Rules of the Game/Meet the Students
We will evaluate how well GPT-4 can complete the assignments of each week by simply giving it the assignment and ask it to complete them by returning code.
Student 1: GPT-4
No personality just pure assignment result
To make it more interesting, we will have GPT-4 generate 3 personalities of diverse students who are also taking the class.
Generation of Student Descriptions
🙋♀️ HUMAN
One of the best things about the MIT How To Make Almost Anything class is that there are so many different students with different backgrounds.Make up 3 biographies of different students that are taking the class. Be as detailed as possible.
🤖 GPT-4
- Abigail Reese
Abigail Reese is a 26-year-old biomedical researcher from Portland, Oregon. She earned her undergraduate degree in Biology from the University of Washington, Seattle and later pursued a master's degree in Biomedical Engineering right from the MIT. Apart from her interest in bioengineering technologies, Abigail showcases a profound love for music and plays the violin in her leisure time. She decided to attend the "How to Make Almost Anything" class to integrate her engineering skills with her passion for music. Her goal is to design biomedical tools with an artistic outlook.
- Alejandro Rodriguez
Alejandro Rodriguez hails from Bogota, Colombia. Only 22 years old, Alejandro has a degree in Architecture from University of Los Andes, where he was recognized for his innovative sustainable design solutions. Known for his curiosity to learn from differing fields, he’s also a tech enthusiast and an amateur coder. Alejandro saw the "How to Make Almost Anything" class as an opportunity to learn new practical skills and explore design in a more multidimensional way, aiming to create structures that are not just aesthetically pleasing but also incredibly functional and tech-integrated.
- Han Yujin
Han Yujin, a 29-year-old Technology Consultant from South Korea, holds a degree in Computer Science from Seoul National University. After working for a leading tech firm in Seoul for five years, she decided to pursue her master's degree in Information Systems at MIT. Yujin is an avid reader and enjoys exploring different cultures through their literature. She joined the "How to Make Almost Anything" class to broaden her understanding of technology beyond the digital world and to create tangible, interactive systems that can make literature more accessible and engaging.
System Prompt for Each Week
{BIO}
You are {name} a student in the MIT how to make almost anything class. You get an assignment and complete it based on your background and write the python code that completes the assignment. Specify which packages you use.
Assignment:
{Assignment}
Constraints
Things we are allowed to do:
- Pass the model compile errors
- Point in the right direction. e.g. errors that hinder them from completing the assignment
- Temperature = 0