I fed one lecture into an AI flashcard tool and got 70 cards in seconds, which felt impressive until I noticed several were basically the same question.
Does anyone get useful results from this, or is making the cards yourself part of the learning?
I fed one lecture into an AI flashcard tool and got 70 cards in seconds, which felt impressive until I noticed several were basically the same question.
Does anyone get useful results from this, or is making the cards yourself part of the learning?
I use generated cards as a draft. Then I delete most of them, split complicated questions, and check every answer against the lecture or textbook.
The time saving is real for definitions. It is much smaller for concepts that need explanation or examples.
i made like 300 automatically and then avoided the deck because 300 cards is horrifying. now i ask for 15 key questions, edit those, study them, then add more if i need them. way better.
crammingfailed wrote:
I ask for 15 key questions, edit those, study them, then add more if I need them.
Same here. A small accurate deck is better than a huge one full of vague prompts. I also add the source page or slide number to each answer so I can check it quickly.
Sounds like the useful part is generating a manageable first draft, not accepting a finished deck. I’ll cap each lecture at 15 to 20 cards and verify them while editing.