Lessons from teching CS50 in the age of AI | by Jess Li | Harvard in Tech | Medium

David J. Malan, Harvard’s Gordon McKay Professor of Practice of Computer Science and the instructor behind CS50, recently hosted an event with HAA on how modern AI works. The session was part of the HAA Speaker Bureau Spotlight Program, co-sponsored by 65 Harvard alumni groups and attended by over a thousand people. Here are the key takeaways. Many people have interacted with machine learning for years, even if they did not label it “AI” at the time. The shift we’re experiencing now is that newer systems are bigger, stronger, and easier to use, so the same underlying ideas now show up directly in consumer tools rather than only in background systems like filtering and recommendations. Professor Malan started with examples where you can write explicit logic (hard coded versus not). In the classic game Breakout, you can encode a direct rule: if the ball is left of the paddle, move left; if it’s right, move right. This works because the environment is simple enough that hand-written decis

saved by

• Linda