We anthropomorphize everything.
Cars get names.
Roombas become pets.
Soldiers mourn bomb disposal robots.
Children punish tables after bumping against them.
We do this automatically. Against our will, sometimes.
Which is why the Turing Test was probably doomed from the beginning.
We treated it, in a way, as the summit of machine intelligence: “Can a machine fool us? Will it ever? Can we talk to a machine and not know it’s a machine?”
But that was never the test. As it has been revealed, it was mostly a test of human projection. And LLMs exposed this in an uncanny fashion. Once they became mundane, the game was up.
Not because they became conscious (more on that later) but because they became socially convincing enough.
Arguing otherwise is like trying to argue that a computer will never outmatch a human at chess while the Deep Blue vs Kasparov matches are happening. That argument is already over. What you are arguing for is now an engineering problem, not a mathematics or scientific problem.
The remaining objections are mostly engineering constraints: speed, memory, latence, cadency, fidelity, ...
That’s it. That’s done.
We humans infer agency extremely aggressively. We are meaning-generating machines.
We look for intention in weather, order in tea leaves, morality in chaos and consciousness in fluent language use.
And we mistake limitations in our conceptual frameworks as limitations in reality itself.
Reality is doing fine. Nature is under no obligation to conform to the limitations of our models.
We’re the ones that need to adapt.
Fortunately we do just that, and we do it *fast*.
Paradox: the arrow cannot go from start to finish because it always needs to cross half the distance, and then another half of the missing half and so on and so forth?
Nope. Please get better math. Create the theory of limits, and you’ll understand why the arrow is always reaching its target.
“The first principle is that you must not fool yourself – and you are the easiest person to fool”
Richard Feynman
I always found Searle’s Chinese Room intellectually lazy.
Not because the intuition is wrong, but because the mechanism felt hand-waved into existence.
Peter Watts took that intuition seriously, though.
He weaponized Searle’s intuition and made the Chinese room into something more than hypothetical. It felt optimal and self-sustaining.
Watts posits in Blindsight not only that Consciousness isn’t necessary for intelligence, but that it can actually be detrimental, an evolutionary slip, a mishap.
I don’t want to go too much into this, because I think anyone who loves Sci Fi should absolutely read Peter Watts’ Blindsight.
Back to LLMs, the Turing test and of human inevitability.
The unsettling possibility is that the Turing test was never a test on machines but a demonstration on human limitations waiting to happen.
We require far less complexity and depth than we imagined before attributing understanding, intention and mind to language output systems.
What matters then? Should we just wave our arms in the air like we just don’t care?
I think that logic, rigor, curiosity and scientific method matter now more than ever.
Because, once more, we realize that humans remain deeply vulnerable to narrative, projection and linguistic fluency.
LLMs did not suddenly create self-deception. They industrialized it.
Many of the frictions that forced us to slow down, investigate, compare, doubt and eventually learn, have been bypassed.
The real question is whether our capacity to adapt can keep pace with the mirrors we are building, or whether we will simply become fascinated by our own reflection.
