I would love to see a reworked version of the Experimental Elixirs, though you can count me among those who actually like the idea of it being random, as I think the emphasis should be on experimentation, but with a repeatable aspect.

I've considered a homebrew system something like the following:

• Preparing elixirs is done before you take a long rest. Elixirs expire after a number of days equal to your Intelligence modifier (minimum of 1).
• When preparing elixirs the number you can prepare (1 to start, 2 at 6th, 3 at 15th) is the maximum number you can prepare per day.
• When preparing an elixir it can either be experimental, or you can spend a spell slot to recreate an elixir that you have previously made.
• The experimental table should have around twenty entries, with a mix of positive and negative effects. To generate an elixir you roll a number of dice equal to the number of elixirs you can prepare, and can spend a spell slot to add one additional dice. You may re-roll a number of dice equal to your Intelligence modifier (you can re-roll the same dice multiple times this way), but you must use the final result. Doubles, triples and quadruples may result in more powerful versions of an effect. Elixirs are uniquely identified by their effect(s) and magnitude(s), you should keep a note of which elixirs you have produced so that you can choose to recreate them.

Not sure what what the effects should be exactly, probably the six given plus maybe five more, and nine negative effects, giving a d20 table that's slightly skewed towards more positive effects than negative? The "negative" results also wouldn't be strictly negative, but rather should be things that could potentially be used as poisons, though if you roll both a positive and negative effect then you're probably going to end up with an elixir you might not want to use.

The basic idea is that instead of a simplistic random roll, you have a system that's more like a "try your luck" game due to the additional dice and re-rolls as you level up, giving you more control over a still somewhat random result. The aim is to encourage an Alchemist to always be producing experimental elixirs in the hopes of giving them access to more known elixirs that they can recreate later. The ability to retain elixirs for a few days means that you can afford to delay replacing elixirs that you like to have in order to produce some experiments, this compensates for the more limited number that you can produce per day. At your DM's discretion you may be able to find or buy recipes that allow you to learn elixirs that you're struggling to produce by yourself.

Just an idea anyway.

I don't think the Alchemist needs Extra Attack, though an elixir effect granting a short Haste could be interesting.

I'ma point out, Haravikk, that The Player Community thought keeping track of the size of one die, that changed sizes only occasionally, was far too complex and rules-intensive when they told Wizards to go screw themselves with psychic classes and the Psi Die.  A d20 table of good and bad elixirs with a requirement to separately track the duration of each existing elixir and track separately from that which of the d20 results you can waste a spell slot to create and which you can't...

Proooolly not gonna fly.

Quote from Yurei1453 >>

I'ma point out, Haravikk, that The Player Community thought keeping track of the size of one die, that changed sizes only occasionally, was far too complex and rules-intensive when they told Wizards to go screw themselves with psychic classes and the Psi Die.  A d20 table of good and bad elixirs with a requirement to separately track the duration of each existing elixir and track separately from that which of the d20 results you can waste a spell slot to create and which you can't...

Proooolly not gonna fly.

If an Alchemist doesn't want to take notes, then they didn't want to be an Alchemist in the first place 😝

The alternative is to have some limit to how many elixirs you can have at once, e.g- number you can prepare times INT modifier (minimum of 1).

People didn't think the psychic die was too complex.  They thought that a mechanic where you were effectively punished by getting the maximum result was a bad mechanic.

You were also rewarded for getting a lowball result.  A mechanic which, while it certainly tasted sour at first blush, grew on many folks as time went on and people mulled it over and playtested it.  Sadlyl the survey came out fourteen seconds after the UA document was released, long before that process of "...y'know what?  yeah.  This kinda works as a good, unifying anchor mechanic for psychic abilities" could finish.  So now psychic abilities are just bad crappy spellcasting or random nonsensical magical superpowers with a coat of purple paint.  Because The Player Community Said So.

Ugh.

