Zaid Al Kazemi

The Extinction of Thinking

Thinking is not a passive activity. It is a high-voltage biological event.

Most of the time, your brain runs on autopilot. Your vision, your memory, and your habits all live in isolated silos, performing tasks automatically without talking to each other. Genuine thinking only occurs when these silos are forced onto a single central stage to share their data at the exact same moment.

In 1998, Dehaene, Kerszberg, and Changeux called this architectural alignment the “Global Workspace.” Their paper proved that consciousness is not a place in the brain but a temporary integration of specialized modules. This peak state feels like a rare coincidence if you stumble into it, but it becomes an intentional practice once you understand the five circuits that must fire in sync:

The Evaluative. This is your sense of stakes. It provides the “why” and determines if a thought might lead to a breakthrough. Without this internal value signal, your brain refuses to spend the energy required to think.

The Memory. This is your past. Thinking requires pulling your deep history and old lessons onto the stage to compare them against the present. Without this, you aren’t thinking; you are just reacting.

The Perceptual. This is your present. You must selectively amplify the data right in front of you, the specific problem you are solving, while muting the notification-heavy noise of the world.

The Attentional. This is the conductor. It manages the intense focus required to keep the other processors firing in sync. It decides what stays in the spotlight and what gets kicked off the stage.

The Motor. This is the commitment. Thinking requires a bottleneck. It must be aimed at a specific output: a sentence you type, a word you speak, or a hard decision you make.

Knowing how to activate the Global Workspace that’s necessary for high-quality thinking is only useful if you understand how it fails.

The Three Failure Modes of Thinking and How Modern AI Triggers All Three

The paper outlines three specific failure modes that shut down the workspace. Current AI products trigger each one with surgical precision.

Premature Habituation. The Global Workspace fires only when a task is new, difficult, or critical. The moment a task feels easy, your brain drops into autopilot. AI tools are engineered to make every task feel easy. By removing the “new” and “hard” signals, they convince your brain that nothing requires your attention. You stop thinking because the machine emulates competence.

Modular Isolation. For thinking to happen, your past (Memory) and your present (Perceptual) must meet on the same internal stage. When an AI drafts a response by “retrieving” your notes, it performs this integration for you. It simulates your thinking by connecting your data points on its own hardware. Because the integration happened inside the machine, your own neural circuits never fire in sync. You are left with output that looks like your voice but never actually passed through your mind.

Evaluation Failure. This is the terminal failure. The beginning of the extinction event. The workspace requires a reward signal to stay active—a gut sense that the work matters. When you offload the judgment of “is this good?” to a machine, your own Evaluative circuit goes silent. You lose the ability to detect that you have stopped thinking.

The Extinction Event

A species goes extinct when it loses the environmental pressure that made it adapt. The pressure of thinking is friction. The friction of struggling to remember, to evaluate, to commit a single sentence to the page.

When a machine handles all three, the pressure disappears. Your five circuits stop firing in sync because they are never called to the stage. Over months, you lose the muscle memory of what thinking actually feels like. You mistake fluent prose for your own insight. You publish ideas you never integrated. You make decisions you never consciously weighed.

The extinction is quiet. There is no final thought, no last broadcast. The stage just stays dark while the modules continue running in their silos. You become fluent, productive, and empty.

To protect the thinker, you have to protect the friction. You have to mute the machine so your past and your values are forced to collide. The effort is not a cost. It is the only signal your brain recognizes as thinking. And we can build these AI tools to amplify our Global Workspace and super power our ability to think by introducing more effort instead of lobotomizing ourselves via tokens.

Remove the effort and you remove the thinker. Enrich the environment with dense stimulation of the five circuits and you elevate them.

Read A Neuronal Model of a Global Workspace

Home