Teaching Computer Science in the AI Age - discuss!
# thinking-togetherd
Dave Mason
05/07/2025, 10:09 PMTeaching Computer Science in the AI Age - discuss!
Dave Mason
05/07/2025, 10:11 PMOur department is having a retreat to seriously consider what we should be teaching and how. I’m the Chair of the department and trying to figure out how to orient the conversation. In this video, (), Derek makes the case that what we need to do is get students to put the time in to do the slow thinking that would give them the fast-thinking skills to build their capacity. I would love to hear the thoughts from this community.
Dave Mason
05/07/2025, 10:11 PMI have pulled together some resources to inform our discussion in the department. Some of you may be interested: https://docs.google.com/document/d/1T3S3mNfL-xvViZAvnSvyejqToowwsuVjWlvUSUhwRQ0/ If you have any other suggestions, I’d love to hear them
Dave Mason
05/07/2025, 10:34 PMI started this in corecursive.slack.com which might be interesting for those in this space. (I heard about this space from an interview on CoRecursive.)
n
Nilesh Trivedi
05/08/2025, 5:31 AMWhich age group?
For teaching CS to middle-schoolers (12-16 yr olds), I believe that best tool is Berkeley's Snap!. Unlike MIT Scratch, it has all the right CS primitives (first order lists, functions, vectors etc) and therefore encourages computational thinking, but unlike the "industrial" programming languages - like Python, it does not have a high cliff for beginners before they can be productive.
Here is a library that adds AI building blocks to Snap language.
k
Konrad Hinsen
05/08/2025, 8:06 AM"Teaching CS" already looks like a much too narrow framing to me. It's of course what a CS department must think about short-term, but what really needs a big revision is "teaching about computing, including AI, to the next generation of citizens", with a special focus on future intellectuals, entrepreneurs, and political leaders.
d
Dave Mason
05/08/2025, 12:08 PMSorry, university undergrad. I certainly agree there is a general education aspect, but we have 2000 undergrads who are looking for a “Computer Science” education. The question is, what should that look like in the face of LLMs, cursor.io etc.
g
guitarvydas
05/10/2025, 12:21 PMAt one point in the video, a question is asked "if they don't craft those sentences what happens to their brains?". My answer: they get to think about things they can't think about now, because they are currently consumed by thinking about noise issues (at least, those who are inspired to think about specific problems - I don't think that the ratio of inspired-to-think:biological-robots will change). IMO, coding and writing-editing essays/videos is noise, something that stands in the way of thinking about more interesting issues. The currently popular trend is to emphasize linear thinking and specialization, I feel that generalization is more fruitful. ChatGPT is trained on linear thinking, specialization, synchronous programming, which isn't sufficiently broad IMO. ChatGPT can dutifully work on the noise, while humans are freed to think more broadly. IIUC, Kasparov was the first human to be put out of a job by AI. One of the things he does now is to run tournaments involving a new chess discipline, known as "Advanced Chess," aiming at bringing together human skills and machine intelligence.
d
Dave Mason
05/12/2025, 4:47 PMThanks, @guitarvydas for your thoughts and the link!