Should we have a [AI dev] tag, and on the benefits of AI in modern computer wargaming coding (invitation to debate - but let's keep it civil)
Yeah, well, alright, AI is baaaad. I still kind of look back fondly on my past innocence, back when our original demo for TFA was held up by Steam because I'd mentioned using AI for spellchecking and felt obligated to explain why and how. Nowadays that would hardly even qualify as "AI use." But beyond a few abuses, is it really such a bad thing?
Nearly all the recent game concepts you've seen on the board lately have been made, in one way or another, using AI as an all-in-one assistant - finally giving wings to guys out there who had ideas but didn't personally have the skills to code them. Some of these were, to be honest, very promising. Claude is doing a great deal on the prototyping front, helping tremendously with proofs of concept. You can tell where it's coming from right away, with those telltale golden default fonts and frames that make it recognizable from the get-go.
But should our designers be ashamed of using that approach? They're not taking a coder's job away - a coder they wouldn't have been able to afford in the first place. Refusing to let them post, or criticizing their approach, only makes them more reluctant to post at all. And when they do post, they dread the inevitable question: which engine? Which language?
I'm following this wargaming resurgence with a lot of interest. A LLM isn't going to imagine something the designer hasn't already thought of. Sure, a LLM can make a game that looks at least as good as Armoured Commander nowadays... But no AI can truly "make" Armoured Commander without being told what it's about and how it's supposed to work. Should we really frown on the next game of that kind being just around the corner, and decide to nail the guy to the wall because we're anti-Claude?
Nobody who's actually trying to make a buck is interested in our niche. The people using AI to help them make games are mostly people like us.
I'd say: give them a chance, let them have it. We should be focusing on the mechanics, and pushing forward the systems that show the most promise. Then maybe, down the line, a coder joins, a publisher takes an interest: you never know. But at least let's be nice to them. Whether you like it or not, they're the designers of tomorrow, and even though we need to set standards, we should judge the substance rather than the polish. My 2 cents.