Looking around, within the scientific community, free software is today stronger than ever, and going even stronger. All the basic tools used by scientists today are 100% free software: R, gfortran, octave, python, torch, numpy, lapack, fftw3, gmp, overleaf, jupyter, git... The only remaining non-free software pieces of infrastructure are github and some proprietary chat rooms that will go unnamed.
GPL : R, gfortran, octave, fftw3, git
LGPL : gmp (dual-licensed with gpl2)
AGPL : overleaf
BSD-like: python, numpy, torch, jupyter, lapack
What is your point?
> Big corporate ran a campaign to dirty the name of GPL, Richard Stallman and the FSF. They won.
"Hey, can we use this GPL software?"
"Better not. We don't have the infrastructure to publish all our code publicly, and legal definitely doesn't want to be in the business of deciding what code we are required to publish by the license and what code we can keep proprietary. Adding GPL code is adding a time bomb to our project."
"Good point. There's some MIT-licensed code that does basically the same thing; I'll use that."
The GPL itself was the best argument not to use the GPL. It placed impositions on the code's consumer that other licenses don't, and network effects will tend to optimize to the path of least resistance (which in this case is the saddle-point between "We want to use other people's code" and "We don't want it to be a hard legal question what code of ours must be published").