What Do You Think About This Comment?

I agree with passive speaker designs - speaker can performance can be all over the map from terrible with the wrong amplifier pairing to excellent with the right match.

However active speakers are indeed optimized in a holistic sense. Manufacturers of active speakers are giving you their holistic result based on the manufacturers criteria (price point, reliability, sound quality etc)

Well yes indeed
if you don’t know where you are going, any road will get you there....
i would say start with a deep love of live unamplified music so you have a reference, chasing your tail otherwise
research the physical  and mental ways in which we hear - for example as a designer do you believe human hearing is more/equal/less sensitive to time information vs amplitude ?
system objectives come next which might with thought provide what we call the balanced trade space: cost , size, efficiency, full range, portable, etc....
then develop some design principles like : pre-ring should be avoided, wavelaunch that looks like the test signal, pistonic behavior in the driver, breakup x cycles away from crossover, cabinet nodes x dB down

probably a decent book in this, might look like designing, building, testing , and improving over time an airplane....
.......

Firstly, a lot of products are designed rather than engineered.

For example some loudspeakers have computer designed crossovers that take into account the dynamic and LRC parameters of the drivers. These designs are listener tweaked with multiple sets of electronics and cables. Others have 'cookbook' crossovers predicated only on the nominal driver R and receive minimal tweaking. Some of the these can be very good and achieve cult status simply because the designer got lucky.

OTOH, the engineered design could suffer from all manner of things beyond the engineers control: voice coil former, diaphragm, surround, capacitor, glue, gasket, laminate, etc. deterioration or a bad business model and simply fade away.

Some designs are so egregious as to be almost unlistenable but are still capable of resolving differences between preceding components and so, as they get better, the improvements are detected.

It is uncharitable to lay blame for one's inability to accurately predict the future