This week were were thinking about how the combination of good archaeology and well provenanced texts can tell richer stories. Ea-naṣir's house and letters provide an excellent case-study, a 4,000 year old micro-history of one merchants life and work in the ancient city of Ur.

I hadn't seen this mentioned yet, but I garden for all of those reasons and also to reduce my reliance on fossil fuel shipped fruits and veggies. It's so nice to eat a fresh cherry tomato from my backyard rather than something that came from halfway across the planet.

I treat as an opportunity to practice accountability, and like you I've never seen in abused. Students generally ask for a few days extra and end up handing in better work often before I've even started grading the things that came in on time.

This has been my policy for a while, I tell the students that the due dates on the syllabus are determined by me, but each of them has their own course/work schedules and that they should identify the crunch times in their semester and propose realistic times that they can hand in their best work.

Wow, not good!

The absolute encouragement and boosterism to do whatever it is you ask is just crazy... seems so dangerous. Not only does it tell you to go for it, it rationalizes it, and predicts future payoff. I mean why wouldn't you?! It's stops short of calling you a coward...

I've rented a dirt cheap VPS from DigitalOcean for ages and have run various CMS type things on it... wordpress, hugo, now I've settled on @11ty.dev, having the server has been handy for running other experiments/websites when the mood strikes.

Oh I'm sure the actual categorization is pretty far off... but I give them credit for putting the data out there... I'd be fascinated what this looks like for ChatGPT™ since that seems to be the one students are most familiar with... but I doubt we'll ever see that.

As further evidence of future crisis check out this graph (also from the whitepaper), the discipline where direct output creation was most common: "Education". What will that do to our education system in 3-4 years when these graduates are teachers?!

I give the authors a lot of credit for putting this all out there, they dance a little bit around the issue of "cheating" but I think this whitepaper gives a lot of important data that shows how students are using AI right now is contrary to learning (full stop).

There's a lot in there, but this figure is the standout. Students use AI to create work, not to understand tricky problems or figure out how to apply concepts. They're just using AI as a shortcut not a tool. Quoting the authors of the white paper: "An inverted pyramid, after all, can topple over."

They divide the questions in a four part (2-axis) categorization... direct vs collaborative and problem solving vs output creation, this useful because it gets at the issue of whether AI is a tool to understand problems or just solve them and move on... you can probably see where this is going.

Caveat this is all anonymized data which they tried to assign to various disciplines in higher-ed, but it probably tracks pretty well... they start off looking at the proportion of subjects using AI compared to percentage of undergrads enrolled in those degrees: not surprisingly CS stands out