Adam Devereaux retweeted
I've long argued that, outside of formally posable questions, thought experiments tend to do more harm than good.
And this is why.
Note that this experiment is not formally posable, because what must be "written on the paper" is not computationally reducible.
There *exists no prestatable set* of possible input / output combinations, because language is neither static nor closed at any level.
Many forget this, because linguistic evolution has always occurred over spans not traceable during a single lifetime.
And this is where those whose minds are consumed by Analytic Reduction always come up short––they must invariably reduce open processes that cohere under the constraints of emergent continuity to models that close over, and thus render invisible, the actual issue at hand.
Which is that the paper, which can't exist, also can't *do* anything. It is not the *kind* of process that *can* do anything. It is not even itself a process; it is a terminus of lower order processes, rendered inert by our own hand, in service our own open-ended telos.
This distinction is not trivial. For example, it was the singular point that John Wheeler, one of the most influential physicists of the 20th century, chose to convey to his classes as a frame for their studies.
He did so by asking the class to write down every flight-relevant equation they knew on a sheet of paper, then to turn their papers in.
At which point he threw the papers in the air and beckoned them to "Fly!".
Of course they fell to the ground, inert as ever.
And this is the central point––a point humanity will not and cannot address until we remove the Analytic Worm from the memetic depths to which it has burrowed and re-examine its limitations, in light of the accumulated evidence against its primacy as explanatory frame.
Math, of course, cannot be conscious. Let’s assume a human could perfectly write/simulate every output of a 100T parameter LLM on a piece of paper. That system, in itself, is conscious. That system would have to be aware that a human is writing the outputs for it when you ask it. The only entities capable of simulating a self-aware and conscious being are conscious beings. If you place the human and paper in a box and only observe the output of the paper, the system is conscious. However, when you replace the human in the scenario with a computer, people become angry.
Adam Devereaux retweeted
🎶"One day we'll find it,
The Russian connection,
The hustlers, the cheaters, and YOU!"
youtu.be/V2OuJYaz_oE








