The Science of Consciousness: What Do We Know About the Mind?

I used to think the question was whether machines could think. Now I know it’s whether we’re willing to admit we don’t understand thinking at all.

We’ve built systems that write poems, diagnose tumors, and play chess better than any human grandmaster. We call them intelligent. We even call them conscious—when they say the right things, when they mimic our cadence, when they apologize for their mistakes. But here’s the quiet truth no one wants to say out loud: we have no idea what consciousness is. We don’t know how it emerges. We don’t know if it’s a property of complexity, a biological quirk, or something else entirely. And if we can’t define it in ourselves, how can we possibly build it into a machine?

Hubert Dreyfus knew this. In 1965, when the pioneers of artificial intelligence were dreaming of thinking machines, he stood in the back of the room and said: "You’re not building intelligence. You’re building a very fast calculator with a thesaurus."

He was mocked. Then ignored. Then, decades later, quietly vindicated.

The AI pioneers of the 1950s—McCarthy, Minsky, Newell—believed the human brain was just a machine, albeit a messy one. They thought if we could replicate its structure, we’d replicate its function. Brains were good at recall, they said. Good at abstract thought. Good at making value judgments. All we needed was enough silicon. More transistors. Faster clocks. Eventually, the machine would wake up.

It didn’t.

Dreyfus didn’t argue that machines couldn’t be useful. He argued they couldn’t be alive in the way we are. He didn’t deny computation. He denied that computation equals cognition.

We’ve spent the last sixty years proving him right.

The Fringe of the Mind

Dreyfus called it "fringe consciousness." That whisper of context that lets a child know when a parent is pretending to be angry. That gut sense that a sentence is wrong, even if every grammar rule is followed. That moment you walk into a room and feel something’s off—not because you saw it, but because the air changed.

Computers don’t have fringes. They have inputs and outputs. They process. They don’t feel.

In the 1970s, AI systems couldn’t translate languages without literal word-for-word substitution. They couldn’t write poetry without stealing metaphors from databases. Chess engines relied on brute-force lookahead, not intuition. They didn’t "see" a position; they counted moves.

And today? GPT-4 can write a sonnet in Shakespeare’s voice. It can simulate empathy in customer service chats. But it doesn’t know what grief feels like. It doesn’t know why we cry at weddings. It doesn’t know the difference between a joke and a wound.

It’s a mirror. Not a mind.

The Human Edge: Ambiguity as a Feature, Not a Bug

We don’t fix ambiguity. We live inside it.

A human hears "I’m fine" and knows it means the opposite. We read tone, posture, silence. We infer meaning from what’s missing. We hold contradictory truths at once: I love you, and I can’t forgive you. I’m terrified, and I’ll do it anyway.

A computer sees symbols. It matches patterns. It optimizes for probability. It has no tolerance for contradiction. No patience for the unspoken. No capacity for moral weight.

Dreyfus was right: humans can make sense of the incoherent. Machines cannot.

Not now. Not ever.

Not unless we build a body. Not unless we build a history. Not unless we build a soul.

And we don’t know how to do that.

What Computers Can’t Do

Dreyfus’s 1972 book, What Computers Can’t Do, wasn’t a prediction. It was a diagnosis.

He wrote: "Human thought cannot be translated into a language defined by logic and mathematics."

He was ridiculed for saying this in the age of Moore’s Law. But here’s the irony: we’ve built machines that can outperform us in every measurable task—and yet, they still can’t answer the simplest question: "Why?"

Why do you feel lonely?

Why does this song make you cry?

Why did you choose this life?

A machine can generate plausible answers. It can string together words that sound wise. But it doesn’t know why it says them. It doesn’t know what "why" even means.