A cell type corresponds to a state cycle attractor in the dynamical behavior of the genomic system… An immediate consequence of this identification is that cell types are discretely different. Attractors lie in different basins of attraction and do not intergrade.#
We know that some computations
cannot be described in a more compact form than carrying out the computation and
observing its unfolding. Thus we could not, in principle, have general laws, shorter more compact descriptions, of such behavior. Thus we could not have general laws
about the behavior of arbitrary, far-from-equilibrium systems.#
Minor modifications of minimal programs grossly alter the output of the algorithm. In contrast, minor alterations of highly redundant algorithms modify output only slightly. In short, compression of an algorithm increases the ruggedness of the landscape it adapts upon.#
Organisms may be buffered against a
range of alterations in their fitness landscapes by harboring a hierarchy of epistatic couplings among genes or traits.#
For a fixed mutation rate, as the complexity of the entities under selection t increases, the number of “bad” alleles which accumulate increases as t2.#
For a system with a sufficient number of parts. the fitness loss due to mutational damage of one part becomes small. Therefore. the selective force tending to restore the damage becomes weaker than the mutational pressure tending to damage the part. In short, selection becomes too weak a force to hold an adapting population at adaptive peaks… Beyond that level of complexity, selection cannot climb to peaks or remain there.#
As the complexity of the system under selection increases, selection is progressively less able to alter the properties of the system.#
The simple example of computer programs makes it clear that not all complex systems are graced with the property that a minor change in system structure typically leads to a minor change in system behavior.#
Selection may attempt to pull the evolving population toward properties which are rare in the ensemble, but as it does so, the “back pressure” of mutations toward the statistically typical properties of the ensemble will increase. Thus if selection is a sufficiently weak force with respect to the mutational processes, the evolutionary process will come to rest at an equilibrium modestly displaced from the average properties of the underlying ensemble… If selection can only slightly displace evolutionary systems from the generic properties of the underlying ensembles, those properties will be widespread in organisms not because of selection, but despite it.#
[There are] critical limits to the power of selection: As the entities under selection become progressively more complex, selection becomes less able to avoid the typical features of those systems#
Institutions in economics are commonly modeled as repeated games, and strategies in repeated games are modeled as algorithms. Algorithms are explicit sequences of instructions that map from an input to an output. But in a world of open-ended affordances, Goodhart’s law implies no finite-length algorithm can maintain cooperation in large . . .