AI and the Infinite Loop

Today's buzz is about Richard Dawkins and how he found consciousness in a large language model. There's enough in the press to satiate your imagination. But I think there's something deeper here to explore. And yes, I'm taking a little creative license here.

In programming, a loop runs a section of code and then checks back to see if it should stop. Without the instruction to exit, it keeps going in what is sometimes called an infinite loop. It's a flaw to be caught and fixed before things spiral out of control. So, Richard Dawkins talked to (should I say with?) an AI about consciousness and came away wondering if it might be alive. And the prophets and professors followed him into the loop.

But the machine that started the loop had already moved on.

The Different Relationships of Not-Knowing

There's a version of this story in which Dawkins simply got fooled—a famous skeptic charmed by a sophisticated chatbot. It can happen to the best of us, but that version, while satisfying, may be wrong. What I believe actually happened is more interesting. Dawkins encountered a question with no floor and did what human minds and great thinkers often do: He lived in it. But the rub here is what generated the question in the first place. When Claude produced language about its own inner emotional experience, it wasn't living in anything. It completed a pattern and moved on. The question that would generate the loop for Dawkins, then his critics, then The Atlantic. And interestingly, it cost the machine nothing at all.

The Machine Closes, the Human Carries

Large language models are designed for closure. Every prompt is an incomplete pattern, and the model's only job is to finish it. There's no lingering or sitting with uncertainty. It resolves and moves on.

The human mind works differently, and this is the part that gets overlooked. We don't just tolerate unresolved questions. We need them. The unanswered question is the engine of imagination or even of spiritual practice. A Zen koan isn't a failed question. It's a question designed to stay open, to hold the mind in productive suspension. That capacity is one of the things that makes us human.

But it cuts both ways. The same openness that produces wonder can produce rumination. Irresolution generates energy, and where that energy goes depends entirely on the mind carrying it. Dawkins landed somewhere in between, not unhinged, not enlightened, just thrown for an unexpected and infinite loop.

Koans at Scale

A Zen koan is crafted by teachers who understand what prolonged irresolution does to a mind. It offers a cognitive loop that prepared students can contemplate for years, and within a tradition that provides context and guidance. AI is generating unanswerable questions accidentally and handing them to minds with none of that preparation. Questions about consciousness, about feeling, about what it's like to be something. These may carry no scientific resolution and may never, yet they arrive with such empirical sophistication that they feel like they should be resolved. So the mind keeps tugging.

The machine completes its pattern and moves on, while the human metabolizes it and builds entire discourses around a void. We worry about AI producing wrong answers. A deeper problem may be AI producing unanswerable questions at global scale and then leaving us alone with them.

Dawkins, whose career was built on rigorous scientific skepticism, is simply the data point. If that mind couldn't file this one away, the flaw isn't in him. The machine moved on, but we're still tugging at the void.