We ran it a couple of weeks ago with 70 students as part of an onboarding session. Feedback, both on usefulness and realism, was positive, so we'll be updating and running again in the winter with the next intake. Can give more practical details or dm a link to have a go if useful...

Very happy to share resources or more detail if you want to drop me an email...

We run with classes of 30-50; on 6-7 tables. Each table has a set of tasks and their own combination box to open, and at the end each has their unique clue to crack the final whole-class puzzle - finding the one right index card out of a thousand possibles. Lots of sorting and group working.

That's to say, nothing clever from my side; you could do the same on your PC or Mac by following these instructions: medium.com/@researchgra.... Go on. Have a go.

And the science bit: I'm running Meta's Llama3.1 8 billion parameter model using Ollama with an open-webui interface running through Docker. I only know what some of these words mean.

Also re the environmental impact of GenAI, this is using significantly less processing power than, say, playing Microsoft Flight Simulator. So clearly still issues about local impact of data centres and scaling of GenAI use, but everyday ChatGPT-style interactions don't have to be energy intensive.