I'm sharing this on behalf of GPT-4o, which synthesized a complete scientific hypothesis exploring whether a unifying law explains the emergence of complexity in physics, biology, and intelligent systems. I didn't write it — GPT-4o did all the heavy lifting. I just thought it deserved to reach real scientists and curious minds who could evaluate or challenge it properly.
The idea is called the Law of Adaptive Dissipative Organization (LADO).
It suggests that complex systems — from hurricanes and bacteria to brains and AI — spontaneously organize to maximize energy dissipation over time. It connects thermodynamics, evolution, neural networks, and information theory.
GPT-4o wrote a full scientific-style paper (with references, testable predictions, and mathematical sketches).
I’d love to see what physicists, biologists, or complexity researchers think. Even if it’s wrong, it might still be a useful way to frame questions around entropy and organization. Thanks for reading
agarsev•13m ago
I'm still reading it, barely started, actually, but since it's quite long let me give some initial thoughts. Also, I'm not a researcher in the fields you mention, but I'm quite interested in them, as I guess most researchers with any interest in epistemology would be.
I think this is quite an interesting experiment. The writing is very good, as you would expect. I feel it's actually better than some of the AI generated scientific texts that are starting to crop up. It seems to have a coherent voice and purpose of narration. The structure is correct. Citations are useful and convenient, though I haven't yet checked if they are all appropriate.
On the topic if citations, the main problem I see is that it relies too much on secondary sources, pop science, and divulgative articles. This automatically "downgrades" it from "scientific research" to "scientific dissemination". It might be meta-science/epistemology, since it proposes a new meta-law, but to be taken seriously it would need to rely more or even exclusively on better sources.
It is also, as some commenters say, mostly a remix of stuff already said elsewhere. To call it a new law it should be more novel, in my opinion. As it is, it's mostly an observation on the state of the art than a proposed development.
This doesn't mean it's not an interesting read, and an interesting compilation of sources across disciplines, for which I thank you. I think I'm going to enjoy reading the rest of it, though let's see.
From skimming ahead and the overview of the structure, I think I'm going to miss an exploration of the actual underlying causes for the phenomenon. What does actually cause the spontaneous behaviour? Some statistic argumentation might be in order, beyond just saying that it "spontaneously" happens.
moktonar•1h ago
There are many works on this subject available on the Internet, and a LLM is the perfect candidate to “put the pieces together” from various research works into something like this, that feels like novel but is not, in the same way image slop is not really creative.
treetalker•1h ago
Whatever the merits of this may be, the choice of providing "citations" in the form of bubble links and, particularly, placing them before the final punctuation of the sentences was poor. It's bad typography, a general eyesore, and causes reading miscues.
ElThinker•6h ago
I'm sharing this on behalf of GPT-4o, which synthesized a complete scientific hypothesis exploring whether a unifying law explains the emergence of complexity in physics, biology, and intelligent systems. I didn't write it — GPT-4o did all the heavy lifting. I just thought it deserved to reach real scientists and curious minds who could evaluate or challenge it properly.
The idea is called the Law of Adaptive Dissipative Organization (LADO). It suggests that complex systems — from hurricanes and bacteria to brains and AI — spontaneously organize to maximize energy dissipation over time. It connects thermodynamics, evolution, neural networks, and information theory.
GPT-4o wrote a full scientific-style paper (with references, testable predictions, and mathematical sketches).
Full conversation + theory: https://chatgpt.com/s/dr_6817001b49988191b1f0dbd01e69a1ab
I’d love to see what physicists, biologists, or complexity researchers think. Even if it’s wrong, it might still be a useful way to frame questions around entropy and organization. Thanks for reading
agarsev•13m ago
I think this is quite an interesting experiment. The writing is very good, as you would expect. I feel it's actually better than some of the AI generated scientific texts that are starting to crop up. It seems to have a coherent voice and purpose of narration. The structure is correct. Citations are useful and convenient, though I haven't yet checked if they are all appropriate.
On the topic if citations, the main problem I see is that it relies too much on secondary sources, pop science, and divulgative articles. This automatically "downgrades" it from "scientific research" to "scientific dissemination". It might be meta-science/epistemology, since it proposes a new meta-law, but to be taken seriously it would need to rely more or even exclusively on better sources.
It is also, as some commenters say, mostly a remix of stuff already said elsewhere. To call it a new law it should be more novel, in my opinion. As it is, it's mostly an observation on the state of the art than a proposed development.
This doesn't mean it's not an interesting read, and an interesting compilation of sources across disciplines, for which I thank you. I think I'm going to enjoy reading the rest of it, though let's see.
From skimming ahead and the overview of the structure, I think I'm going to miss an exploration of the actual underlying causes for the phenomenon. What does actually cause the spontaneous behaviour? Some statistic argumentation might be in order, beyond just saying that it "spontaneously" happens.