The New Digital Humanism: From the End of Writing to the Age of Reading

For a few months now, I have been wondering what will happen to our writing ability in the coming years with the arrival of Generative AI.

This article is a personal, shared exploration of how Generative AI is changing how we read and how we are ceasing to write. It is an open reflection that touches on the evolution of LLMs, the new role of text, and my hope for reawakening critical thinking.

In short, LLMs, language models that can generate better text than most of us would write, have significantly changed the outlook, especially now that we have "reasoning" models like OpenAI o3, o4, DeepSeek, Grok, Claude, etc. They show us the logical steps they take before answering, and, at least in my case, this is often the most interesting part of their output. Not to mention DeepSearch features that, in effect, produce 10-20 page text 'bricks' at the speed of light.

My starting point is simple: never before have we had so much new, personalized, yet imperfect text to read.

Even as we unlearn how to write, this enables many people to read a large amount of new text that they would never have read before.

How we got here

But before we understand where all this will lead, it's worth considering how we got here.

Once ChatGPT was discovered, students jumped in first, followed by those who work in sales marketing: not necessarily for noble purposes at the beginning but, indeed, with a bit of amazement deriving from the fact that we were finally able to quickly produce written texts without fear of the judgment of those who read them. But, above all, the first models were decidedly imperfect, forcing us to focus on the quality of the text, often transforming us from writers to editors.

Writing becomes accessible to everyone

In the meantime, we realized this was a great advantage for those with writing difficulties, dyslexia problems, grammatical problems from childhood, or physical difficulties due to various pathologies. And that, thanks to these tools, there is a concrete possibility of solving the problem. Today, everyone has the potential to produce a text of a medium-high quality level, have fun with more or less colorful tones, imitate any character and writer, and make ideas stand out through the production of an increasingly good written language. But generating text is not writing.

The change of scale: infinite contents

Generating text is a change of scale in the quantity and quality of content generated from a prompt, a draft, an idea.

LLMs, as we know, are big producers of text for whatever questions we ask. Digital interns, as I like to call them, like to write and produce a lot of content.

By simply writing in an AI Chat, "the cat is": instead of finding a simple answer like "on the table", we find ourselves reading a story and an explanation of felines worthy of the best episodes of Super Quark (for my non-Italian friends: a well-known popular science program that aired in Italy).

For those who are very used to reading, this thing is not noticeable at first. But little by little, it comes out: the LLMs, initially with the excuse of editing and now with the fact that they are increasingly verbose, are pushing more and more people to read more, even without wanting to.

Have we really become readers?

I think there is a cultural paradox that is worth addressing.

I'm pretty annoyed that we have been constantly reminded, when we produce any form of content, that people don't like to read for the last twenty years.

Books and newspapers are no longer sold. We must be brief, with a few clear words. It seems to me that this is a self-fulfilling prophecy: by repeating to ourselves that we do not know how to dedicate time and attention to reading, we are making it true.

The Myth of the End of Reading

I really can't stand hearing this complaint anymore. Being told that 'young people' can't pay attention for more than 3 minutes, that 'you lose the reader' if you write too much. Come on!

Neurobiologically, we are always the same: we are simply more distracted. Instead of feeling sorry for ourselves, we should take note of this and find new ways to reactivate our attention. We didn't all suddenly get sick!

In an age where we have access to so much knowledge, thinking about not reading is like going outside on a sunny day, in a beautiful place... but blindfolded.

I understand that a video is more straightforward, direct, and emotional, and makes more in terms of involvement. But I find depth of thought only in reading.

This is changing even more since we have Deep Search tools enabled by the agentic systems present in ChatGPT, Gemini, Grok, DeepSeek, etc. With these tools, we are talking about searching instead of simply searching, as we did with search engines. We are talking about texts thousands of words long, detailed, structured, and with sources at hand.

And this is a different cognitive work.

Previously, with search engines, you would read two or three titles after a simple search query, with their abstract, and then go to the link, quickly scroll down the page to see if it contained something interesting, and continue.

Here, it is a matter of waiting up to 25-30 minutes to have a long and elaborate text that contains, presumably, accurate, new, structured information with possible links to sources. Talk about abundance!

It's something we've never had the chance to experience before, and it has the potential to profoundly change the way we learn.

My New Way of Learning with AI

I have to say that in the last three months, I have changed a lot in the way I inform myself, study, and learn.

Now, I work with one or more Deep Search tools for every minimally important topic. I think of a good prompt and start with models like o3 to structure the basic concepts, maybe citing links I've read or attaching documents, I let it work for a few minutes on the structure of the result, I fix the steps a bit, and then I move on to a Deep Search.

As I was saying, the latter can take up to 30 minutes.

So I let it go and find a dedicated space to go back and reread it, away from the frenzy of the days, usually in the evening. It is my daily review of meaningful content.

It's a 'disposable' text that I use only for myself. Those who follow me know how much I don't like publishing things written by an AI model, and therefore, I'm not interested in having content 'as I would do it'. I'm interested in learning.

