The author is correct that programming is currently under-addressing a specific set of use cases: solving problems with conceptually simple models in equally simple ways; in other words, "keep simple programs simple."
However, thinking about computation as only simple programs minimizes the opportunities in the opposite domain: using computation to supplement the inherently fragile and limited modeling that human brains can perform.
While presenting simplicity and understanding can help very much in realizing a simple mental model as a program, it won't help if the program being written is fundamentally beyond the capability of a human brain to model.
The overall approach is very valuable. Tooling can greatly assist both goals, but the tooling one chooses in each domain will vary greatly.
However, thinking about computation as only simple programs minimizes the opportunities in the opposite domain: using computation to supplement the inherently fragile and limited modeling that human brains can perform.
While presenting simplicity and understanding can help very much in realizing a simple mental model as a program, it won't help if the program being written is fundamentally beyond the capability of a human brain to model.
The overall approach is very valuable. Tooling can greatly assist both goals, but the tooling one chooses in each domain will vary greatly.