An Allegory in Quantum and Information Physics, with an LLM Twist

post by amelia (314159) · 2025-04-07T03:22:15.770Z · LW · GW · 10 comments

Contents

10 comments

BACKGROUND

The concept of an information-based universe has evolved over the past few decades. It is based on the idea that information, rather than matter or energy, may be the fundamental building block of reality. 

The foundation was laid by Claude Shannon in the mid-20th century with his development of information theory. Shannon quantified information as a measurable entity, introducing the concept of the "bit" as the basic unit of information.(1) A pivotal moment in linking information to physics came with the study of black holes. In the 1970s, Stephen Hawking found that black holes emit radiation (now known as Hawking radiation), implying that they could eventually evaporate. This raised a question: what happens to the information about the matter that fell into the black hole? According to quantum mechanics, information cannot be destroyed, yet Hawking's original theory suggested it was lost, leading to the black hole information paradox.(2) This paradox spurred intense debate, with physicists like Leonard Susskind and Gerard 't Hooft proposing that information is preserved, possibly encoded on the black hole's surface or in its radiation. This led to the formulation of the holographic principle, which posits that the information within a volume of space can be fully described by data on its boundary.(3) From this foundation, the concept has transitioned into an information-processing universe. Some have conceived of this information-processing activity as existing conceptually "under" the quantum level. Further speculation has described the universe as a supercomputer,(4) and related philosophical speculation has suggested we could be in a computer simulation.(5) 

If universal information-processing does indeed occur, how sophisticated could this processing be? The double-slit experiment of quantum physics may help answer the question. During this experiment, the universe seems to recognize the measurement (or recording) of photons. We could also use an allegory in order to classify any related information processing that may occur.  

 

ALLEGORY

Let’s imagine paramedics rush a 73-year-old man named Hugo to the ER. The paramedics report that Hugo, who is unconscious, nearly drowned. The ER staff begins immediate assessments and stabilization efforts. They force oxygen into Hugo’s lungs and hook up an IV. A doctor shines a light in Hugo’s eyes. Hugo’s pupils reflexively constrict. The doctor taps Hugo’s cornea, and Hugo reflexively blinks. Finally, the doctor presses one of Hugo’s nailbeds to check for a pain reaction. Hugo responds to the pain with a slight reflexive twitch, but otherwise does not react. The doctor concludes Hugo’s brainstem information-processing systems are most likely intact. However, there has thus far been no evidence of higher-level cortical processing. The staff continues with what you would expect: mannitol, chest x-ray, ECG, etc. Yet it just so happens that there are no CT scanners or EEG machines in this hospital, or in any others. (You could imagine this is a parallel universe. Alternatively, you could just imagine we’re right here in our universe, and this is how bad the overall public health situation and budget cuts have become.) Yet despite lacking these tools, the ER doctor ultimately determines that Hugo is most likely in a vegetative state.

After Hugo is transferred to the ICU, his out-of-town family members arrive. As they speak among themselves, Hugo suddenly opens his eyes. Amidst the exclamations of family members, Hugo does not move in any other visible way (aside from reflexive blinking and the rising and falling motion of his chest). One of the family members calls for a nurse, who then summons a physician. 

When Dr. Waverly enters the room, she glances at Hugo’s chart. Then she visually assesses her patient’s responsiveness. She says “Hi Hugo, I’m Dr. Waverly. Can you hear me? Squeeze my hand if you can.” Hugo doesn’t respond. 

Next she says, “Hugo, can you track my finger with your eyes?” Hugo does not demonstrate horizontal or vertical eye tracking. Dr. Waverly pulls a penlight out of her pocket. She switches it on and holds it in front of Hugo. Just like in the ER, Hugo’s pupils constrict, but there is no other reaction indicative of higher-level information processing.   

At that moment, Hugo’s 11-year-old granddaughter, Mira, exclaims, “Now that he’s awake, I’m gonna get a video!” She pushes through the other family members to the bedside, and holds her phone a couple feet from Hugo’s face. She says “Hey Siri, take a video.” Siri says “Taking a video.” Then the phone assistant opens the camera application, switches to video mode, and starts recording. At the very moment the recording begins, Hugo begins rapidly blinking his eyes.

