chatGPT part 2: how we change our teaching and student learning

My seniors students have discovered chatGPT. They (like me) were stunned when they saw it. One student asked me why they should study software engineering when there is a tool like this. 

I agree, this tool changes things, but while it can provide snippets of code, I don't think it can write whole systems with various input, processing and output. Much like visual studio code copiliot, I see systems like chatGPT offering support for experienced programmers.

The key thing is evaluation. How will we know the answers we get from chatGPT (and copilot) are correct? I spoke with our school director, always intellectually curious, who suggested teachers ask students to evaluate a reply given by chatGPT. I think this is where we need to be in relation to chatGPT: a smart assistant who needs to be watched carefully. 

That being said, it looks like intelligent agents are going to become better (much better) at being assistants. This seems to be a plausible near-term future. 

Our job is to teach our students well enough so they can evaluate / analyze the output to determine if the solution is correct / best fit for the problem they are trying to solve.