Can Physics Predict Evolution? - Assembly Theory Explained

“Infinite balls, which is basically what it takes to propose such a theory” peak math humor

Now it's time to test this formula by calculating its values for different well-studied objects and see how well it allows to classify them.

“Statistically Impossible” changes as conditions change. What was “statistically impossible” before a certain molecule assembled became highly probable once that molecule was present. What seems to be missing from the theory is any concept of initial conditions. I suspect there must be classes of phenomena that give false signals of “life” by having just the right amount of orderliness.

The complexity for a “phospholipid ball” to acquire a GLUT transporter and start the tremendous steps to make ATP, to me is mind boggling

Is this really a theory? For it to be a theory it would need to make testable predictions that are not already made by other theories. I think this is at best a model, or perhaps a way to construct models.

Didn't Kenyon write the book on this and then found out it was impossible? Information theory is different from complexity. It's a whole other level.

Thank you for a nice summary of this latest physics theory trend. From your video and the original papers, the main point of assembly theory appears to be that successful evolution is necessarily multi-level, with each level having its own ability to select, remember, and replicate valuable entities. Software designers call this modularity.

Regarding the SOS or Friends radio signal, if my understanding of Assembly theory as you describe is correct, then it depends on how likely it is for their complexity to exist based on random processes. The example of the SOS has lower assembly complexity than the Friends radio signal though so it would be easier to determine that the latter is more likely from life than the former unless there were enough copies of SOS made out of rocks that it couldn't be produced randomly.

The more we learn the further we are from solving this question. Lee Cronin has pitched his life's work to saying "I will create life in 18 months. 30 years of 18 months away. I took 3 Synthetic Chemistry courses for fun to better understand this question.

I think many biologists have had this idea of "start with simple pieces, combine them into more complex things, check what's able to reproduce more and build upon that" for quite a while now, and it's not hard to see how that process can slowly turn seemingly impossible events into very likely ones. However, it's really nice to see that idea formalized into a mathematical model, kudos to Professors Sara I. Walker and Leroy Cronin! Thanks to Dr. Ben Miles for the nice presentation of the subject, too. I'd just like to point out that the title of the video can be a bit misleading with that "Predict Evolution" part. Assembly theory does not let us predict if, say, zebras will grow horns one day or things like that. It's still nice, nonetheless.

This is the most exciting theory I have heard in years. Of course, I like to measure things, that said however, being able to recognise something is of great benefit.

very interesting topic! It reminds me of Kauffman's Theory of Adjacent Possible (also on combinatorial). I've also considered similar ideas, but from a physics and information theory perspective, in the paper "The Universe as a Telecommunication Network" (J. Phys.: Conf. Ser. 2533 012045 - DOI 10.1088/1742-6596/2533/1/012045)

I think the Banarch Tarski theorem is quite different from Euclid's parallel axiom. In fact, the situation is practically the opposite: With the parallel axiom, it turned out that there are other, slightly different valid axioms that also fit with all the other things you want Euclid-style geometry to do Anything that does not rely on that axiom is simultaneously a proof for all forms of geometry, and the parallel axiom tells us about the situations where we must special-case. So dropping or relaxing the axiom gave use new universes of mathematics to look at, which is great. However, Banarch Tarski has basically the opposite situation: It is a direct consequence of the Axiom of Choice and it's one of the reasons why that axiom is considered suspect by some, with many relying instead on weakened variations that can no longer be used for deriving Banach Tarski. It's basically one of several strange artefacts by just assuming unbounded choice. So it's a result of a very particular, specific mathematical world that probably doesn't have much bearing on reality. In fact, if you want to guarantee stuff to be actually realizable, you're going to have to drop one more axiom beyond choice: The Law of the Excluded Middle is also causing some strange things (though fewer than the Axiom of Choice) and makes it impossible to say how to get something by simply following the proof that you get something. By dropping these two axioms and looking at various alternatives, you discover many different mathematical worlds, some of which have weird stuff like Banach Tarski in them, others weird in other ways. For instance, in one such world you have things of which you can not prove them different from zero nor can you prove them to be zero. They are "confused with" zero, and it's strange to have them at first, but these objects are very useful. In particular, they allow you to quite trivially redefine how differentiation works entirely algebraically, no limits needed: These strange new objects are one particular flavor of infinitesimals! And there are many many such worlds, hidden away by overly strict axioms that can often be avoided. And that's what the discovery of spherical and hyperbolic space broke through.

I like this theory. Any attempt to quantify a phenomenon is a step in the right direction. It is long overdue for biology.

These mathematicians have clearly have no idea just how complex a cell is…imagine an entire city that builds copies of itself over and over, then multiply that complexity times a million or so, except imagine that all of the parts are made outside of the city walls and brought in by machine’s specifically designed to go search out exactly the right pieces and bring them back. That should give you an idea of a single cell.

8:49 - interesting. I was just listening to a podcast about “bioelectricity & the blueprints of life” - trying to explain (among other things) cell differentiation when the cells start with identical genetic code. Feels pretty layered.

11:10 "Of course mule is alive metabolically." -- But isn't that the important part? Technically it is a multicellular eukaryotic organism and the individual cells of which it consists are reproducing all the time.

My posdoctoral mentor, who earned his PhD in biophysics with a Nobel laureate, once said something to the efeect of, physics and biophysics have uniquely helped illuminate some crucially important questions, such as the mechanism of neruronal excitation and the structures of macromolecules. Unfortunately, it also caused too many smart people to waste their lives. I think this might be an example.

Dr James Tour would argue that we don’t have sufficient current understanding of the chemical process to fully explain how life started on this planet let alone on other worlds.

Lee Cronin: "I've got no clue...".   Sir, you are correct about that.