I have always been drawn to the moment when information technologies transform how minds meet. I feel very lucky to live at a time when it seems we are encountering two seismic systems changes. I was 16 in 1993 when internet became available for everyone in the Netherlands through pioneering and aptly named Xs4all. And I was 45 in 2022 when ChatGPT made GenAI available for everyone.
I have worked in the national library of the Netherlands for most of my working life, mostly because this type of change fascinates me. People meeting across time and space, through words. I work in an institution that exists to keep the stories and knowledge of the past available in the present and to pass on the stories and knowledge we have to the future. ‘Without words, who knows humanity’, as an ancient proverb says.
For libraries, keeping up with information technology has been core business. The beautiful Genealogy of culture and technology since 1500, created by Kate Crawford and Vladan Joler (2023) shows this under ‘data organization’. Starting with the printing press, it now sees an explosion of AI datasets and semantic networks. The internet made information available across space. AI is now changing the way we see and use information, with human-mediated access rapidly becoming a smaller part of our information consumption.
Provocatively, I now often say: ‘We used to store knowledge and information in long form texts and create metadata to access that information; now we store information in metadata and graphs and generate long form text to access it again’. Which means: we used to preserve books and other texts and created links to them in catalogues, this made it possible to look for books. But increasingly, we are harnessing the knowledge contained in those books independently of their original content. And using LLM’s to access more information than any one human can process. How knowledge graphs and LLMs might interact to create hybrid intelligence is beautifully expressed by Dutch scientist Frank van Harmelen on the Knowledge Graph Insight Podcast.
When I started working at the National Library in 2001, I was taught the concept of Universal Bibliographic Control. It was a library concept from the seventies: if every country would take care of its own output, together national libraries would carry the knowledge of the world. National Bibliographies showed what every country published. Universal Access to Everything was the complementary concept, implying that with inter-library loan, we would know everything that was published and have a way to access it.
Universal Bibliographic Control assumed the unit of knowledge was the published manifestation, and that completeness was achievable through coordination between institutions with shared norms. Both assumptions are now strained. We now need other concepts to achieve the aim of knowing what there is to know. Might national institutions take the responsibility for knowing what there is to know about their countries and putting that into a knowledge graph? A national knowledge graph as the successor to the national bibliography? And if we connect these together, is the lofty ambition of universal access to everything still possible? Or is the national level no longer relevant in times when the digital realm is not divided geographically?
We live in times of epistemological turbulence. In the past, publishing content was costly, and publishers were gatekeepers providing selection and amplification (as Michael Bhaskar argues in The Content Machine). Libraries trusted that process and preserved what publishers found worth publishing. All these assumptions about knowing what there is to know are strained.
The Gutenberg transition lasted around fifty years, from manuscript to printed text. The products of this era are called incunabula, or ‘cradle books’. The web has not reached that fifty year limit yet and already we see a new challenge to epistemology (what we can know) and ontology (what exists). And also a clear challenge to humanity – whether it matters whether humans are part of the information equation.
Working in a library to me has always been about conversing with those who lived before as much as with the living. ‘Breaking bread with the dead’, as Alan Jacobs put it in his book on how ‘temporal bandwidth’ helps to live a more stable and relaxed life. I call it ‘temporal inclusiveness’. At a time when information technologies are transforming how minds meet, keeping alive this idea of human minds meeting across time and space is what centres my attention.
May 2026