What exactly is he suggesting using AI for? A busy, thankless, unpaid, DM using AI to flesh out the details of a campaign is a good use of ai. But what is he - the CEO of Hasbro - going to do with AI?
WotC desperately wants to get customers comfortable with paying for AI generated images. They could effectively gut the expensive art departments. A major use for AI would be for Hasbro's card games. They already pay like shit. If the option to make "art" for nothing becomes available, they'll do it.
Gut the art departments, arguably some of what very little value actually remains in WotC outside of their clawing corporate avarice. Great idea great idea
IIRC what they're really angling for is getting out of publishing books to begin with and instead selling AI GMs as a service who just play what may as well be calvinball with the players based on a secret ruleset if any at all. Edit: to add onto this, IIRC the reason 5e lacked the massive amount of splatbooks they relied on in earlier editions was because they really wanted to shift towards selling digital services and wanted to get away from things like "people playing the game having a copy of the rules" in favor of them just like, buying an NPC or an item or some shit for a virtual tabletop, because they really want to be a shitty microtransaction-filled MMO but without all the "investing resources into actually making and running that sort of thing" part.
So basically they're trying to be AIDungeon even though that failed miserably because people only wanted it for porn and also the concept was untenable as a serious thing and also still is completely untenable.
In other words they should go for it, put all their eggs in that incredibly stupid basket, and stop making RPGs at all. I am saying this because I want what's best for them, obviously, and not because I want to see D&D finally come to an end and make way for other, better systems. Definitely.
4E was the one where they wanted to shift to digital services, but they didn't ever materialize. (In case you're wondering why it never happened - I was on the job market in Seattle at the time; they were offering a 60k salary for a senior engineer with a combo of rare talents in the same city Amazon was offering 90k to a junior engineer.)
5E was a reaction to 4E being hated by a large portion of their fans, and then Stranger Things and Critical Role resulted in them lucking into the most successful D&D edition yet, so I think their policy was mostly "Don't do anything, you might break it again."
Latest Pathfinder, though chunky rules wise given it's origin, has been doing a good job taking the 2nd place of best known RPG especially after the whole WoTC going after third party content creators and pissing all over the community. There's also a new boom in indie TTRPGs which is nice.
How's Pathfinder? People seem to like it, but I've also heard it described as number-crunchy like 3.5e which I think would be a step in the wrong direction.
Pathfinder 2e isn't number crunchy, but it is rules-heavy.
With numbers, the specifically max out the counting by limiting every roll to d20 + Circumstance bonus + Status Bonus + Item Bonus. The exceptions to this are quite rare. If you get more than one bonus of a given type, pick the highest number. So you don't get huge number stacks you have to add.
There are lots of little rules like falling damage, or grappling rules, or counterspelling, that are generally easily referenced with robust fan-made officially sanctioned reference tools.
People who tell you it's less complicated than 5e are lying, but its core elements are super elegant and play really well. Expect a tactical game in combat, nicely streamlined exploration, and lotsa fun choices for players to make during chargen and during combat
Yeah it's more complicated that 5e, but 5e is often not complicated enough where it needs to be, which is still why my group is largely still on 3.5e (and also we have all the books
Pathfinder 2e is basically just D&D done better than D&D itself has ever done. It's like if 3.5e had been refined into a more streamlined and modern game with a bunch of the bad-design-decisions that are baked into the fabric of D&D mitigated as much as they possibly can without just ripping them out, and they found a way to keep the variety and customizability of 3.5e but without as many exponentially compounding consequences.
Basically, where D&D went down this path of flattening out the splat and turning characters into basically straight line level up progressions with as few choices as possible and no room for character building (that is, you choose your class, your major class feature, and then that's it you're done you just level up and never think about anything again unless you're a magic user who gets to choose which spell you want to be allowed to cast once per day), Pathfinder 2e went down the path of pruning the sprawl of options down to a few major branches for each class, and then fleshed that out with a lot of little choices that aren't particularly meaningful but which leave tons of room for flavorful character building and which make sure that every level everyone is making some sort of choice no matter how small and fluffy that choice is.
AIDungeon was such a hilarious shitshow, for all that it was an entertaining toy for a while. Honestly I feel a baffling sort of nostalgia for the GPT-2 era of chatbots where it was new and good enough to be interesting but bad enough to be incredibly funny. It was so much better than the bland slop modern chatbots are where they're bad enough to be useless and good enough to be boring. It's like schlocky grindhouse B-movie slop vs modern overcooked Hollywood slop.
I hate proprietary AI models so much. They're all the material and social problems of open source ones, but entirely in the hands of the dullest and most ontologically evil corporate suits alive. Trying to make a machine that's the sum of decades or centuries of human culture and which extrudes a fine paste somewhat resembling the same, that's obviously a very mixed bag at best, but when that paste is just going straight into a selection of corporate injection molds there are no possible positives left, only the complete enclosure of human culture and its replacement with injection molded slop.