Well, as for backwards engineering to produce more selective benzos Girisopam does exactly that. If one was to further backwards engineer such compounds we would be able to find more selective compounds that affect both GABA and AMPA receptors. This would be incredibly useful for selective agonists/antagonists as they could be used to figure out which receptors the non-selective compounds bind to.
If we look at
Tofisopam we can see another selective 2, 3-benzo. It is also a PDE10A inhibitor, so this link could make a path to other PDE receptor ligands. These attributes make 2,3-benzos extrememly useful in working out link between the Glutamate receptors and the GABA receptors. If more work and research went into these compounds, the perhaps the whole field could be joined by selective and non selective compounds, thus opening up the way for non selective anxiolytiocs and nootropics. Such compounds would dwarf the nootropics we have at the moment providing both anxiolysis and nootropic properties.
Imagine, if you will, a compound that relaxes you enough to be more confident and enhance cognitive processes without the flaws of traditional benzos and with cognitive enhancing abilities far surpassing the compounds that are available today. This would bring us closer to such a compound that could be adequate to increase work-power by a multitude of factors (hesitant to say but..) similar to NZT from the film Limitless, a goal which every nootropic researcher wishes to find.
These compounds must be fully explored to unlock their true potential but as of now they are limited in their field to very few compounds (the three mentioned previously, mainly). In my opinion this is an incredibly exciting field, which we know very little about.