There's a whole other post to be written about my hopes for the use of resistance, immunity, and vulnerability moving forward in D&D, but I think there's a special case to be considered: Poison damage.
In the Monster Manual, there are 343 stat blocks that do not have any resistance or immunity to poison. The Monster Manual has a total of 450 stat blocks, giving us 107 that do have these resistances or immunities, or about 24% - nearly a quarter of all monsters in the book.
Among those 107 that have either resistance or immunity to poison, a full 102 of them have full immunity - the handful that do take half-damage include the single "Duergar" stat block, the Assassin (who I guess has worked with poison so much that they've built a tolerance to it?) and the Incubus and Succubus (which are actually the same creature, just with different, gendered names) and the Cambion.
The point here is that of all damage types, poison is the least reliable. If you are facing the vast majority of fiends, the vast majority of undead (oddly, vampires are susceptible,) or the vast majority of constructs (only Modrons are not immune,) your Poison Spray cantrip is going to be utterly worthless.
Now, poison is often coded as a "bad guy" damage type. While Radiant damage is associated mostly with angelic, holy forces, poison is the classic "underhanded" way to harm someone.
But it's also something that player characters can often avoid as well. Dwarves (including Duergar), Green Dragonborn, Autognomes, Stout Halflings, Plasmoids, Reborn, Warforged, and Yuan-Ti are all resistant (the old version of Yuan-Ti Pureblood is fully immune) and Monks all get full immunity to poison damage eventually.
Now, logically, these usually make perfect sense. Fiends are beings of the hellish lower planes, where one imagines poison suffuses everything. Undead are already dead, so the biological mechanisms that a poison would attack or inhibit are already nonfunctional anyway. And Constructs don't have organic tissue that a poison would usually wreck.
But the end result is that this one damage type is singled out for utter uselessness in a ton of scenarios. And I think that leaves some design space on the table.
In World of Warcraft, the Rogue class has as one of its core ideas the use of poison - you apply these to your weapons, and it adds damage and some other effects. Of its three specializations (the WoW equivalent of subclasses,) the Assassination spec is built around abilities that enhance the effectiveness of your poisons.
Now, WoW kind of dispensed with the idea of damage resistance and immunity long ago (the original raid dungeon, Molten Core, was filled with fire elementals, forcing Mages who had chosen the Fire spec to swap over to Frost in order to be effective - today, you can hurt a fire elemental with fireballs even if that doesn't make a ton of sense, for game balance reasons).
But as it stands, unless you can be guaranteed that you won't be fighting a lot of constructs, undead, or fiends in your campaign, you really shouldn't build a character concept around using poison.
As it stands, the DMG has a ton of interesting poisons with different effects that you can employ against your players or NPCs. Many of these are ingested poisons - something you'd sneak into someone's meal rather than coating a weapon with it. But in terms of adventuring gear in the PHB, "basic poison" is the only thing that's really available, and it adds a piddling 1d4 poison damage - meaning that no one in their right mind would bother spending the 100 gold required for it (oh, and the target needs to only make a DC 10 Con save to not take any damage). For 100 gold, you could get two healing potions that each heal at absolute minimum the maximum amount of damage that this could do.
Now, as it stands, I think the only reasonable way for a player character to get poison as a piece of equipment is by using a Poisoner's Kit. The thing is, 5E has been kind of odd about its use of tools in the first place, and surely Alchemist's Supplies should also cover this, right?
One D&D does make the use of tools for ability checks far, far clearer (I've been using the new rules in my campaigns, and it works out great - our Artificer often gets 30s on Intelligence Arcana checks that can use Tinker's Tools, having 22 Intelligence, proficiency in Arcana, and Tool Expertise, which means they get a +16 to their checks and roll them with advantage).
Still, what we don't have are good crafting rules. The amount of downtime required for using the crafting rules found in Xanathar's is basically prohibitive in any campaign where there's any kind of ticking clock (which is most, I think). As it stands, I believe that it would take you two whole weeks to craft a single vial of basic poison - again, not remotely worth it. (Also, it would take 30 weeks - over half a year - to make a set of plate armor).
But we also don't want poison to become this necessity for anyone who fights with a weapon - if it's too good, character concepts that would never use poison will feel obligated to do so.
So, as it currently stands, player characters basically ignore, or should ignore, any spell or item that deals poison damage (there might be edge cases like spells such as Cloudkill when you know you're facing living, non-fiendish monsters). And even opting into it as a character who, logically, it would make sense to be all about that poison (such as an Assassin Rogue - who do automatically gain proficiency in Poisoner's Kits) is a fool's errand given how hard it is to squeeze any significant damage out of it and how costly it will be to get that tiny amount of damage.
