D&D General The DM Shortage

[Reynard]

That's how the book you cited in the OP told people to run them.

You keep repeating this like it is true when it is demonstrably not so.

 

[Paul Farquhar]

Talk to a chat bot for a while. After a while, you'll be able to tell whether it's giving you a random answer or a reasoned one. The more you know about how they work, the faster you can pin it down.

The same thing would happen even faster if you put a chat bot type "intelligence" in charge of your dungeon.

Would you tell if a human DM was giving you a random answer? There plenty of occasions a player will ask me something I don't know the answer to so I just say something random. Like the troll question, which I would have been perfectly happy to let the player answer themselves.

 

[Paul Farquhar]

You keep repeating this like it is true when it is demonstrably not so.

It says that in the RULES section, but the example dungeon you praise so highly TEACHES the opposite (You might almost think more than one person worked on the book).

 

[Celebrim]

Unless you're a certain ex-Google employee, no one believes that AI actually "thinks" in any way that we would consider it thinking. Without that it will never feel "real".

The big problem we have right now with the neural network approaches is that they can't really remember any of the results of their own thoughts, and stashing those facts actually runs contrary to the way the existing algorithms work. The existing algorithms are basically stateless and it's not clear how we establish and remember new state - certainly not at any speed.

It's clear that we are getting where we can fake human creativity, but it's equally clear to me that most of the ways we are doing it is a dead end. We're going to need another improvement in our algorithm before we're going to get close to being able to consistently trick humans for any length of time.

I'm sure the guys doing great work with generative art right now would love to be able to establish as a fact that humans have five fingers, but it's not at all clear how they could do that because the computer artist isn't reasoning in the sense we understand it.

 

[Greg Benage]

Would you tell if a human DM was giving you a random answer?

Over time, you can tell the difference between a DM using judgment and a DM just giving random answers, if for no other reason than consistency.

 

[Celebrim]

Would you tell if a human DM was giving you a random answer?

With a sample set of one, no. But I could then ask you more questions about that answer to see if you were self-consistent in your answer, which, being a human you could achieve if you tried. An AI like we have now can't be.

Let me try to explain. If you talk to a chat bot:

Me: "What food do you not like?"
Bot: "I'm not fond of beets" <- answer could be random or reasonable

If you go something like:

Me: "Why do you not like beets?"
Bot: "They are disgusting" <- answer could be random or reasonable

But if you were instead to go:

Me: "What do beets taste like to you?"
Bot: "They are delicious" <- answer not reasonable and probably random

Ok, but there is a good percentage chance you'd get "They taste disgusting" randomly, which is self-consistent. You could test the bot a few times and if it keeps changing it's mind, well it's a bot not a person.

But there is an even faster way to get at this. You answer the bot like you were a bot.

Me: "You have always enjoyed beets."
Bot: "Yes I have. I love beets." <- answer not reasonable

 

[Reynard]

AI can emulate simple conversations because it has a databank of questions and responses gleaned from some source. An AI could be taught to learn the rules of D&D, just like it's been taught to learn chess. But holding a real conversation? You'll hit limits. Actually creating entire campaigns from scratch? Logically responding to what the PCs do with the flexibility that a DM can? Nope. I don't see it happening any time soon.

I was actually thinking being able to run a group of players througha module sufficient to entertain them, not become sentient or be a perfect simulacra of a decent human DM. I don't think most players have that high of expectations of the game and therefore if they had to choose between "no game" and "Alexa as DM" they would choose the latter.

 

[Paul Farquhar]

With a sample set of one, no. But I could then ask you more questions about that answer to see if you were self-consistent in your answer, which, being a human you could achieve if you tried. An AI like we have now can't be.

An AI could store it's answers and regurgitate them. A human forgets. My memory is terrible, and I write too slowly to make detailed notes.

Of course, what are you doing here? Playing D&D or trying to tell if the DM is human or not?

 

[Reynard]

It says that in the RULES section, but the example dungeon you praise so highly TEACHES the opposite (You might almost think more than one person worked on the book).

In what way? Be specific.

 

[Paul Farquhar]

In what way? Be specific.

It has a dungeon. With rooms. Containing monsters. It doesn't suggest any way to interact with the monsters other than fighting them. If you have zero hit points, you die. If you fail a poison save, you die.

As presented in the example, D&D is purely a game of fighting monsters until you die.