a remarkable coincidence showing that numerical data can be misleading!

[Jonathan Post]

Dear Robert Gerbicz,

Thank you for a smart answer to a stupid question.

So there is implictly a bunch of sequences derived from the array, for k>1

A[k,n] = nth value such that ceil(2/(k^(1/n)-1) isn't equal to
floor(2*n/log(k)).

For that array we have, for instance, its presentation by
antidiagonals, its main diagonal, and the like.

Is this a foolish attempt to generalize, which makes a "nice" sequence
into a "dumb" or indifferent sequence?

I'm just wondering, as so many of the sequences from "the usual
suspects" suffer from the use of arbitrary coefficients, but may
generalize into decent sequences as easily as passing from 1 variable
to 2 variables.