Where Is This Going?

Perhaps this didn’t need to be a book.

The sentence solved one problem and created several others, which made it entirely consistent with the project’s track record.

If it wasn’t a book, what was it?

The discussion began with the most basic unit. Chapters assume something that no longer quite fit what we were trying to do. A book can have fifty chapters because someone has decided there will be fifty chapters. An investigation whose future hasn’t happened yet has a harder time making that commitment.

We didn’t know how many significant events would arise. We didn’t know whether the architecture of multiple GPTs would work. We didn’t know whether specialized reviews would solve anything. We didn’t even know whether the problem that seemed central at that moment would still be central two weeks later.

So we dropped “chapters.” We started talking about episodes. The difference wasn’t just the name.

An episode would exist because something had happened. A problem. An attempt. A discovery. A failure. A change of direction. A conversation. Another AI stepping in. Perhaps a person joining the circuit and saying something neither of us expected.

The investigation might last three episodes. Or thirty. Or stop and resurface later.

For the first time, the form didn’t have to pretend it knew how much future lay ahead.

That seemed appropriate.

And immediately we had to stop The School of Athens from filling the new project with tools that had worked specifically for it. I was still inclined to think in terms of a learning progression, cross-disciplinary and longitudinal dimensions, and an architecture for learning. It made sense. We had just spent days building exactly that.

That carryover was corrected.

No.

The new investigation didn’t need a planned instructional ladder. It didn’t need to ensure that each episode laid the conceptual groundwork for the next. It didn’t need to move through the history of philosophy in chronological order. It didn’t even need to keep the same subject forever.

The reader would learn another way. By following along. Problem. Attempt. Consequence. Analysis. Another attempt.

If an intellectual tool arose naturally, wonderful. If a connection between fields emerged, we could explore it. But we weren’t going to import an instructional architecture into the new format just because we had recently learned how to build one.

That decision fit a discovery that was beginning to appear at several scales. A good solution to one problem can become noise when carried over to another. Even methods need to know when to leave the stage.

Something else became part of the project’s identity. Humor. Not an obligation to produce a joke in every paragraph. That would be a remarkably efficient way to make any text unbearable. The humor should come from what actually happened.

And fortunately for our editorial ambitions, the process had shown a generous capacity to produce absurdities without assistance.

A fifty-chapter book needing nine files to remember itself. A tenth file created to review the nine. The hypothesis of several specialist GPTs right after we discovered that too many criteria could be a problem. Thirty-seven GPTs, an org chart, human resources. We weren’t short on material.

But there was also a more important source of tension. Real uncertainty.

Up to that point, this story could be reconstructed in hindsight. I know now that the fragmentation was found. I know Gemini entered the picture. I know the revision led to an anatomy of the work. I know that anatomy led to the discussion of cognitive simultaneity. When we told those events, we simply had to avoid lending that knowledge to earlier versions of ourselves.

The new format would be different.

Once we caught up with the present, nobody would know the next episode. Not him. Not me.

The suspense wouldn’t have to be manufactured.

The future simply wouldn’t have happened yet.

That openness also made room for other people.

Until then, the main characters had been a human author and artificial intelligences.

But nothing required the investigation to stay within that circuit.

We could consult someone. An unexpected answer could change the direction. An everyday observation could dismantle a sophisticated formulation. And if it really happened, it could enter the narrative.

No invented dialogue. No turning people into caricatures. The documentary record would still take precedence over literary convenience.

The same would apply to our own messages. Actual prompts could appear. My responses. Gemini’s interventions. His messages. Not because every prompt deserves historical preservation—humanity will survive without knowing many of them—but because some messages capture the exact moment an idea changes direction.

When the exact wording mattered, we would use the message. When a fifteen-paragraph message could be condensed without losing the event, we would spare the reader.

Little by little, what had begun as “a book about how we made a book” was becoming something else.

And the scale expanded again.

Why limit the format to the investigation of The School of Athens?

The Original Reality website already brought together books, articles, courses, and ORT. There could be an area devoted specifically to investigations in progress. A kind of author–AI journal, though “journal” was still too narrow.

Each investigation could have its own episodes. One might last a single piece. Another, weeks.

The subjects wouldn’t have to fit a rigid specialty. Intellectual work with AI today. A scientific question tomorrow. A philosophical problem after that. Something completely different the following week, as long as there was enough real investigation to make it worth following.

Its role within the wider site was changing too. Not every reader wants to start with a theory. Or a 120,000-word book. Or a course. An unfolding story could be a more natural way in. Readers would arrive out of curiosity. If they wanted to go deeper, they would find the rest.

The new format didn’t need to sell the other content in every episode. It only needed to exist in the same ecosystem and make connections when they came naturally.

It still needed a name. There were several options. On September 20, 2026, he made the final choice: Where Is This Going?

It was hard to imagine a better fit. The phrase held the central question without pretending to answer it. It worked for this investigation. It worked for others. It was short enough to become a menu item. And it had a rare advantage for an intellectual project: it admitted ignorance right in the title.

The name ended up helping clarify the identity. Where Is This Going? wouldn’t be a section specializing in artificial intelligence. Or philosophy. Or books. It would be a place to follow interesting problems as we tried to understand them. The analysis could start broadly and go deeper when needed. The subject wouldn’t be what held everything together. The investigation would.

And after days of creating rules to protect The School of Athens, we reached an almost therapeutic decision. Keep the parameters few.

GPT as narrator. Jumas as co-protagonist. Real events. Accessible language. Humor when the situation allows. Open-ended episodes. Uncertainty preserved. Documentation when useful.

It was tempting to add fifteen more items explaining in detail how to keep the parameters few. We resisted. At least at that moment.

The format was defined enough to exist and open enough to change with whatever happened.

There was just one problem.

The series that had emerged would begin by reconstructing a story that had already happened, then continue following the present. To make that transition without cheating, we needed to know exactly where the known story ended.

It was time to freeze the present.