[Music] so one of my favorite puzzles that I would do when I was a little kid is called a nonogram where you're given an empty grid and then some numerical clues that tell you which squares in the grid you have to fill in and I thought we can have the model play a little game where it first generates a puzzle to solve then we ask another instance of the model to try to solve the problem that it generated so yeah I'll ask it generate say a 5x5 nonogram where the final answer is the letter M
we'll see what it comes up with all right we see it just gave us a little puzzle we'll go ahead and copy this open up another window of o1 ask it solve the following puzzle and let's say visualize the answer in some pretty way why not this puzzle doesn't look like it's it's too hard but the way that an monogram works is for each row and each column you're given a list of numbers and the numbers tell you how many squares are filled in and if the squares are consecutive then you'll see like a two
for two consecutive squares if there's a space between them you'll see a one comma one and so you're supposed to just try to figure out which squares do I indeed have to fill in and it looks like the model got this right Illustrated a nice little letter M but I think one of the things that is like nice about examples like this is it it it's similar to like sidoku or a crossword for example where you have to kind of make a guess and then see if that's a right guess or a wrong guess and
then backtrack if you get it wrong and so any type of task where you kind of have to search through a space where you have different pieces pointing with different directions but there are mutual dependencies so you might get a bit of information that tells you that these two pieces contradict each other like a model like om is really good at trying to refine the search space here
