Vincent A Murphy
How we are witnessing the fastest mass cognitive transition in recorded history
We are living at the twilight of the Printocene.
For roughly five hundred years, print has been the dominant medium through which civilisation met information. It made knowledge feel solid, scarce, and sacral. Books were heavy objects, expensively made and reverently stored. Even when the paperback and the photocopier arrived, the underlying intuition barely shifted: the printed page was still the final word. We scribble in the margin of a newspaper and some small part of us half-expects to be struck by lightning.
That long era is now being overtaken at remarkable speed. The shift from manuscript to print unfolded over decades. The shift from print to generative AI is happening roughly an order of magnitude faster. Within three years of the first widely-available conversational models, it is already increasingly difficult to deny that we are in the early phase of the fastest mass cognitive transition in recorded history.
We have been here before, though not often.
To understand what is happening to text now, we need to understand the rare class of media that caused similar upheavals in the past.
Cognologies: media that rewrite what information is
Most technologies rearrange logistics. A steam engine hauls more coal; a telegraph moves messages faster; a smartphone bundles several tools into one pocket rectangle. Useful, but fundamentally incremental.
A tiny handful of technologies do something stranger. They don’t just move information; they re-define what information feels like in the first place. They alter the mental “default setting” of a civilisation. I call these rare creatures cognologies.
Across human history I argue there have been four:
- Language– the first shared, symbolic medium. Information feels ephemeral and embodied, existing in breath and gesture.
- Writing– marks on clay, wax, parchment. Information becomes externalisable and enumerable; debts, laws, and stories can outlive the speaker.
- Print– industrialised writing. Information becomes mass-produced, stable, and standardised. You can build bureaucracies, universities, and nation-states on that.
- Generative AI– text that answers back. Information becomes soft, suffused, and secular: fluid, abundant, remix-able, and stripped of the mystique that clung to ink.
Each cognology installs a new cultural “operating system” for thinking. Language birthed orality; writing demanded scribal skills; print required mass literacy. Generative AI, as a cognology, will demand its own literacy too – one that looks less like exam prep and more like disciplined play.
To see why, we need to visit the almost-future that came before this one.
Hypertext: the almost-future of text
The twentieth century gave us a preview of what a post-print world might look like. Vannevar Bush imagined the Memex; Doug Engelbart demoed the mouse, windows, and collaborative editing; Ted Nelson named hypertext; Tim Berners-Lee quietly wired the Web together. They were building roads across the desert long before there were many travellers.
Their shared dream was simple and radical: text that links, branches, annotates, and recombines, rather than marching in a single line from margin to margin. Hypertext promised to free us from the tyranny of the linear page. Instead of a book as a sealed container, we could have webs, trails, and living documents.
Culturally, though, we mostly stayed print-brained on glowing rectangles. Web pages became fast pamphlets. Comment boxes became the margins of a very large, frequently angry book. Our mental model of information – solid, scarce, sacral – remained rooted in the Printocene OS.
Hypertext turned text clickable. It did not yet make text conversational.
Generative AI does.
Where hypertext let us jump between fixed sentences, generative models let us negotiate the sentences themselves. We no longer merely traverse text; we co-produce it. That is a different order of shift, and one that demands a different stance from us if we want to stay sane and sovereign.
From Literacy to Ludicity
Every cognology forces a new literacy.
- Orality demanded the arts of memory and rhetoric.
- Writing demanded the slow, specialised training of scribes.
- Print demanded mass literacy so that citizens could read, write, and verify texts.
- Generative AI now demands something new.
Print literacy taught us how to treat books as workbenches, not shrines. You learn to skim, to annotate, to cross-check sources, to move from decoding letters to judging arguments.
AI, as a cognology, asks us to do something parallel but not identical. We are no longer just reading and writing about a world; we are co-steering live flows of generated text.
For that, we need what I call Ludicity.
Ludicity is structured, playful, verifiable collaboration with live, evolving flows of text and thought. It is to AI what literacy was to print: a new default competence that lets ordinary people use a dominant medium safely and powerfully, rather than be used by it.
Two simple reframes capture the difference:
In literacy, the model is the book. It becomes a workbench, not a shrine.
In Ludicity, the model is AI. It becomes a lab partner, not an oracle.
The core competencies actually rhyme:
- Comprehension– not just “what does this text say?” but “what did the system do to produce it?”
- Expression– not just writing essays, but framing problems and prompts that shape the dialogue.
- Evaluation– not just critical reading, but explicit testing, red-teaming, and cross-checking.
- Ethics– not just citation, but disclosure of AI assistance, data consent, and provenance.
Ludicity is the name for practising these on purpose, rather than stumbling into them by accident or outsourcing them to “the AI”.
Of course, an idea this grand needs to get its hands dirty. So I built a small, portable way to make Ludicity practical.
The Ludicity Field Card: literacy for AI sessions
Philosophy is lovely; teams need checklists.
The Ludicity Field Card is a one-page pattern for structuring any serious session with AI. It is deliberately modest: a thing you can run in a classroom, a meeting, or alone at a laptop. Its goal for any given session is simple:
Produce one useful outcome with effective surprise and clean ethics.
It does this through four components.
- The PLAY loop
A short loop that gives shape to “play” so it produces outcomes, not chaos:
- P – Problem Write a one-line brief with audience, constraints, and a clear “done = …”.
- This stops prompt sprawl. It is the AI-era equivalent of setting a research question before you start skimming books.
- L – Layout List 3–5 plausible approaches, with obvious risks and limits. Choose a path.
- This forces option-generation before fixation, like scanning the table of contents and deciding which chapters to read.
