Advice to the CEO in an age of AI
- Malcolm Maxwell
- May 4
- 5 min read

The Age of Connected Ideas
Why the curious generalist—armed with AI—is finally ready to inherit the earth.
Glance around the room.
Not dramatically. No need to stare at the kettle as if it houses Alan Turing’s ghost. But notice the objects: the laptop, the phone, the light switch, the router blinking in the corner like a tiny, anxious priest of modernity. Each seems simple because it has been made simple at the point of use. Press the button. Open the app. Ask the question. Receive the answer.
Yet behind each object stretches a chain of dependencies so intricate that no single human understands it end to end. The chip designer does not run the undersea cable. The undersea cable engineer does not write the compiler. The compiler engineer does not manufacture the lithium battery. And the person using the device is trying to remember where they left their glasses…
This is modern life: we are surrounded by systems we can operate but cannot fully understand.
The Hidden Complexity
Artificial intelligence does not reduce that complexity. It adds a new layer. But it also does something far more interesting: it gives us, for the first time, a practical interface to the connections between those fragments.
That is the shift.
The lazy version of the AI story claims that machines now answer questions. That is true, but not important. Answering questions was what we asked of search engines, consultants, analysts, interns, and whichever unfortunate colleague made eye contact in a meeting. Answers are useful. Answers are not transformation.
The deeper change? AI connects fragments.
A thought from agriculture. A pattern from software. A metaphor from biology. A regulatory constraint from finance. A user complaint buried in a support ticket. A half-forgotten article about attention. A design principle from architecture. Alone, each is just information. Together, arranged correctly, they become insight.
AI is not merely another productivity tool. It is a connection engine.
The Specialist’s Prison
For most of the industrial era, knowledge work rewarded specialisation—and rightly so. The world grew too complicated for general comprehension. Someone had to know tax law. Someone had to know transformers. Someone had to know warehouse logistics, trade settlement, shipping workflows, or the strange internal politics of enterprise procurement, may God have mercy on them all.
Specialisation was the only rational response to complexity.
But it has a weakness. It sees deeply and narrowly. It optimises within categories. It improves the canal, the plough, the relay, the spreadsheet, the workflow. It rarely asks what happens when the canal changes distribution, which changes the calendar, which changes what time people agree is ‘now’, which then changes laws, which changes the shape of civilisation.
That kind of question belongs to the generalist.
The Useful Generalist
I should be precise. I do not mean the vague generalist, the conference-panel generalist, the person who has skimmed three airport books and now has a framework. We have all met this person. Some of us have been this person. There should probably be a helpline.
The useful generalist is different. Not shallow, but wide. They collect domains the way engineers collect cables they will definitely need one day. They are comfortable being temporarily stupid in unfamiliar territory. They ask basic questions without embarrassment. They notice when two fields use different words for the same problem. They do not know everything. They know how things might connect.
Until recently, this was a difficult posture to monetise. Organisations love generalists in theory and punish them in practice. Job descriptions demand ten years of experience in a tool that has existed for four. Departments defend their boundaries. Specialists speak in dialects. The person who moves between worlds is often treated as insufficiently committed to any one of them.
AI Changes the Economics
AI changes the economics of that movement.
A generalist with AI can enter a new domain faster than before. Not master it—that distinction matters. Mastery still requires time, feedback, practice, and the occasional humiliating correction from someone who has actually done the work. But entry becomes cheaper. Translation becomes faster. Pattern comparison becomes routine. The generalist can now ask: what is the equivalent of technical debt in law? What does supply-chain resilience teach us about software architecture? What does neural diversity reveal about idea generation when more formal function is no longer the bottleneck?
That last point matters more than it first appears.
The old world often punished associative thinking. The person whose mind jumped from one idea to another was told to focus, specialise, stay in lane, finish the deck. Sometimes that advice was correct. Many unfinished decks have died so that civilisation may live.
But associative thinking is also the source of many useful connections. It is the mind noticing that two unrelated things rhyme. AI gives that mind a stabilising partner. It can capture the fragment, hold the thread, organise the mess, test the analogy, produce the outline, and return the thinker to the path without demanding that the path be straight.
This is not magic. It is not therapy. It is not the arrival of a benevolent robot butler with a clipboard and superb emotional boundaries. It is simply a new cognitive arrangement: human curiosity coupled to machine-scale retrieval, synthesis, and reformulation.
The result? The scattered thinker, the curious generalist, the person with too many tabs open both in the browser and in the soul, suddenly has leverage.
The Trap of Elegant Nonsense
But leverage is dangerous when confused with wisdom.
AI can connect ideas badly as well as brilliantly. It can produce elegant nonsense. It can flatten expertise into plausible mush. It can make the incompetent feel fluent and the fluent feel omniscient. This is why the generalist who thrives will not be the one who merely prompts. Prompting is table stakes—and not at an interesting table even then.
The successful generalist will verify.
They will cross-check. They will ask experts. They will understand that a connection is a hypothesis before it is a conclusion. They will treat AI outputs as maps, not territory; accelerants, not authorities. They will know when to zoom out and, just as importantly, when to shut up and let the domain expert speak.
The New Bottleneck
The future, then, does not belong to specialists or generalists in the crude sense. It belongs to people who can move between the two modes: broad enough to see the network, disciplined enough to respect its parts.
This is the organisational problem now emerging.
Most companies are still asking where AI can save time. That is a reasonable question, but not the most important one. Saving time is the first-order effect. The second-order effect is changing who can participate in problem-solving. The third-order effect is changing which problems are visible in the first place.
When ideas become easier to connect, the bottleneck moves.
It’s no longer access to information. It is not even access to expertise, though expertise remains precious. The bottleneck is sense-making. Who can look across the blinking room of modernity and ask what these objects are doing to us just because they are there? Who can see that the relay, the plough, the calendar, the chatbot, and the distracted mind are part of the same story?
The generalist can. Not because they know more, but because they are willing to look sideways.
Imagination Becomes Operational
The question is not whether AI will replace expertise. It will not—except where the expertise was mostly expensive formatting. The question is whether expertise, once made more accessible, becomes usable by people who can connect it to other things.
That is where the advantage moves.
The world is not changing because AI can write emails, summarise meetings, or generate suspiciously cheerful strategy documents. It is changing because AI makes connection cheap. And when connection becomes cheap, imagination becomes operational.
The winners will not be those who know one thing and defend it to the death.
They will be those who can gather many things, test the links, discard the nonsense, and build something useful from the pattern.
The future belongs to the generalists.
Which is inconvenient for tidy org charts, but rather good news for the rest of us.




Comments