Are We There Yet?

Not that I want to obsess about it but the news just  keeps  coming. I know these hype trains roll around with tedious regularity, only to end up shunted off into some cultural siding, rusting reminders of promises unfulfilled, but occasionally one just keeps picking up speed and momentum until it's hard to hear anything above the singing of the rails as it speeds by on its way to the future.

While many of the claims being made may be overstated, many of the reports exaggerated, there are already so many active applications of the process in play it would take a major effort of will to deny the reality that's already upon us. The future can look after itself; this is the present talking.

Of course, just as with most of the insanely over-optimistic claims made on behalf of the metaverse, much of what's being said and written about AI is pure nonsense. Whatever merits either may have are not helped by the crazed, evangelistic language and wildly hyperbolic assertions made by their most fervent supporters.

The evidence, however, speaks for itself. For the imaginary metaverse so heavily promoted by shills for crypto and blockchain, sad, deluded failures like Decentraland do service as the very opposite of proof of concept,offering up ruined monuments to greed, hubris and self-delusion, as Dan Olson's lengthy and entertaining takedown takes almost too much pleasure in describing. Meanwhile, anguished, aggrandized pleas for recognition of AI sentience come across more as cries for help from those that make them than any remotely objective assessment of current technological capability.

It's all too easy in the current, fevered climate to make assumptions that lurch too far to the extreme, to believe it's either all a crock or a scam or a farce or the singularity or the apocalypse. It's none of these things. Then again, I'm not going to be the one to assure you there aren't crooks, scammers and farceurs at work here too. They're always with us.

Backtracking just a little, where I would take issue with Dan Olson is in his rubbishing of the notion that to believe the metaverse is already here is in some way to give in to vaguery that renders the term meaningless. Nope. It's here alright. Like AI, the metaverse is observably present. It just doesn't look like the photos in its press pack.

The metaverse is clearly the worldwide web we all use every day. It already fulfills all of the necessary criteria - instant communication anywhere at any time with anyone; immediate access to services and products; communal virtual spaces; avatars, abstracts, anonymity. The pieces are in place. The only thing that's likely or that needs to change substantively is our mode of access which, as Dan rightly explains, will not magically arrive as a world filled with people happily strapping half-bricks to their foreheads.

Virtual reality of the kind imagined by its zealots will have to wait for technical solutions several orders of magnitude more strange and rare than anything we've seen so far. My own feeling is that we're far more likely to adopt something along the lines of the augmented reality seen in The Expanse, where schematics replace flat screens via a form of projection, than anything like Star Trek's holodeck, but there's no need for either. The flat screens we have now will do just fine.

As the metaverse turns out to be nothing more than a fanciful rebranding of the digital revolution that began thirty years ago, so AI may end up being nothing more exotic than a collection of software we employ to reduce the amount of tedious busywork we've been mislabeling "creativity" for far too long. Ever since the industrial revolution, engineers have been automating processes that used to require manual labor; now the software engineers are getting in on the act.

Much of the negativity I've seen concerning the advance of generative AI into areas previously handled by low-wage, low-status workers has revolved around the inherent uselessness of the product itself. The argument seems to be that AIs are only capable of churning out reams of twaddle, which might well be used to replace the twaddle previously produced by people at a much slower and costlier rate, and that this is somehow a bad thing because it will put the twaddlers out of a job.  

At best it seems like something of a circular argument. If we don't value the output, why do we value the means of its production? And if we consider producing it a worthwhile way for someone to earn a living, why do we not value the work they produce? Is it humanitarian to want to reserve an occupation for human use so those without better prospects can keep themselves alive by producing work no-one wants or values?

Don't look at me. I did English. I think you want a philosophy grad or a sociologist or maybe even an economist. I just find it a bit unconvincing as an argument, that's all, which may well be because I don't know what I'm talking about. Wouldn't be the first time.

I also don't find it all that interesting. I'm not that bothered about who writes copy I don't read for  websites I don't visit. AIs or interns, it's all a bit too abstract for me.

A fight in which I do have more of a dog, and one which ought to be of particular interest to residents of this part of the blogosphere, might be the potential incursion of AI into gaming. There's been some talk of that of late, not least in the abrupt redirection of resources intended for the "social mmorpg" Galaxies of Eden into a business dedicated instead to "the creation of cybernetic beings and eventually infinite virtual universes."

I first read about this at MassivelyOP, where the tone of the reporting was decidedly cool. I'm used to MOP being heavily down on anything AI or metaverse adjacent, so I went directly to the source to make up my own mind. In this case I think I'd have to acknowledge the sarcasm and disdain is probably justified. I refer you to the embedded video in the linked page if you require further convincing.

