that’s really not the case though … if you’ve ever looked at the way e.g. the sysex dumps are packaged and reshuffled every time a feature is added you’d appreciate there’s often a lot of consequential work beyond what may seem simple conceptually - this applies on so many layers
it’s disingenuous to believe it took about half an hour to do the randomizer although that’s significantly simpler - we shouldn’t trivialise how much time/effort/resources it would really take to implement a second LFO or other feature … also it may be best not to overlook whether it would just go against their philosophy for the instrument - i.e. keeping it 1 layer deep (distinct from the A4 with its greater complexity) (note also that the AR eschews much of the resolution available on the A4 for fine tuning lots of stuff - it’s a lot simpler)
Having said all that a second lfo (or even just a second ‘destination’ ) would open up so much more - but it won’t happen at the cost of anything - the fx levels ‘fix’ shows that !
but then again, changing the sample start/end paradigm and offering start/len or offering a new modulation destination of start&len would improve things too … there’s a tonne of feature requests that will never happen including some with no ‘cost’ - like binning the terrible default attribution of +18 to filter freq for pad pressure - to me this seems like it’s a no brainer, takes no time, but I’m probably under-thinking the consequence to suit my preferred outcome