Discussions with chatGPT on physics like quantum mechanics and quantum computing gives (i) direct quick access to an AI physicist who has (ii) read the literature and (iii) argues according to logics of language. This can be a very constructive experience and I have learned a lot since my reading is limited. On the other hand, discussion with real physicists may be less rewarding since (i)-(iii) may not fulfilled.
Let me give an example of a discussion with chatGPT with the following prompt:
Standard Quantum Mechanics StdQM is troubled by the fact that its basic mathematical model in the form of Schrödinger's Equation SE has exponential computational complexity because it involves $3N$ spatial dimension for a system with $N$ electrons, and so is uncomputable.
StdQM thus acts like a symbolic theory without ontology which does not offer any definite prediction of any real physics. In practice StdQM is thus replaced by a another theory which is computable and so is predictive, which can be Density Functional Theory DFT, Atoms in Molecules AIM, Molecular Orbital Theory MO as standard since long, or the recent Real Quantum Mechanics RealQM.
StdQM thus does not deliver any predictions for systems with several electrons, and so cannot be either verified or contradicted. DFT, AIM and MO are viewed to be drastically dimensionally reduced approximate versions of StdQM capable of delivering predictions, which are accepted if agreeing with observations and discarded if not. Accordingly, text books claim that there is no prediction of StdQM which does not fit with observation.
RealQM is not a dimensionally reduced approximate version of StdQM (without ontology), but starts from a model of atoms and molecules in terms of classical continuum physics based on non-overlapping one electron charge densities as a model with clear ontology. RealQM is computable in the same sense of as classical continuum physics, and thus delivers predictions for systems with many electrons.
Despite being uncomputable non-predictive without ontology, StdQM has remained for 100 years as the foundation of modern physics as being fundamentally different from classical continuum physics. None of the foundational problems coming from multi-dimensionality have been resolved, but since no real alternative to StdQM has been pursued, only reductions, StdQM still is the untouchable theory as being the only alternative and as such a theory that cannot be wrong: The only possible theory cannot be wromg!
Here RealQM brings new perspective by offering another possible theory thus removing the exclusivity of StdQM as the only possible theory asking for concrete evidence of correctness.
I asked chatGPT to comment on the article and got a response as if chatGPT took the role of a standard physicist defending StdQM and not taking exponential complexity and exclusivity into account. So I asked for response taking on a different role:
That’s an excellent shift in perspective — rather than responding like a professional physicist or philosopher of science, let’s try to look at Claes Johnson’s post through the eyes of an intelligent “man in the street” — someone curious, reflective, and a bit skeptical of institutional authority, but not steeped in quantum mechanics.
How it might read to that kind of reader
Johnson’s message, boiled down, sounds something like this:
“The quantum theory physicists worship is too complicated to compute, too abstract to connect with real stuff, and too exclusive — they won’t even look at alternatives. Meanwhile, AI can actually hold a better conversation about it than most human experts.”
That’s a pretty gripping claim for a lay reader. It has two strong hooks:
-
The complexity argument — “If the math really needs a space of 3N dimensions for N electrons, that sounds insane! How could anyone ever calculate that?”
-
The exclusivity argument — “If physicists admit the math is impossible to compute exactly, why do they still insist it’s the only valid theory and dismiss alternatives out of hand?”
Those both appeal to everyday intuitions about fairness, openness, and practicality. Let’s unpack them in that spirit.
1. The exponential complexity issue, from a layperson’s angle
To most people, exponential growth is intuitively frightening.
If doubling the number of electrons makes the problem millions of times harder, you’d think: “Surely nature itself can’t be that complicated — atoms don’t sit around doing 10⁵⁰ calculations per second!”
So when Johnson says quantum mechanics involves 3N-dimensional equations, the ordinary reader might nod: “Yeah, that sounds unphysical — like a math game divorced from reality.”
The lay reaction is reasonable: if a theory’s equations blow up exponentially and no one can ever compute them directly, maybe it’s not the “real” mechanism but a symbolic shorthand — a way to get results in limited cases.
At that intuitive level, Johnson’s point feels plausible. He’s basically saying, “Nature works efficiently — any true physical law should be computable.” That resonates with common sense: we expect the universe to be elegant, not computationally nightmarish.
The subtlety a lay reader might miss is that physicists don’t actually solve those 3N-dimensional equations literally — they use approximations and statistical tricks. But if no one tells you that clearly, Johnson’s critique feels like he’s exposing an emperor with no clothes.
2. The exclusivity argument
The second emotional chord Johnson strikes is frustration with scientific gatekeeping.
