• Together with the group you worked with, finish the negotiating game sketch you started in class, turning it into an polished interactive experience. (I thought the way Jiapei embedded UI elements and graphics was quite successful. You could also consider using sound or simple animation.)

<aside> <img src="/icons/chat_lightgray.svg" alt="/icons/chat_lightgray.svg" width="40px" /> Sketch with Jiapei!!! here

I really want a dog, but always fail to convince my mom and dad to have one. They think it's enough to raise two kids, my brother and I. Puppies don't talk, and my parents sometimes get the wrong idea even with me, let alone a puppy that doesn't talk.🥺

Jiapei and I aims to develop this simple but warm game. I hope that every child with such a good wish will get a puppy who will accompany them and spend countless happy hours with them! Forever!

Jiapei really taught me a lot in the coding process. Like how to set a specific button in the following format:

And also how to utilize the text we input in the box and the message the OpenAI sent to us. When looking at Jiapei code the most challenging part, I also quietly learned how to make the entire framework load more smoothly (don't put everything inside the 'draw' function, which I did a lot in the past)!

Because Jiapei has already done a pretty good job on embedding UI elements and graphics, we decided to continue to let the users have a better using experience. The painting shows what we have done with it to reach our goal.

</aside>

This is my “puppy”, who has been with me since I was born. I love my puppy, forever.

• Document some of the most funny, deep, meaningful or surprising dialog snippets that you encountered playing this game.

<aside> <img src="/icons/chat_lightgray.svg" alt="/icons/chat_lightgray.svg" width="40px" /> After complete our coding, Jiapei and I both tried to convince the OpenAI.

</aside>

Jiapei’s process

Lisa’s (which seems pretty easy)

Wow It was really surprised that even with the same code, same requirement, the difficulty level can not always stay the same. We are wondering for the reason.

To make this game easier, we first tried this to imply the difficulty level of the game.

But when trying, we found it is very difficult to persuade the mom in the game.

Then I asked ChatGPT about how it comprehend this sentence:

That could be one reason why the game is still very hard when we apply the instructions. In that case, we changed it to “You are very likely to be convinced.”

This instruction is much more simple and straightforward, causing little misunderstanding.

One funny thing is that when talking to OpenAI, all the things can be made up. Since OpenAI know nothing about the players, we can say that we are pretty rich with stable jobs to be able to raise a dog. Or I just own lottery so don’t worry about the money (that’s my dream ;-D).

But this also somehow generate some problems. If all the information can be made up, then what’s the point to create such games to help players under different situations? I’m sure not everyone have enough money or eligible enough to take care of their pets. Buying a pet is such an important decision that you need to be responsible for their whole life. Any abandonment or abuse of pets is not acceptable!

• Watch the trailer for Papers, Please, an award-winning indie puzzle game by Lucas Pope (no LLMs or AI involved!) some people called a “dystopian political sim”. (We didn’t get to watch this in class - wanted to share such scenario as example for where LLMs could be considered in more darker, critical contexts.)