Duke Ellington is remembered, among other things, as a composer who wrote for specific musicians and personalities rather than for some nameless abstract conception of what given instruments could or should sound like. If this is so, he must have worried more about the present than the future, since others would continue to play his music long after he and the musicians he wrote for were gone. Perhaps the advent of recording technology (which happened within his lifetime) made this approach more palatable, since the work of the specific bands he wrote for could then be documented conclusively; or perhaps it was simply his intuitive way of writing and he never gave a second thought to how he would be remembered. The least likely explanation, though, would seem to be that he wished to subvert the traditional hierarchy that existed in European music between composer and performer; that it really bothered him when the person who wrote the music got all the credit; that he set out from the beginning to mitigate the influence of the composer's ego from the music making process in a quasi-Cagean way.
I'm not an Ellington expert, nor a worshipper, nor a detractor, and have no axe to grind with his work or legacy, but because no one (and I mean no one) misses an opportunity to invoke this textbook one-liner about his compositional style, I think it's worth problematizing. To be a bit less rosy about it for just a second, one might also say that allowing your players to simply "do their thing" means not asking them to venture outside of their comfort zone. It means setting parameters that make for predicable results and minimal difficulty in achieving them, the polar opposite of a full-fledged experiment where the outcome is not known at the outset.
Musicians tend to spend the majority of their time working somewhere in the middle of this spectrum, in a space where the musical parameters are overwhelmingly familiar but not entirely so. Where we feel compelled to move away from this more common mode of working and towards either extreme (i.e. merely going entirely with what we know, or conversely, more thoroughly subjecting ourselves to the whims of chance), this is usually a consequence of externally imposed conditions, such as being hired to play a wedding ceremony, or being thrown into a jam session comprised of people you've never heard or played with. The difference between these two types of contexts is not that one is destined to succeed and the other to fail, but merely that we know what to expect from one but not the other. One offers comfort without the potential for growth; the other offers the opposite.
A blind allegiance to unfamiliar, difficult situations in the name of musical and personal growth is meaningless in absence of an eventual outlet for cashing in on that growth, the products of which are necessarily obtained by way of tasks which have become familiar through experience. Similarly, the outright absence from one's musical menu of any potential for growth in the form of unfamiliar or experimental musical contexts is a sure recipe for a hardening of the arteries. As musicians, we do our best to balance these two concerns in our work, striving to offer safe but competent musical products when that response is appropriate, as well as throwing caution to the wind when we see the potential for something new and exciting to take shape.
Knowing when to do which, though, can be tricky and contentious, and there's great danger in distilling the approach of an artist like Ellington, who had a long, eminent, and prolific career, into single-sentence pieces of advice. I doubt very much that Ellington himself would approve of the way many musicians now invoke his most famous compositional tenet merely as an excuse for not rocking the boat, for not making people uncomfortable, not forcing them to grow, not putting them in a position where they might fail; all of this in the service of some vague and ultimately overly idealistic aspiration to an all-encompassing collectivity where individual artistic voice is neutered in the name of group consensus.
Ellington indeed wrote for people rather than conceits, but the music is undeniably his. Just because voice might be a source of egotistical attachment to our own music doesn't mean it's worth subduing in polite musical company. To revert to one of my favorite metaphors for a second, I consider voice to be the essential source of biodiversity in a musical ecosystem: without a diversity of voice surrounding us, we become creatively inbred and defective. The ego is, indeed, equally destructive in the social realm of music making, but there are more attractive ways around that inconvenient fact than simply abandoning any compositional device that might harm our colleagues' buzz.