Assignment: The Picture Ladder

Take one picture you photographed and one you drew. Ask AI to do as many of the five things from the lecture as each picture allows: extract, interpret, transform, simulate, extend. Not every picture allows every one. Part of the exercise is finding out where each picture stops.

Do not photograph people or anything with personal information.

Student instructions

1
Step 1: Choose the two pictures.
One photograph of something real with a lot in it. Some options: your closet; your desk or workspace; a dish you like; a moment from a sports game; a piece of equipment with controls on it; the inside of your fridge or pantry; a shelf of books, games, or supplies; a bike, a skateboard, or an instrument. One drawing you make yourself: a game, a machine, a plan for something, a diagram of how something works, a screen for an app.
2
Step 2: Extract.
Ask for an inventory, a transcription, or a list of parts and connections, whichever fits. Note what it got right, missed, and added.
3
Step 3: Interpret.
For the photo, ask AI how to use what is in it, or what in it fits a goal you name, or what is wrong or about to fail. For the drawing, ask which step or part is weakest and why.
4
Step 4: Interpret as a reviewer.
Send it with two different roles: a skeptic, and the person who would have to use or build it. Keep the strongest point from each.
5
Step 5: Transform.
Ask AI to turn what it sees into something of a different kind. For example: turn a picture of your closet into a better way of organizing the clothes in it; turn a series of pictures of a football play into a playbook entry, with a description of the play and an X's-and-O's diagram in ASCII; turn a photo of a dish into a recipe; turn a photo of an object into build instructions from materials you choose. Write one sentence on what AI had to understand about the picture to do it.
6
Step 6: Simulate and extend.
Ask AI to run the drawing, in narration, in frames, or as a game you play move by move, and to stop at any step that fails. Then ask it to extend the drawing: add the next three steps, the next level, or the missing half, and draw the result in ASCII.
7
Step 7: Find the stopping point.
For each picture, name the first thing that did not work well, and say why you think that picture stopped there.

What to turn in

The pictures. Both, as sent.

The five uses. For each one you did: the prompt, the reply, and one or two sentences on what came back that you could not have typed.

The stopping points. Where each picture stopped, and why.

One paragraph. The lecture said AI can use details it sees in a picture that we would never have typed, and that a picture shows AI our physical situation. Which of the five things changed how you think about what you can do with a camera or a pen and AI?

We are looking at whether you can get AI to do more with a picture than describe it, and whether you can tell which things a given picture allows.

Submitting conversations

Follow your teacher's instructions for sharing conversations. Include each picture, the prompt, and the reply.