- A – Ask Use 2–4 targeted prompts with roles, constraints, and formats. Keep a simple delta log of versions: v1, v2, v3, noting what changed and why.
- Small, testable steps beat one mega-prompt; the log makes learning inspectable.
- Y – Yes-but-verify Define acceptance tests before generation. Verify important claims with at least two independent sources. Disclose AI assistance and provenance when sharing. Play without verification is wishful thinking. Verification turns it into work.
Where print literacy taught drafting, revising, and citing, the PLAY loop teaches prompting, iterating, and verifying.
- Five Pillars (habit cues)
Each session, we cue five habits:
- Intent– clarify purpose before prompting. Who is this for, and why?
- Iteration– prefer visible versions (v1, v2) over invisible tinkering.
- Information Geometry– deliberately switch shapes (table ↔prose ↔code ↔image) to reveal structure and break stuck states.
- Verification– tests, citations, red-team passes on high-stakes claims.
- Ethics– provenance, consent, and data minimisation as default, not afterthought.
These mirror the “study skills” and “editing checklists” of traditional literacy, but tuned for dynamic text.
- KPIs (so leaders can care)
Play is fun; budgets require numbers. Each session lightly tracks:
- Velocity– did cycle time shrink?
- Surprise– did we achieve both novelty and fit (a 4/5 or better)?
- Coverage– are big claims checked against more than one source?
- Rework– hours avoided versus the old baseline?
- Hygiene– is provenance disclosed when we share?
This is a polite way of saying: if you want organisations to invest in better play, show them the ROI.
- Social Contract (quick ethics check)
Finally, a four-point “are we being grown-ups?” list:
- Do we have consent to use these inputs?
- Have we minimised sensitive data?
- Are AI-assisted outputs labelled when shared?
- Has anything high-stakes had a peer or red-team check?
The Social Contract keeps the power humane. You can generate almost anything; that doesn’t mean you should publish it.
Put together, the Field Card turns Ludicity into something you can run on a Tuesday afternoon, not just a noble word in a keynote.
Hypertext inside the Printocene, Ludicity inside the Cognology
At this point, it is worth placing Hypertext and Ludicity on the same map.
Hypertext was the last great innovation inside the Printocene operating system. It took printed ideas of citation, footnoting, and cross-reference and gave them a literal implementation. The blue underlined hyperlink is a mechanical incarnation of “see page 213” and “ibid.” It made text clickable.
Ludicity is the default posture inside the new AI operating system. It takes generative models as a given and asks: how can ordinary people co-create with them without drowning in nonsense or handing them their agency?
One way to think about the shift is in those three adjectives:
- Printocene / Hypertext– information feels stable, bounded, somewhat sacred. Scarcity lives in page count and shelf space. Authority is linked to publication.
- AI / Ludicity– information feels fluid, abundant, secular. Scarcity lives in attention, judgment, and compute. Authority is linked to behaviour: how you test, attribute, and interact.
- Hypertext gave us better maps of a fixed landscape. Ludicity assumes the landscape itself is in motion and trains the traveller accordingly.
What then of the future of text?
All of this raises the question the symposium itself is asking: what becomes of text now?
If we look back, the pattern is sobering. After Gutenberg, it took decades before print was widely used for anything beyond reproducing familiar manuscripts and religious works. For a considerable time, people simply poured old wine into new bottles. Only later did the textbook, the newspaper, the scientific journal, the cheap novel, and the bureaucratic form emerge as native print genres.
We should expect something similar now. For a while, we will keep using AI to generate press releases, student essays, and fake PDFs that look suspiciously like the old world. But our children will not experience it that way. For them, interactive, generative text will be the default, and print will be the strange, solid relic.
They will assume text is editable, negotiable, re-rollable – that information obeys them, not the other way round.
Our task, as late-Printocene natives, is twofold:
- Steward the elder cognology.
Orality, handwriting, print – these remain extraordinary achievements. A handwritten letter still carries an intimacy that no generated paragraph can touch. A well-made book still commands a kind of quiet that a glowing rectangle struggles to maintain. These older media deserve care, not dismissal. - Teach Ludicity as the new default.
We do our children no favours by treating AI as an aberration or a passing fad. To send them into an AI-saturated world armed only with print-era literacy is to ask them to fight with training swords while the battlefield fills with drones. Ludicity is how we give them live steel and proper safety gear.
Text, in other words, is no longer just a product. It is a behavioural space– a place where humans and machines negotiate meaning in real time. The future of text will be decided less by fonts and file formats than by the habits we cultivate around this negotiation.
Coda: awe, bewilderment, and recognition
To live at the hinge of cognologies is to be a little stunned.
It is natural to feel awe in the old sense of the word: transfixed in the presence of something that feels god-like. It is natural to feel bewildered – literally, to feel lost in a wood. We are wandering in a forest whose trees are made of text that moves.
But we should remember two things.
First, we have walked this path before. Our ancestors who watched the first printed Bibles roll off the press felt similar vertigo. The monks watching their world of scriptoria shrink must have felt their skills both devalued and strangely elevated. We are their descendants, not an entirely new species.
Second, the angel and the demon in this story are not nestled inside the machine. They are nested inside us. Generative AI is a mirror with a very fast refresh rate. It amplifies our capacity for kindness, cruelty, curiosity, and laziness alike. No amount of code can save us from the work of recognising which is which.
Ludicity, at heart, is a modest proposal:
that we meet this new cognology the way our wiseancestors met the last one
with play, with rigour, with ethics, and with the absolute stubborn refusal
to hand our judgment to anything that does not bleed.
If we can do that, the twilight of the Printocene need not be a funeral.
It can be a strange, luminous dusk, one in which we honour thecrepuscularity of theold light,even as we learn how to see by the dawning of the new.