Hacker News new | past | comments | ask | show | jobs | submit login

This one's my favourite: https://www.youtube.com/watch?v=gUiGEDnf5e4

Crow keeps undoing the guys shoe laces so he can steal the frying pan while the guy's doing his shoe laces up. Classic.




To be fair, that's pretty basic reinforcement learning. The other video, on the other hand, shows adaptation of existing knowledge to solve a problem, along with multi-step planning, which is really incredible!


That may be true, but the more advanced one isn't nearly as funny :P


Only if the crow had significant experience with untying shoes as an opportunity to swipe something interesting... bit of a long shot, no?


You're over thinking it.

On its own shoelaces are interesting. Pulling on the loose end may have started as curiosity.

The person then puts the pan down to tie them. Crow grabs the lace again because, again, curious and playful (birds can be incredibly mischievous), and once again the pan reappears!

Now we're in reinforcement territory.

Now don't get me wrong, the crow learns fast. But it still could be just basic reinforcement of a natural behavior, which is the same technique used to teach various animals "tricks".


It makes me want to tame a crow because it's animals with these kinds of personalities I love. They make life fun.




Consider applying for YC's Summer 2025 batch! Applications are open till May 13

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: