CSE AI trivia, history, and tips
Black and white photograph from 1952 showing Donald M. Weidenbach with a large Engineering Research Associates (ERA) magnetic drum
The magnetic drum was an early storage system that helped computers store instructions and data electronically. 1952. "Engineering Research Associates - Systems Components." University of Minnesota Libraries, Charles Babbage Institute., Accessed May 19, 2026. https://umedia.lib.umn.edu/item/p16022coll91:205

From room-sized computers to AI 

Long before today’s AI breakthroughs, College of Science and Engineering (CSE) alumni were helping build the future of computing. In 1950, alumnus Arnold A. Cohen helped develop the ERA 1101, one of the earliest stored-program electronic computers delivered to a customer site. Modern AI chips can weigh just a few grams, but early computer hardware filled rooms. 

Test your knowledge with the question below. 

AI trivia look back question

How much did the ERA 1101 weigh?
A) 1,500 pounds 
B) 4,200 pounds 
C) 17,000 pounds 
D) 45,000 pounds

Expand all

Click here to reveal the answer

Answer C) 17,000 pounds

ERA 1101 was no lightweight 

Cohen earned a Bachelor of Electrical Engineering in 1935, a Master's in Physics in 1938, and a Ph.D. in Physics in 1947 from the University of Minnesota and later became CSE assistant dean. The ERA 1101 used about 2,700 vacuum tubes and magnetic drum memory. It could store instructions and data in memory instead of being physically rewired for every new task — a breakthrough that helped shape modern computing. 

Today’s AI systems are dramatically faster and smaller, but they still rely on computers that can store, update, and process enormous amounts of information quickly. Cohen also helped preserve the early history of computing in Minnesota as an early supporter of the Charles Babbage Institute for Computing, Information and Culture at the University of Minnesota. 

Today, the University builds on this legacy with its initiative to lead the future of AI. 

Vipin Kumar standing in Keller Hall

Vipin Kumar

Data Science Initiative Director & William Norris Land Grant Chair in Large-Scale Computing

"If you're not a leader in this field, you will lose relevance pretty soon. It’s absolutely critical for us to be at the top."

Ready for the next round?

See how your prompting strategy holds up against an AI model's logic 

If an AI model gives you a vague or "lazy" answer, there’s a simple prompt trick that can improve its reasoning. Research shows that one short phrase can encourage Chain of Thought prompting (CoT), helping AI models reason through problems more carefully and reduce logic errors.

AI pro tip question

Which phrase encourages better AI reasoning?

A) Run a recursive diagnostic
B) Let's think step by step
C) Double-check your response
D) Explain your answer carefully

Expand all

Click here to reveal the answer

B) Let's think step by step

This phrase encourages AI systems to break their reasoning into smaller steps, which can improve logic and reduce mistakes. 

Standard prompts often force AI to predict the final answer too quickly. 

“‘Let’s take this step by step,’ also works with people, including students,” said Joe Konstan, Associate Dean for Research, College of Science and Engineering, and Distinguished McKnight Professor and Distinguished University Teaching Professor in the Department of Computer Science and Engineering. 

“There’s a reason we ask students to show their work and take things ‘step by step.’”

Help us shape the future of AI

What is your favorite AI prompt hack — or your biggest question, concern, or curiosity about AI today? Share your thoughts with the CSE community here. We’ll share selected responses in the next edition of the CSE Alumni and Friends e-newsletter.

Want to expand your AI skills?

Register online for the summer short course, “Beyond the Prompt: Building Custom Tools with AI” before May 31, 2026.