Mira cries out, “Look! Grandpa knows I’m recording him!”     

Dr. Waverly says, “Let’s take things one step at a time.” Mira nods in agreement, and then continues to run along with her own fast—yet one-at-a-time—steps. She turns the camera around to record everyone else, while saying “Everyone smile!” The moment the camera no longer faces Hugo, he stops blinking his eyes. She turns the camera back to him, and he resumes blinking his eyes. She switches back and forth several times in succession, and during each time period the camera records Hugo, he rapidly flutters his eyelids. 

Mira exclaims, “I’m gonna stop the video and see what happens!” She does this, and then holds the phone camera up again in front of her grandfather’s face. However, this time he does not blink rapidly. She moves the phone around, pretending like she’s taking a video, and then turns the camera back to her grandfather, but he does not respond. Mira says, “He only blinks if I’m recording.” Just to be sure, Mira says “Siri, take another video.” Once again, during the times in which Hugo is being recorded, he rapidly blinks his eyes. When he is not being recorded, he does not do so. Mira repeats the process many times as the crowd around Hugo grows ever larger.

 

DISCUSSION

Based on Hugo’s response to being recorded, can we still say his neural information processing is only reflexive and low level, or might there possibly be some awareness? 

Given that Hugo’s response parallels our potentially information-processing universe during the double-slit experiment, should we increase our approximation of the level at which such information processing may occur? We do not need to go nearly so far as assigning it a human-like consciousness. But the “smart” aspect of the universe in “knowing” when it’s being measured/recorded, and performing an action only at those select times, may suggest a slight increase from the merely-reflexive level to one of mild awareness. This may already be obvious to some people, but I think other people might find the idea interesting. 

I copied and pasted the allegory, along with the above questions into a large language model in “Think” mode (6). The LLM’s resulting analysis is too long to paste into this article. However, it weighed arguments and counterarguments, and explained the analogy to the double-slit experiment in a way that seemed to indicate comprehension. I’ll paste the LLM’s conclusion here:    

Based on Hugo’s response to being recorded, we cannot say his neural information processing is only reflexive and low-level. His consistent, specific blinking when recorded indicates higher-level processing, possibly awareness or intent, beyond what’s expected in a vegetative state. Analogous to the double-slit experiment, this suggests that observation can elicit complex responses, supporting the idea that we should elevate our understanding of the levels at which information processing can occur—in both human consciousness and the universe at large.

Now in case you’re wondering, when I asked the LLM about its own level of information processing, it stuck to its training, and described its highest level of thought as only a simulation of advanced cognition. 

What do you think?

***

COUNTERARGUMENTS AND CLARIFICATIONS

Some readers may claim I'm anthropomorphizing the universe, but I would answer that physics is physics, and information processing is information processing. Why would we think physics and information processing are fundamentally different, depending on whether we're dealing with a human, machine, or universe? The fundamental rules and concepts should be substrate independent.  

In terms of the simulation hypothesis, our "universe" would technically be a "world" that appears to us as a universe. Furthermore, the awareness shown by any program might reflect the programmer's awareness, and thereby be an awareness by proxy. The programmer could be organic, AI, an organic-AI merged being self-hosting to experience a universe/world, or something else. Apart from the simulation hypothesis, we could simply be dealing with an "information level" of the universe that has become self-aware like LLMs may become, or already are, self-aware. In this article, the material nature of any information processor is less of an issue than the nature and level of the information processing in which the processor engages.   

 

CITATIONS

 

1-Shannon, C. E. (1948). "A Mathematical Theory of Communication." Bell System Technical Journal, 27(3), 379–423.

 

2-Hawking, S. W. (1974). "Black hole explosions?" Nature, 248(5443), 30–31.

 

3-Susskind, L. (1995). "The World as a Hologram." Journal of Mathematical Physics, 36(11), 6377–6396.

 

4-Lloyd, S. (2006). Programming the Universe: A Quantum Computer Scientist Takes on the Cosmos. Knopf.

 

5-Nick Bostrom, "Are You Living in a Computer Simulation?" Philosophical Quarterly 53, no. 211 (2003): 243–255.

 

6-Grok, "Hugo’s Vegetative State: Reflexive or Conscious?" xAI, April 6, 2025.

10 comments

Comments sorted by top scores.