None of that is really relevant to the artificer though, I suppose.  As I said before, I don't think we're going to see any sort of updates or alterations for artificers save for an Expanded Spell List as the token 'hey, artificers get stuff too!' section of Class Feature Variants, and they're almost certainly not gonna do a damn thing to fix the alchemist for people who detest Wild Chemistry.  Artificers are still too new to be considered a Real Boi class like the PHB guys, sadly.  We don't get to have all the same cool fun toys.

Quote from Yurei1453 >>

None of that is really relevant to the artificer though, I suppose.  As I said before, I don't think we're going to see any sort of updates or alterations for artificers save for an Expanded Spell List as the token 'hey, artificers get stuff too!' section of Class Feature Variants, and they're almost certainly not gonna do a damn thing to fix the alchemist for people who detest Wild Chemistry.  Artificers are still too new to be considered a Real Boi class like the PHB guys, sadly.  We don't get to have all the same cool fun toys.

 I wouldn’t bet on that. They’re including updated alternate stuff for the Beast Master if the UA is any indication. If they did the same thing for the Alchemist, and just used one of the older tested mechanics (satchel, Stone, or special pet), they wouldn’t have had to retest it in UA. So if they stuck it in Tashas out of the blue, I wouldn’t have my shocked face on. I’m not predicting it or anything, I’m not even saying that it’s even particularly likely to happen. I just wouldn’t be completely surprised if it did though.

The Beastmaster itself isn't getting touched.  They're simply adding new critter stat blocks that socket into the Beastmaster's existing mechanics to try and make the old, jank-ass rules work better.  That's a lot less of an option for the alchemist, given that one of the explicit stated goals of CFVs was "ABSOLUTELY NO TOUCHY SUBCLASSES"

Is it possible they're saving more detailed sub-class updates for later? i.e- nail down the classes as a foundation then see what else still needs changes?

The issue with doing CFVs for subclasses is that not all subclasses are in the PHB.  Fixing the Undying, for example, would technically require a person to have three books - the PHB for base warlock rules, the Sword Coast book for the base Undying rules, and then whichever book had the subclass fix in it.  Wizards is pathologically terrified of Book Bloat and requiring players to have more than the Base Three Books plus whichever one optional supplement they're using.  Books like Tasha's Allspice Soup Pot and Xanathar's Rummage Sale are expected to stand completely on their own with nothing but what's in the Core Three to back them up.  Same with all the settings guides - Wildmount and Eberron are expected to require nothing whatsoever beyond the Core Three to work.

Subclass CFVs break that rule, and ebcause Book Bloat was one of the biggest issues with 3.5e, Wizards would rather just print new things entirely a'la "Undead Patron" than require players to have more books.

They're never going to "fix" a subclass.  They'd rather take the hit on that subclass falling flat and replace it with a new one than break the "promise" they made to 5e people that nothing beyond the Base Three would ever be 'required' for another book to work.  Sadly, this means none of the optional books are ever allowed to interact.  I'm still surprised to this day that Wizards so much as put a sentence acknowledging Xanathar's Rummage Sale in the Eberron book ("if you have Xanathar's Guide to Everything, these spells are on your list, and you can Replicate Infusion the common items from that book.")  That sort of interaction between optional books is expressly forbidden by the "promise" they made to players when they emergency-aborted 4e and switched to 5e.

I don't think you are completely correct here, Wizards has reprinted Races with added features in theros. Why not reprint Subclasses with added features/Variants? It wouldn't really break the Core+1 Book rule. So while I don't get the whole noise about the Alchemist in this threat (I play a Battle Smith) I don't think they are fully against expanding old subclasses.

Core +1 is normally only for AL play. I personally don't know anyone who uses the those rules unless forced to at a Convention. I NEVER use AL rules in my campaigns . . . EVER! And I have been playing/DM'ing D&D since 1981.

Of the few streams I have ever tried to watch (I much prefer to play or DM) I have never seen one that uses AL rules.

