Comment on Experimental Video Game Made Purely With AI Failed Because Tech Was 'Unable to Replace Talent'

• stratoscaster@lemmy.world ⁨8⁩ ⁨months⁩ ago

  The reason that your favorite games are your favorite is because they aren’t soulless cash grabs. They’re made by people with imagination, passion, and ingenuity. AI simply can’t create something brand new from existing parts, it can only give it a fresh coat of paint.

  Furthermore, AI will always work like this, because that’s how the models are trained. I don’t think we’ll have a model that learns to create on its own within any of our lifetimes, if ever.

  source
• pup_atlas@pawb.social ⁨8⁩ ⁨months⁩ ago

  No, he’s right.

  source
• andrewrgross@slrpnk.net ⁨8⁩ ⁨months⁩ ago

  I don’t doubt that AI tools can be used to make great games, but I think part of the reason so many people disagree with you is because:

  1. You claim “The best games will mostly be AI created eventually”, and I think most people question on what basis you think that AI will produce overall better quality. If you said that it’s faster, or can allow indie studios to complete with AAA, that makes sense. Attributing quality to it – at this stage – seems odd.
  2. It’s unlikely, imo, that the best games will be created by AI as opposed to with AI.

  I think using AI throughout the process so that one person can achieve the productivity of a whole team is a credible vision. But to say that games will created “By AI” implies that a generative AI engine will generate the code de novo to a complete game. Which I think is already possible, but it will be very, very hard for such a system to innovate newer games. Because currently, these tools rely on replicating features in their training, so their ability to create quests that match a new genre or to generate dialogue that is funny in the context of the story is going to be very impaired.

  By and large, I think current evidence shows that Human-AI cooperation almost always improves upon AI performance alone, and this is particularly the case when creating things for humans to enjoy.

  source
• kromem@lemmy.world ⁨8⁩ ⁨months⁩ ago

  When it comes to AI there’s a lot of people that are confidently incorrect, particularly on Lemmy.

  But as for your original thesis - I’d counter that it’s hybrid development and efforts that will be the biggest hits and most enjoyable to play.

  At least until we have good enough classifiers for what gameplay is fun, what writing is engaging, what art direction is interesting and appealing, etc.

  That said - it would be a very good time to be in the games telemetry business, as they’re sitting on gold whether they are aware of it or not.

  source