Your comment sparked an idea of a conversation with an agent where at any time you could ask, “what’s this word?” and it would respond with an image explaining it. I’m surprised there don't seem to be voice apps that integrate any visual content.

The English intro feels non-native too, and does have a typo (“You’ll be truly amazed of how polished […]”), which makes me think it wasn’t AI.

The downside of that is that cards often have to hook into custom functions that interface with the engine (unless they're sufficiently boring). You can still go with ScriptableObjects if you're willing to e.g. write those functions in Lua and have an interpreter layer, but for all but the largest projects, making each card its own class works well.

Sweeping generalizations about how LLMs will always (someday) be able to do arbitrary X, Y, and Z don't really capture me either

Please don't submit AI-written posts