Nobel prize 2024 - magic just got less wriggle room

https://www.nobelprize.org/prizes/chemis...s-release/

The Nobel Prize in Chemistry 2024 is about pro­teins, life’s ingenious chemical tools. David Baker has succeeded with the almost impossible feat of building entirely new kinds of proteins. Demis Hassabis and John Jumper have developed an AI model to solve a 50-year-old problem: predicting proteins’ complex structures. These discoveries hold enormous potential.

So the protein folding problem turned out to be predictable from purely physical properties after all. No magic realms needed.

[Laird]

I wasn't aware until now that magic realms were being proposed as necessary for protein folding. Can you share more details about this proposal, and who was making it?

I wasn't aware until now that magic realms were being proposed as necessary for protein folding. Can you share more details about this proposal, and who was making it?

Certainly:

https://psiencequest.net/forums/thread-l...7#pid39647

but on the seemingly abstruse and technical argument over the origin of specialized protein folds hinges the entire debate over ID versus Darwinism. If protein evolution really happens by the undirected meaningless purposeless semi-random process of neo-Darwinism, then overall, neo-Darwinism does explain creative macro-evolution, and that fairly well decides the entire debate over ID versus Darwinism in favor of Darwinism. That means the neo-Darwinistic mechanical process underlies the entire origin of life, evolution, and human beings, and all the toxic implications of Darwinism actually underlie our reality.

I think that even if fully understanding the arguments for ID requires an extensive background in biology, the toxic and in part known to be false implications of Darwinism would rule out its really being the truth. And therefore the protein fold argument is resolved simply by its inevitable implications.

[Laird]

@nbtruthman can clarify if I'm wrong, but he doesn't seem to be saying in that quote that protein folding itself necessitates a "magic realm". He seems instead to be making the point that given the available time, a semi-random walk over (genetic) coding instructions is unlikely to the point of impossibility to result in the folded proteins necessary for life.

I’m not claiming the problem of evolution has been resolved by naturalistic causes but I do claim that the god of the gaps has been further reduced.

[Laird]

You still haven't shown anybody claiming that particular gap, but anyhow, we all misfire from time to time.

You still haven't shown anybody claiming that particular gap, but anyhow, we all misfire from time to time.

It suggests that while the total number of possible amino acid combinations is enormous, the number of physiologically relevant or beneficial proteins (protein space) may not be as large as we are led to believe as the totalt set of foldings must be quite limited to be captured by an ai model.

[Laird]

I don't understand what you're saying. That's not a rhetorical gibe: I really don't understand. Perhaps you can rephrase, and in particular (1) make clear what the referent of "It" is in "It suggests", and (2) (a) make clear how what you're saying is relevant to your original implicit contention that somebody was claiming that protein folding requires a "magic realm", or at least (b) how what you're saying is relevant to @nbtruthman's argument.

It suggests that while the total number of possible amino acid combinations is enormous, the number of physiologically relevant or beneficial proteins (protein space) may not be as large as we are led to believe as the totalt set of foldings must be quite limited to be captured by an ai model.

I think your mistake is in assuming that the AI approach to the problem of finding functional protein foldings correlated to specific amino acid chains is an undirected unintelligent Darwinesque random search program which necessarily has to be over an inconceivably large protein space in which most configurations are functionless or harmful. This underlying nature of protein space is a fact rather than being drastically limited in the total number of functional adaptive configurations, and is very extensive in its containment of very many of them. It's just that this large number of desired specifically functional proteins are randomly spread out over a huge volume of mostly nonfunctional sequences separated on average by large distances in protein configuration space. A random search program  as implemented by undirected semi-random Darwinistic processes is mostly useless. 

So the AI system approach instead, rather than trying futilely to capture the huge total set of foldlings. is instead (using very powerful computer systems to handle the large volume of data processing required) to intelligently abandon Darwinesque undirected semi-random search and examine the multitude of known functional protein folding configurations for structural commonalities and functional correlations and using statistical means predict other related foldings with the desired properties (many medical-related). This is an "intelligent design" sort of method rejecting Darwinism in order to work. 

So, AI techniques as applied to finding desired protein folding amino acid configurations reject Darwinistic processes in favor of a very intelligent-design like model.

@nbtruthman can clarify if I'm wrong, but he doesn't seem to be saying in that quote that protein folding itself necessitates a "magic realm". He seems instead to be making the point that given the available time, a semi-random walk over (genetic) coding instructions is unlikely to the point of impossibility to result in the folded proteins necessary for life.

Yes