So, what do we do?
First, I think, we need to be more liberal with what poison can effect.
Fiends are, I think, the prime candidate to lose this sort of immunity. Yes, the lower planes are filled with noxious fumes and fetid waters, but we could easily create environmental aspects to these places that fiends are simply immune to. Indeed, if we say that Devils are no longer immune to poison, it creates all manner of potential intrigue if rivals for power in the Nine Circles are engaged in a constant war of assassinations, including via poison.
If we don't want to go so far as to remove poison immunity from all fiends, we could limit it, instead, to, for example, demons. Demons, after all, tend to be the gross, unsanitary kind of fiends, and it could create some interesting scenarios in which a party might need to figure out which kind of fiend they're dealing with.
Next, I think we should take a look at undead. I think it's reasonable to say that incorporeal undead like ghosts and specters can have full immunity to poison. But corporeal undead like zombies, ghouls, and the like could either fully take it or perhaps only have resistance to it. I think here we'd also carve out an exception for Skeletons, who don't really have any soft tissue left to be affected.
The point is, if we can bring poison immunity and resistance down to the levels that, say, Fire or Lightning get, or in that ballpark, we could open the door for player characters to start actually wanting to use it.
Now, how should they do that?
Again, I don't think we want to make this so good that everyone who uses weapons will have to stock up on this stuff. So let's place some limitations:
Poison, as an item, should only be able to be applied to certain weapons. It should't be practical to slather one's greatsword or maul with it. But a dagger? A blowdart? An arrow? Sure, these all make sense to me.
Broadly, we could say that it only goes on finesse weapons or ranged weapons. We might limit it further, though: while I think the Pistol and Musket are not so powerful that they need the insanely higher cost they currently have in the playtest, I also think it would be fairly absurd to try to poison a bullet.
Another consideration is that usually, the more beneficial properties a weapons has, the lower its damage die (while detrimental properties, like Heavy, Two-Handed, or Loading, tend to increase them,) and if we are only allowing poison on certain types of weapon, we are effectively giving it a new property, albeit one that requires frequent expense to use.
So let's start very conservatively and say we can only use this with d4 weapons that deal piercing damage. That's just Daggers and Darts. As it stands, the only easily-available poison to player characters is a DC 10 Con save to take none of 1d4 poison damage on a hit. That's a save that will very frequently succeed. So we're adding barely any damage to our weapon.
If we make the save be for half damage, this does turn it into, on average, about 1 damage even on a success. Still doesn't seem worth the 100 gold.
So, I think we could take this one of two directions: one is that we keep poison expensive, but make it worth its cost. For 100 gold, we can get two healing potions, each of which heals for an average of 7 hit points (2d4+2), meaning a total of 14 restored hit points. So, it would seem reasonable that, for an equal investment, we should get at bare minimum a comparable amount of damage.
So, we spend 100 gold on our basic poison, and we should get an average of 14 damage from it. We could either get that from 4d4+4 or 4d6, or even 3d8 (which is 13.5, but pretty close). But what of the Con save?
The thing is, as it currently stands, there are two d20 tests involved in letting a Basic Poison actually do its thing. And that thing it does is a piddling 1d4 damage - not enough unless rolling max damage to kill a Commoner. We're not inflicting the poisoned condition - a much more potentially powerful effect - so here's a question: should there even be a saving throw?
Player character who get advantage on saves against the poisoned condition also tend to have resistance to poison damage, so I think in a lot of ways this is kind of double-dipping. We spent a hundred freaking gold on this! Again, a Commoner will avoid all of this damage 55% of the time, given the low DC for the save.
So I'm tempted to say that the damage should just come with the hit - if you get stabbed with a poisoned dagger, guess what? You're taking that poison damage.
Now, we could also imagine a scenario in which the poison is cheaper and does less damage, but I think then we start to get into "required gear" territory. If it's only 1 gold for poison that does a little bit of damage, you'll swiftly hit a point in which you will want to apply it to your weapons on every fight.
I think keeping it expensive makes it special - we now need to save that poison for the big fight.
Indeed, I think we could even have higher-level poisons show up, but like healing potions, we're not expecting to be able to buy these at just any adventuring supply store. And, like healing potions, I think these are going to be one-and-done when used. If you have it on a melee weapon, you'll have time (say an hour) to make use of it, but once it strikes a target, the poison is no longer on the blade.
Now, this is also where subclasses could come in: I think you could have a Rogue subclass that specializes in poisons. They might be able to more easily produce it - making use of a Poisoner's Kit, for example. The subclass could even derive special toxins that do other types of damage for the (now) rare cases where poison doesn't do anything.
The point is, I think there's some potential left on the table here for interesting and very thematic gameplay.