gosokkyu

エンド

  • 戦う人間発電所

owatte shimatta


IGN Japan just published a two-part roundtable with Yuji Horii, Hiroyuki Imabayashi (founder of Thinking Rabbit and creator of early ADV like Kagiana Satsujin Jiken, and also Sokoban!) and Takemaru Abiko (writer of Kamaitachi no Yoru and contributor to games like 428), plus Square-Enix AI researcher Yusuke Mori; the first part is a discussion about their classic ADV works and the parallels between old-school test-parser ADV and what's commonly known as "AI", and the second part being an examination of the modern applications, strengths and weaknesses of AI as applied to ADV design, with references to the latest iteration of Square's "NLP Adventure" project. The first part's not especially distasteful and well worth reading; the second part is an actual discussion and not just old dudes ooh-ahh-ing at a new toy, but I can't say I found it especially compelling.

As you may have suspected, this is essentially a sponsored promotional piece for S-E's AI division, but the circumstances are a little more complicated: y'see, S-E recently started an internal initiative to catalogue and archive their legacy dev materials (including documents, art assets and source code) across their various companies with the goal of capitalising on them in some way, in a manner not dissimilar to recent initiatives by Bandai-Namco, Sega and Capcom. Unlike those companies, S-E saw fit to leave this task to their AI research division, with the explicit angle of framing these assets as representative of significant technical milestones that underpin contemporary research efforts, and it's why they're doing stuff like releasing AI apps with a Portopia skin, or why their "look at these docs we found!" CEDEC presentations from the last few years all have fig-leaf connections to contemporary tech trends, or why their recent paid promotion with denfami on the dev suite used to make Tactics Ogre carried the immediately ridiculous assertion that it was the world's first game engine—sometimes neat ol' junk is just neat ol' junk, and even the people involved are fully aware of this fact, but they've been given an order and they're sticking to it.

I wonder, is it working? By that I mean, do the people who control the fate of this initiative believe it's worth whatever meagre budget they've afforded the people doing the work? From the outside, it's hard to gauge whether any of them are "successful": Namco seems content to just slap all their new high-res scans onto print-on-demand shirts, but their video feature initiative didn't last long; Capcom's latest initiative seems to begin and end at using classic art to farm for retweets; Sega Forever's recent archival projects seem less like a bona fide initiative and more like a directionless brand scrambling for purpose, and one that didn't even do a great job of just properly explaining what it was they were doing. (The recent layoffs and other changes at Sega Europe seem to have killed Sega Forever, but who knows.)

Square's AI team has been at this nonsense for a minute so I presume somebody sees it as an effective means of commoditising their archival material, and if that's what it takes to get the work done then I'm willing to put up with the silliness, but it'd be nice if these big orgs would commit to genuine archival efforts instead of mere value extraction—or, at the very least, would be more supportive of independent orgs that might do this work for them.


You must log in to comment.