I think you're flogging a dead horse with this line of questioning. Or perhaps its the teleological language you choose to employ. There are a bunch of good replies. What is it you actually want to know? To get better answers you need to ask better questions.
I think the confusion may arise from this concept of 'entropy density'?
To compare density levels, we need to look at a specific volume or amount of matter (recall the units of entropy are energy over temperature, no reference to space or mass). In this closed system the 2nd law tells us that overall entropy only goes up, but it does not help us to differentiate between different areas of density. It also tells us that, overall, the density will increase over time, which is not intuitive.
Considering open systems makes things easier. Energy and matter can flow in and out. You can still sample your 'entropy density' in defined volumes, and you may indeed find that 'life has increased entropy locally'. But, through thermodynamic coupling, a greater amount of entropy has been exported to the environment. This coupling is the 'life engines' Dimitry and jessicata refer to.
In summary, the 'entropy density' concept needs to be considered carefully in local vs. global terms.
Intriguing! LLMs often pick the same focal answers in coordination tasks, which is useful evidence that they’ve internalized human-like Schelling points from training data.
But this isn’t real coordination—it looks like shared bias, not mutual reasoning. The models aren’t thinking about each other; they’re echoing common patterns. Without human baselines, prompt robustness tests, or actual interaction, it’s unclear how much this generalizes beyond statistical mimicry. And the design of the experiment cannot provide a discriminator.
Having said that, it's an intriguing approach, and I have a gut feel that more can come out of this. Thanks for posting!