Note of discomfort

A side effect of this abundance is the cognitive overload that comes with it and, at times, the confusion in front of the complexity of the reasoning and analysis that is done. Sometimes I feel overwhelmed, and I really have to force myself to get to the heart of it and understand if the AI reasoning holds up. If I think of AI as an intern, at certain moments I realize that the intern… is me.

That is, I don't always have the skills and experience to guide me in judging the result. Knowing that I can't take everything that comes out of an AI at face value, not even the most advanced ones that will come out in the coming months, this forces me to do an analytical workload that, I must say, is exhausting.

The important thing, however, is to remember that it is a question of roles and never forget it, as I wrote some time ago.

Practical examples of augmented reading

To make this reasoning concrete, I'll give you a few examples.

An example of personal study

A detailed, research-based forecasting project has recently been published that outlines a plausible scenario for the evolution of AI up to 2027. It can be found at https://ai-2027.com . It is a long, structured document, full of references, but that makes deep reflections on our next years. After reading it, I realized that I should also study it. And so I applied the method described above: first interactions with o3 and then deep search.

You can see the result here : after a good twenty-five minutes of 'purely AI' work, I had a 7,500-word essay(!!!) that took me more than half an hour to read.

A brand new gem

From April 25th, it is possible, with o3 and o4, to schedule activities. This means that you can ask ChatGPT to send you content on a regular basis, even if you are not connected, with a simple notification.

For some years, I have been using tools that bring me back concepts in constant and daily pills so as not to forget the thousands of things I read and study.

Try doing a deep search and using the prompt below, and you'll see... then let me know.

Okay, now I would like you to bring back to me every morning at 8 o'clock for the next 10 days a reflection from the essay you just wrote. A pill of reinforced learning by creating a task

For those wondering if all this impacts business, here is a concrete example from a project I am working on.

A business example: Green Ride

To be practical, I have to do a little promotion and explain the context. Green Ride is a connected sustainable mobility project that we are carrying out to offer companies a fleet of e-bikes for bike-to-work or company missions.

We are currently carrying out an exciting project with Jakala Civitas for a structured business model for industrial districts. The project is called GRIND and, among other activities, involves an intense analysis of qualitative data from online research, papers, and various studies. One of the steps was a competitor analysis. It required a lot of data and analysis that was long, complex, and difficult to find.

Without boring you too much, the process description alone is a few pages long. I'll summarize it with a graph that should explain it (maybe I'll do technical content just on this later).

We then verified the analysis with several experts, and it was incredibly solid and accurate. I think 80% of the time was spent just reading, fact-checking (which almost always went well, i.e., there were almost no hallucinations), and analyzing the content itself. Doing it traditionally would have taken weeks of work by a large team. We managed to do it in a few days in a much smaller team.

Are we entering the age of reading?

I really like this new way of reading and studying, this journey through new information and thoughts, and continually changing direction, retracing one's steps, starting over.

Or rather, I would like to have an AI agent do it a million times faster than us and generate a very long and complex research document to read, which requires time, patience, and attention. if only to make sure that it does not contain hallucinations or errors or that the agent has not taken a different direction than the one we wanted.

And perhaps this type of deep reading stimulates a new form of dialogue with artificial intelligence and, probably, of thought.

Are we at the beginning of the age of reading or am I too optimistic?

An era in which reading what we have critically constructed through prompts - the instructions we give to Agents - will be a source of new thoughts, new ideas, and new boundaries?

So what?

Those who know me know how much I love having AI ask me questions. How much I love having it prompt me on every topic. To the point that sometimes, using an AI greatly extends the production times of any content because the questions I receive make me question myself and force me to think about new things.

I didn't study classical studies and I must say that at school I wasn't exactly an example for historical-humanistic subjects. But when I stopped going to school, I started studying, researching, until a few months ago, without agents, origins, causes, and effects.

And I believe we are facing a new historical phase that I like to call new digital humanism.

A phase in which we rediscover the value of humans even more through the enhancement offered by the capabilities of the AIs we have at our disposal.

I always talk about this in my workshops: as soon as you start thinking with AI, you must constantly analyze whether you are growing or losing skills by delegating cognition to AI. If that activity is improving you, intellectually, in terms of knowledge, character, open-mindedness, or if it is clouding your mind.

Obviously, I cannot ignore the broader implications, such as the inevitable problem of job losses that is being created. Instead, I would like to begin to understand, also together with you, who read me, if there are characteristics to be optimistic about:

• Greater development of critical thinking

• Trust in human reason

• Valorization of knowledge

• Greater valorization of 'original' content created by humans

If we enter an era of reading renaissance, then it will also be an era of listening, reflection, and deep dialogue between man and machine, and, by extension, between man and man.

And in that light, perhaps, we may truly witness a new renaissance.

Thoughts? Criticisms?

See you soon!

Max

Ps. if you liked this article, share it. Find out more on @maxturazzini or on https://maxturazzini.com