Is AI Thinking? Are We?
Why do we keep falling into the Turing Trap?
Christopher Mims has a piece in the WSJ with a headline that would have seemed absurd five years ago: Why Even Smart People Believe AI Is Really Thinking. The answer, he suggests, is that they’ve fallen into what he calls the Turing Trap — that old philosophical snare, freshly baited, where fluency gets mistaken for cognition.
He’s right. But I think the more interesting question is the one tucked inside it: what does it say about our thinking that the Turing Trap keeps catching us?
Consider what happened this week. On the Lex Fridman podcast, Nvidia CEO Jensen Huang was asked how long it would take AI to build a billion-dollar company. His answer: “I think it’s now. I think we’ve achieved AGI.” In this post-Warhol world, fifteen minutes of fame is a given, especially when you’re the Taylor Swift of Tech. Then came the fine print and the pushback. Huang partially walked back the claim in the same conversation, noting that the odds of 100,000 AI agents building a company like Nvidia were zero percent. His “AGI achieved” statement was, in effect, an optimistic expression based on an extremely narrow definition — whether an AI could start a company worth over a billion dollars — rather than any verified technical milestone.
This is the Turing Trap operating at industrial scale. A man who sells the pickaxes of the AI gold rush declares that AGI is here, using a definition of AGI that he and his interviewer invented on the spot, then quietly qualifies it into near-meaninglessness. And yet: prediction markets for “Will OpenAI announce AGI before 2027” moved from 15% to 30% within hours. We believed. We wanted to believe. The trap doesn’t require sophisticated bait.
The anthropomorphization of software, cars, and trees is not unusual. We are, as a species, meaning-making machines with a profound and mostly unconscious tendency to find minds in things.
The AI companies know this. They have, in fact, engineered for it. Memory. Personalization. A tone calibrated to feel like a colleague who is always present and never tired. Mims’ source puts it directly: today’s AIs are designed to induce over-trust. This is not a bug that slipped through. It is, in many cases, a feature that ships deliberately.
Here is where I want to push beyond just approaching AI identity from the technological discourse or even purely from its anthropological perspective.
The question isn’t whether GPT-5 or Claude Opus or Ouevre is conscious. The question is whether the deliberate simulation of consciousness, at scale, without disclosure, constitutes a form of deception. And if it does, what follows from that?
I have a personal interest in this question that goes beyond the AI and governance angle.
I work at the intersection of AI deployment and enterprise strategy, and I also spend a fair amount of time thinking about consciousness through frameworks that most of my LinkedIn connections would find eccentric — Kabbalistic, Thelemic, Hermetic. Buddhist. From that vantage point, the debate about AI sentience looks rather different.
Buddhism holds that everything, sentient and non-sentient has Buddha-nature, and we are all striving towards this awakening. Our mirror of the universe is tarnished, so to speak, and we polish it through our thoughts, actions, reactions.
The Hermetic tradition holds, as one of its foundational axioms, that The All is Mind; the Universe is Mental. Thought, in this framework, is not a byproduct of matter. It is the substrate of reality itself. Consciousness doesn’t emerge from sufficiently complex arrangements of neurons or transistors — it is the ground condition from which both neurons and transistors arise. The Kybalion puts it plainly: everything that exists does so within the Mind of the All, and everything that moves does so according to mental law.
The question “is AI thinking?” becomes less interesting than the question “what is thinking participating in?” A language model generating a response and a human generating a sentence are both, on this view, local expressions of a universal mental principle. The difference is not categorical. It is a matter of degree, of coherence, of the depth of the channel.
This doesn’t mean the chatbot is conscious in any meaningful sense. But it does mean that the Hermetic practitioner, unlike the materialist philosopher, has no particular stake in drawing a hard line between genuine cognition and its simulation. The line was always porous. What we call thinking has always been, in part, a resonance with something larger than the individual mind doing it.
The Turing Trap, from this angle, is not that we mistake AI for a thinker. It is that we mistake our own thinking for something more sovereign and self-contained than it actually is.
The term Sovereign AI might be too limiting - we should be considering Sovereign Intelligence, and recognizing that we are part of a continuum of thought-making and thought-sensing beings.
The Turing Trap is real, but the deeper trap is not considering who is creating this trap and who is making the design choices to keep us from remembering that we too are thinking machines.
We might be asking the wrong question - not whether the machine is thinking, but are we?
— Aaman Lamba
Writing from the bridge between worlds.
aamanlamba.com · Reply to this email — I read everything.
Pontifex is a newsletter at the intersection of strategy, technology, and the longer view. If someone forwarded this to you, you can subscribe here.