It's my feeling that generative AI will come to play a substantial role in the making of video games, especially rpgs. The two seem to be made for each other. Far from feeling apprehensive about how they might interact, I'm very keen to see them strike sparks off each other. I do not, however, believe that, as claimed by the developers of Galaxies of Eden, "We are now at the precipice of attaining true AI sentience and infinite content generation". In your dreams.

They also claim to have "managed to craft a crudely sentient cybernetic being", something I more than doubt, even with the most generous definitions of "sentient" and "being". Or "crudely", for that matter. As usual, there seems to be a self-destructive desire to run at breakneck speed into the future, not just before mastering walking but at the very moment when the first, tottering steps are being taken.

A much more reasoned and objective experiment with AI integration in an mmo - or in this case a text-based MUD - was reported by New World Notes, a blog that styles itself "The world's longest-running metaverse news and culture site." I won't attempt to precis the full article; it's worth reading in full, including the transcripts of some of the ChatGPT-generated NPC conversations that took place in game. They are impressive, at least if you're the kind of person who'd normally impressed by roleplayers larking around. 

If nothing else, it demonstrates that ChatGPT can certainly improvise as well as most casual roleplayers, admittedly not the highest of bars. As Matt Mihaly of Iron Realms, the company behind Aetolia, the MUD in question, makes abundantly clear, however, any widespread integration between commercially available generative AI programs and mmos would be prohibitively expensive right now. 

What probably wouldn't be either impractical or costly, however, would be asking something like Chat GPT or Bard to write static quests. In fact, why speculate? In the immortal words of Luci from Disenchantment, as recorded on the T-shirt I'm wearing right now... Just Do It!

My prompt: 

 

"Here is an example of how quest dialog is formulated in EverQuest II:

(Which I've cut because it's far too long for the post - I used the first section of dialog here if you really want to know.)

Here is the structure of a short quest in the game:

" Speak to Gherrana Cobbleblork at the Highland Outposts to begin this quest. She asks you to "gather some feathers from the kragploom aviaks".

    Kill 10 of any type of Kragplooms.
        The closest ones are Kragploom chicks in The Highlands.
        More Kragplooms can be found on top of Kragploom Butte north-northeast of The Highlands. ( 91,223,197 )

Return to Gherrana for your reward. "

Please construct a short quest for the game in which a non-player character asks a player character to go somewhere and do something for them in similar style. Include names and some dialog to give the quest flavor.

 

And here's Bard's response:

"Sure, here is a short quest for EverQuest II:

Quest Name: The Lost Key

Quest Giver: Agnar the Smith

Location: The Commonlands

Quest Description:

Agnar the Smith has lost his key to his workshop. He asks you to find it for him.

Quest Steps:

1. Speak to Agnar the Smith in the Commonlands.
2. Go to the nearby cave and search for the key.
3. Return the key to Agnar the Smith.

Reward:

100 platinum pieces

Dialog:

• You say, "Hail, Agnar the Smith."

• Agnar the Smith says, "Greetings, traveler. What can I do for you today?"

• You say, "I've heard you lost your key."

• Agnar the Smith sighs. "Yes, I did. I was working on a new sword when I accidentally dropped it into the nearby cave. I've looked everywhere, but I can't seem to find it."

• You say, "I'll go look for it."

• Agnar the Smith smiles. "Thank you, traveler. I'll be forever in your debt."

• You go to the nearby cave and search for the key. After a few minutes, you find it on the ground near a pile of rocks.

• You return to Agnar the Smith.

• You say, "I found your key."

• Agnar the Smith is overjoyed. "Thank you so much! I don't know what I would have done without you."

• Agnar the Smith gives you 100 platinum pieces.

• Agnar the Smith says, "Please, take this as a token of my gratitude."

• You say, "You're welcome. I'm glad I could help."

That took Bard about five seconds and can anyone honestly say its substantively inferior to any run-of-the-mill mmorpg quest? If anything, it already assumes too much backstory compared to the average "Fetch" quest. 

 

I'm not suggesting the EQII team would be able to copy and paste the whole thing into the game verbatim - you probably wouldn't want to be handing out 100 platinum pieces for a Commonlands quest, for a start and I'm not sure where the cave might be - but I am saying it wouldn't require much more than a little tidying up before it could be slotted in alongside most of the existing low-level quests in that zone without anyone being any the wiser.

To my mind, it's not a matter of having AIs replace designers or writers so much as giving those people a powerful new tool that could make their work a lot easier - and quicker. If nothing else, it ought to make a big difference to those one-dev shops I keep reading about.

 

Maybe we will eventually be able to leave the AIs to it and all just sit back and consume what they produce but long before that paradisaical day arrives I suspect we'll be enjoying their human-moderated output without even realizing.

Much the way we spend much of our time in the metaverse every day of our lives right now.