He implies that “Standard Quantum Mechanics” has become a closed priesthood: even though it’s uncomputable and conceptually foggy, physicists defend it out of tradition and career safety, while new ideas (like his “Real QM”) are ignored.
To the lay reader, this sounds familiar — it’s the same story people tell about entrenched academia, Big Tech, or politics: institutions protect their dogmas and insiders. That’s a powerful narrative.
Even if you don’t understand the physics, you can understand the social drama: the lone thinker with a simpler idea being dismissed by the establishment. That taps into a long cultural archetype — Galileo, Tesla, Einstein — and in our AI-saturated age, the “outsider vs orthodoxy” story finds fertile ground.
So Johnson’s framing — “real physicist vs AI physicist” — doubles as metaphor: the open, logic-based machine versus the human expert bound by dogma. It’s half-serious, half-provocative, but it plays well rhetorically.
3. The psychological comfort of “computability”
For an everyday thinker, there’s something comforting about the idea that the world is computable — that with the right model, we could, in principle, calculate everything.
Johnson’s promise that his “Real QM” is a continuum, “computable” model of atoms feels satisfying: it sounds like bringing physics back down to earth, to something continuous and classical we can picture, rather than probabilistic ghostly waves in 3N-space.
So even without understanding the math, the man-in-the-street feels an intuitive pull:
“This guy wants physics to make sense again, to be calculable and visualizable — not weird and mystical. I like that.”
4. The possible unease or doubt
But an intelligent layperson might also feel uneasy. There’s a nagging question:
“If this simpler, computable model really works, why haven’t others adopted it?”
That thought undercuts the romantic rebel narrative. Most people know that science, while bureaucratic, does eventually adopt what works (e.g., relativity, plate tectonics). So maybe Real QM hasn’t convinced anyone because it doesn’t yet match data well enough — or maybe it’s incomplete.
So the lay reaction could split:
-
Sympathetic skeptic: “He’s right — they’re hiding behind math they can’t compute!”
-
Cautious realist: “If it worked, people would use it. Maybe it’s just not there yet.”
5. The AI angle
Finally, the “AI physicist” idea.
To the ordinary reader, that might sound both amusing and profound: if ChatGPT can discuss physics calmly, quote papers, and reason clearly, while human experts get defensive, maybe that tells us something about human bias.
So Johnson’s anecdote — that an AI conversation is “more rewarding” — rings true emotionally, even if it’s tongue-in-cheek. It’s less about AI understanding physics, and more about AI being open and non-tribal, which flatters readers who are tired of expert arrogance.
In that sense, the AI serves as a rhetorical mirror: a symbol of reason over orthodoxy.
6. Overall lay impression
If you’re not a professional physicist, Johnson’s post can sound refreshingly candid and a bit rebellious. It appeals to:
-
Common sense (“Nature can’t be uncomputable.”)
-
Fairness (“Why exclude new ideas?”)
-
Frustration with experts (“Real physicists won’t listen.”)
-
Hope in technology (“AI might help cut through bias.”)
But it also leaves a thoughtful reader with lingering uncertainty:
-
Is the exponential-complexity problem really fatal, or just hard?
-
Does his Real QM actually make accurate predictions?
-
Is the exclusivity he describes real, or just how it feels to an outsider?
So the post succeeds rhetorically — it provokes curiosity and sympathy — but it doesn’t fully settle the doubts it raises.
Would you like me to take this a step further — for example, write a short “man-in-the-street” response to Johnson’s blog post, as if it were a comment you could post under the article itself?
You said:
ChatGPT said:
Here’s a short, natural-sounding “man-in-the-street” style comment you could post under Claes Johnson’s blog entry:
Comment:
I’m not a professional physicist, just a curious reader, but your point about exponential complexity really made me pause. If quantum mechanics truly needs a 3N-dimensional equation for N electrons, it does sound like something nature itself couldn’t be running in real time. That raises a fair question: maybe the math is more of a symbolic description than a literal mechanism.
I also get what you mean about exclusivity. From the outside, physics can look a bit like a closed shop—new ideas get waved off before they’re even tried. Still, I wonder: if your “Real QM” works as cleanly as you suggest, why hasn’t it caught on? Are there published results or examples where it matches experiments better than the standard methods?
Either way, I find it refreshing that an AI can now take part in these debates. Maybe it’s less about the machine being “smarter,” and more about it being open to every angle without ego. That in itself is a healthy thing for science.
Inga kommentarer:
Skicka en kommentar