Very disappointed that this article ends just before it starts to get into anything concrete and interesting—I'm working on solving (a simplified subset of) very similar problems for a game I'm working on, and would love to learn more about the author's approach!
Inform 7 is actually kind of fun to write once you get the hang of it.
What's interesting though is, at least in my experience, LLMs are terrible at generating Inform 7 code specifically because it's so uncanny-valley-similar to English.
Imagine a world where Microsoft was pushing “Copilot” integration everywhere, just as they are in this one—but the proof was, actually, in the pudding. Windows was categorically improving, without regression, with each subsequent update. Long-standing frustrations with the operating system experience were gradually being ironed out. Parts of the system that were slow, frustrating, convoluted, or all three, were being thoughtfully redesigned without breaking backwards compatibility, and we were watching this all unfold in real time, in awe of the power of “AI”, eyes wide with hope for the future of software, and computing in general.
Think of how dramatically this hypothetical alternate reality differs from the one we live in, and then consider just how galling it is that these people have the nerve to piss on our leg and then tell us it's raining. Things are not getting better. This supposedly-magical new technology isn't observably improving things where it matters most—rather, it's demonstrably hastening the decline of the baseline day-to-day software that we depend upon.
The distance between the promise and the reality really is huge. On some level I wished they'd just promise less, because it's not like LLMs compleatly useless. I don't find much use in them, but some clearly do. They do them. But since the entire economy has apparently bet the farm on AI, underpromising isn't really an option, while underdelivering is a problem for future Microslop and co.
Is that the original source? I've seen it a few different places, but what really sold me on it was a post Foone reposted on Bluesky.[1]
I've called them Microshaft before, but given their whole pivot into AI (might be a stretch to call it that, but close enough), Microslop certainly feels like a good name.
Interesting thought experiment. In that alternate reality, their shareholders would probably be shouting "why would you give competitors access to this awesome tool?!"
I guess you haven't tried ZZK-5.6 with Maverick Agent? What prompt did you use? If it doesn't work, you can always try a swarm of agents with model hot-reload and re-spin. That will solve all your problems, write all your code and then make you a cup of coffee.
This is a 23-year-old game we're talking about here.
The design work was complete long before anyone working on this project was hired by Ubisoft, and proven in the form of a game that shipped several console generations ago. Ubisoft presumably still has all of the original art assets for reference.
All that had to be done was to study the original game code, port it for modern systems, and then polish up the visuals some. Not a trivial amount of work by any means, but much, much easier than starting from scratch and making a game from nothing.
This should've been a layup for any competent studio given SEVEN YEARS(!!!) to work on the project.
That it wasn't, is undeniable evidence of a AAA game development competency crisis.
The process you describe, of porting the code and polishing the visuals, is for a "remaster". Ubisoft's now-cancelled Sands of Time project was described as a "remake" [1], which generally is a bigger project that keeps the same characters and story beats as the original, but revises the art and gameplay almost as if it were a sequel.
I agree that remakes sound straightforward and it's baffling that Ubisoft couldn't get this one out the door: there are rumors that even before its announcement in 2020, the Sands of Time remake had been started, scrapped, and restarted. But even Nightdive Studios, a perennial remaster maker, struggled for many years with their System Shock remake [2] so it must be harder than it sounds.
I'd love to hear the inside story from folks who've worked on remakes (both released and cancelled) about why they aren't as straightforward as fans expect. Nightdive recently discussed the emotional toll from dealing with angry fans [3] but I haven't seen any interviews discussing the development challenges.
Regardless of “what kind of ‘remake’” they were going for, they had about twice the development cycle of the original game to make it. None of the successful AAA “remakes” of the past few years (most of which are Japanese) had development cycles this long.
The very first Prince of Persia game came out in the early 90s. It's at least 35 years old. It was noteworthy as being one of the first games (if not the first) to have ragdoll physics for the movement of the player.
It was a game of playing through 12 levels (IIRC) and if you died, you started over completely. I played it for 20 hours straight at one point and ended up beating it. I think it takes like 30-45 minutes for a full run through normally? I'm a bit fuzzy on this part. It was an amazing game.
What you're talking about is the early 2000s "reboot" that launched a new franchise under the old name and really wasn't that similar to the original other than a setting of being loosely Persian/Arabic. But it wasn't a platform game in the same sense.
The failed remake in question is for Sands of Time (2003–2004) rather than for the original Prince of Persia (1989)—though, judging by this catastrophe, the remake development team wouldn't have been able to remake the original game in seven years either!
The article states it was a remake of The Sands of Time, so the early 2000s game. As the basis for the project it will count as the original in this case.
