However, you're still going to be placed under arrest for violating the electronics fashion police good design practice policies
But ossifer, ossifer!
Let me ask you this anyway:
If the LED runs for a year from now would you really want to do it this way in the future for an indicator that you really had to depend on or even a general purpose indicator, or would you prefer to use a tried and proven safe technique to run the LED?
Well, I'm tending to reply that while I wouldn't use a single LED, for reasons that have been explained here (power factor and wasted power in the reverse direction), I would go ahead and use 2 LEDs back-to-back, with no intervening resistor even.
I'm thinking of the case of a commodity-item power strip; I can now buy these things in my local dollar store. They're practically throw-away items. (Not that I like throwaway items; in fact, I hate them.) So ASS-U-MING that back-to-back LEDs with a capacitor would pass muster with the UL, then yes, I might consider doing just that.
We're not talking about "mission-critical" components here. I would certainly never suggest doing this in, say, a heart-rate monitor used in a hospital or some such.
I believe someone in that other recent thread on this pointed out that they saw just this in a commercial power strip or something.
Please, let's keep a little perspective here.
By the way, I nominate the circuit in
this post as the best example of overkill in lighting a LED directly from the poweline. I mean,
four diodes, two capacitors and a resistor to drive one lousy LED? Heck, why not go all the way and build a dedicated, regulated SMPS for that one little diode?
I suppose that's the way NASA would spec it ...