"Inside" as in "LLM search engine for game tips"
We're over a year into the AI hype, and this is the practical application? A game-adjacent search engine? The next AI Winter is going to give us all frostbite.
"Inside" as in "LLM search engine for game tips"
We're over a year into the AI hype, and this is the practical application? A game-adjacent search engine? The next AI Winter is going to give us all frostbite.
This use case is definitely a solution in search of a problem though. Their stated example is something that can be done quickly and locally - for over a decade, mods could list down crafting/processing recipes, item uses and sometimes even highlight inventories with needed items. Everything else about Minecraft's undiscoverability could be solved by writing a comprehensive in-game guide and redoing the achievement system - something sensible instead of calling an LLM.
The simplest explanation is usually the correct one.
That brings up another point: This particular LLM system will require custom integration by the game developer, development that could go to all those features you mentioned. That sounds like a lot of work for a dubious payoff for either the player or developer.
Another way to frame my question above: Is this feature a good demonstration of AI's value (present or future) in the eyes of Microsoft's investors?
On a larger scale, it also seems to demonstrate the different motivations in playing, or the destination versus the journey. For me part of the reason I'd play Factorio is engaging with problem solving assuming the relevant breadcrumbs are there to follow or puzzle pieces to assemble. Following distinct instructions to goals is a different experience that I'd look to other games to provide.