationships humans missed. We may be approaching a moment where AI accelerates the very process this whole history describes.

Compression of the knowledge lag. The time between insight and application is collapsing. In medicine, materials science, and climate research, that could be genuinely lifesaving.

Democratisation of capability. Backpropagation was obscure for 25 years. Today, a student in Tauranga has access to tools that would have seemed like science fiction to researchers a decade ago. That’s not nothing.

The Question Worth Sitting With

Every equation in history solved one problem and created another nobody anticipated. E=mc² gave us nuclear power and nuclear weapons. Black-Scholes gave us derivatives markets and the 2008 financial crisis. Perhaps all tools reveal something about those who wield them — and those who misuse them.

AI is not an equation. It’s something more like a universal equation-finder — and we don’t yet know what it or we will find, or what those findings will cost.

The history of world-changing ideas suggests we should be neither naively optimistic nor paralysed by fear. But it does suggest we should be paying very close attention, and that the people most worth listening to are often not the ones with the most to gain.

None of this requires you to become an AI expert — especially not in isolation. It does require intellectual and ethical honesty, both about what you’re assuming, and the people you’re impacting.

If you’re being told not to rock the boat, ask yourself if the people reassuring you have more invested in the answer than in the question. The people most likely to tell you what you're getting wrong are usually the ones with the least to lose by saying it. Find them. Listen to them. Before the answer becomes obvious in hindsight.

The AI Scenario Checkup: Nine Questions, No Jargon

Mon, 23 Mar 2026 00:00:00 GMT

Most business owners and leaders I talk to aren't debating whether AI matters — they've moved past that. The question keeping them up at night is more specific: Is this relevant to my business? Where do I actually start? What's the real risk if I get this wrong?

Those questions deserve a straight answer. So I built one.

What Is the AI Scenario Checkup?

It's a free, plain-English planning tool. Nine questions. Seven real-world scenarios. About three minutes of your time.

No jargon. No consultant-speak. No sales pitch at the end.

What you get is genuinely useful:

Your primary AI scenario — an honest picture of where your business sits right now

A secondary scenario if you're sitting close to two positions

Two watch-outs specific to your situation

Your top three priorities for the next 3–18 months

A plain-English read on the financial upside and downside

A one-page summary you can act on immediately

Who Is It For?

Business owners, organisational leaders, and board members who are genuinely trying to figure out whether AI is worth the bother — and what to do about it if it is. Whether you're already running AI tools or still deciding if you should, you'll leave with something concrete.

The checkup is grounded in real-world research: international security incident data, cybersecurity benchmarks, and productivity findings across professional services, retail, trades, and health. Your results reflect what's actually happening for businesses like yours in 2026 — not boilerplate advice written for enterprise clients with unlimited budgets.

A Starting Point, Not a Substitute

This tool won't replace a proper conversation with someone who knows your business well. But it will make that conversation significantly more valuable. Come in informed, not cold.

And just so we're clear: no drip campaigns, no follow-up calls, no sales sequence. Just a useful tool, built by a management consultant who depends on straight dealing.

→ Start the free AI Scenario Checkup Takes about three minutes · No obligation · No sales pitch, ever

Leadership Intelligence

Governing AI: Why Governance Literacy Matters More Than Technical Expertise

Governing AI: Why Governance Literacy Matters More Than Technical Expertise

The stakes are moving faster than the agenda

What boards consistently get wrong

What effective governance actually looks like

A practical starting point

The Question Worth Sitting With

Further Reading

What 10 Cups of Coffee Can Teach You About Data

What 10 Cups of Coffee Can Teach You About Data

The Ten Questions You Could Ask

Why This Matters Beyond Coffee

Drinking from the Hydrant?

The Illusion of Learning

Efficiency Is Not Effectiveness

The Missing Infrastructure

The Question Worth Sitting With

Further Reading

“Everydata: The Misinformation Hidden in the Little Data You Consume Every Day”

Everydata: The Data You Misread Every Day

The Book's Core Argument

The Ladder of Inference

Why This Still Matters

Verdict

Equations Gave Us the Laws. AI May Be Writing New Ones — And We May Not Be Able to Read Them.

A Pattern Across History

The Acceleration Timeline

The Shift Nobody Talks About Enough

The Error Problem

The Risks Worth Naming

The Opportunities Worth Naming

The Question Worth Sitting With

Further Reading

The AI Scenario Checkup: Nine Questions, No Jargon