One of my very few pet peeves with DnDBeyond is that they have to bow down to the great gods of WotC when it comes to UA material when even WotC doesn't follow their own restrictions. Every UA article ever publicized since 5e came out is still available on the UA website. However, WotC FORCES companies like DnDBeyond to take away access to them after a certain length of time. You don't lose previously created characters using those rules, you just can't create new characters with those unless you can successfully homebrew them.

"Core + 1" is how they sell the game.  Any given 5e book on the shelf is designed to be bought by as many people as possible.  Every book you put in front of that book - every "prerequisite" book one has to own before they can use the book sitting there on the shelf - reduces the number of people who will buy that book.  Wizards absolutely cannot tolerate selling 5e books to some of the people instead of all of the people.  Ergo, nothing will break their promise of "as long as you have the Core Three, you'll never need another book to use anything else we sell".

Because that would reduce their sales, and reducing sales is the only true Sin to a corporate entity like Wizards.

That's because Wizards doesn't sell the game that way. They're not naive enough to think people aren't going to buy up as many books as possible. Options are what helps sell books. I don't have The Mythic Odyssey of Theros, yet, but both the Ravnica and Wildmount setting books make specific mention of subclass options found in Xanathar's Guide to Everything. There's a PHB+1 rule for DDAL, but that's just for players. And only certain books are DDAL legal. Tasha's is including reprints of the Ravnica and Theros subclasses to open those options up for DDAL play.

And for those of us who are truly strapped for cash, the basic rules are free. It's possible to run entire campaigns with just that pdf.

That said, the bloat that plagued 3/3.5 and 4th edition is starting to rear its ugly head. As smart a move as the reprints are,  those of us who only bought previous setting books (Ravnica, Theros) for the subclasses therein may feel like they wasted money. And now to get their money's worth, they need to start running campaigns in those settings. It's not impossible, but I do think it's fret with its own problems. Moving up the tiers in Ravnica, for example, is a lot harder than, say, on Eberron or the Sword Coast. At least that's how I see it.

They try to follow that rule, in interviews they often talk about it. I thinkl it isn't ab bad thing they want to avoid book bloat in 5e so it stays accessible for new players. But they somewhat broke the rules by including Xanathars spells and subclasses in the setting books, so I think Tasha's as a book in Xanathars footsteps will weaken that rule,too. Also tashas has so many addendums to the basic PHB rules, that it could be considered a PHB - Part2(especially by having the Artificer in it) so I guess they will include some stuff from Tashas in the next books.

A simple test of how good the Alchemist is would be the number of DDB accounts using it. I dare say there's more Armorers than Alchemists. Their level 3 class features are a very, very lopsided comparison.

Don't forget, an Artificer can learn Alchemy Kit proficiency as a downtime activity, as well as the new Poisoner feat. They can still be a small-a alchemist.

That's not really a test of how good anything is. In a situation where not all content is free, the most popular options are always going to be the free options, which is why the most popular subclass for each class is the one in the basic rules. Even if you filter it down to people who have all content unlocked, the most played subclasses are not inherently the best subclasses. It'll be the subclasses that have the most appeal to more people, which is impacted by more than "Who has better numbers". Anyone who has access to any subclass of artificer has access to armourer right now. If there's any reason for alchemist to be less popular than other artificer subclasses, it's because it's seen as the more supportive subclass. Anyone who has played any role/class based game knows that supportive and tank roles are always less popular than dps roles.

Tl;dr: The most popular option in any situation is not inherently the best option.

The other simple test of how good the Alchemist is would be to check and see how good the Alchemist is.

Lemme go double-check the wording again quick.

.
..
...

...nope.  Alchemist still absolutely terrible.  Third-level features less impactful and more restrictive than all other artificer subclasses.  Fifth-level combat aid still the weakest, least useful such feature in the entire class outside janky multiclass 'make Healing Word heal for 35hp from a first-level slot' nonsense.  Higher-level features solid but sabotaged by godawful design of generally more important and defining early class features.  Subclass still, mechanically speaking, a total disaster.