I don't see why a Darkrai suspect wouldn't be reasonable, or Iron Valiant suspect wouldn't be reasonable. Maybe without them this whole Sleep thing dies down, or heck maybe we leave things as be and in a months time Hypnosis abuse is a thing of the past, much like the month of Quick Claw Quick Draw.
It seems like the level of conversation this has drawn (now and in the past) shows that the topic of Sleep is simmering beneath the surface for a lot of folks even when there aren't notable abusers around. We've gotten used to it because, something close to 100% of the time, sets running non-Spore sleep moves are just fishy bullshit. You run Hypnosis Alotales/Iron Valiant/Darkrai because you want that rush of lulz when it hits, but even with Darkrai, the Hypnosis sets aren't the most consistent, they're just the most obscene when they work.
The reason this gets so much discussion is because of the weirdness of Sleep Clause/Sleep Mod and the way that it creates artificial ways to "counter" sleep, and thereby completely upends with the value of RNG. With no Sleep Clause/Sleep Mod, there is no "sleep sac", so there's no reason to make an otherwise sub-optimal play by switching a less useful 'mon for the MU into a potential sleep-inducing move -- which means there's no upside to having Hypnosis miss, thereby "preserving" your sleep attempt for later.
Because of Sleep Mod, any comparisons between Hypnosis and Focus Blast (or other moves that trade accuracy for power) is meaningless because any other move is allowed to connect more than one time. You don't lose the ability to click Focus Blast on a predicted Gambit just because you connected with one when they switched in Zapdos earlier.
So, basically, folks know in their guts that we're on something of a Slippery Slope Continental Divide*.
On one side, you have "preserving" a mechanic in a way that feels bad no matter what you do -- either letting Sleep run unrestricted and leaving the entire game open to cheese strats or by implementing a rule that fundamentally changes the definition of "good plays", "bad plays", "good results" and "bad results".
On the other side, you have, essentially, the rejection of a long-lived game mechanic due to its fundamental uncompetitiveness and the frustrating questions that come with that decision about other, similar mechanics that are less easy to jettison and the fundamental role of RNG at the heart of Pokemon.
Not gonna pretend I have the "right" answer here, so I'll just echo/expand on something that (I think)
DaddyBuzzwole said: if anybody out there wants to make a mod that's the Pokemon equivalent of "No Items, Fox Only, Final Destination" and see how it turns out, please do! I'd be deeply fascinated to see how folks respond to 'mons with massively reduced random element.
*Note, neither of these are actually slippery slopes, the phrase was just too fun not to use. The choice to implement Sleep Mod as it is, setting the precedent of directly altering game mechanics, is the closest thing in this mess to an actual Slippery Slope, and it makes sense that it's something that many people seem to passionately want un-done in some way.