The Learning Curve #2: Stop assigning one big final paper
A small upgrade starting this issue: every Learning Curve now has the same two or three parts, one teaching move, one AI story translated, and now and then a tool worth a look. Same promise as ever, no hype, no hard sell.
The Teaching Move: Patchwork Assignment
Since I wrote about final papers on LinkedIn a couple of weeks ago, the replies have kept circling one worry: the single deliverable that lands at midnight in week 15. It is the easiest thing in the world to outsource, not the learning, the packaging, because nothing about it is visible until it is done.
The fix is not surveillance. It is making the work show its seams.
Break the paper into a patchwork of smaller artifacts that each get handed in along the way: the question they chose and why, an annotated source, a messy first attempt, a paragraph on what they changed after feedback, then the final piece. Each step is small enough that doing the real work is easier than faking a story about it. And you get to watch the thinking move across the semester instead of guessing at it from one polished block of text.
You are not adding busywork. You are grading the process you used to take on faith.
To set it up this week: take one final paper and pull three checkpoints out of it. Put a date and a tiny point value on each. That is the whole redesign; you can refine the prompts later.
The accessibility note: A semester of small checkpoints serves students with extended-time accommodations far better than one high-stakes midnight deadline, because the flexibility is built in instead of bolted on. Let a checkpoint or two come in as audio or annotated notes and you have also widened the ways a student can show thinking, which is the point of the assignment in the first place.
Claude for Teachers: Translated
On Tuesday, Anthropic launched Claude for Teachers: a free year of its premium AI for verified K-12 educators in the US, with lesson planning tied to state standards and student data terms written to comply with FERPA.
If you teach in higher ed, you cannot sign up, and that is fine, because the feature list is the useful part. Scheduled reviews of exit tickets. A library of teaching skills. Differentiated materials at several levels. Almost none of it is new model capability. It is context, memory, and scheduling wrapped around the twenty minutes of a teacher's week that repeat. The gap between you and that product is not access. It is sitting down and naming which twenty minutes of your own week repeat, then pointing the tools you already have at those.
One caution before you try. The K-12 version ships with signed, FERPA-aligned data terms. Your personal account does not. The workflow patterns travel; student data should not. Build the habit on your own materials, not your rosters.
The announcement is worth ten minutes if only for the feature list: Claude for Teachers.
P.S. If you read something here you disagree with, reply and tell me. I would rather be corrected by someone in the role than be polished and wrong. And if a colleague owns this work, forward it their way.
— Brooks Winchell. 21 years inside higher education, four of those at Quinsigamond Community College.