prompt-design on S Anand

Gemini Sketchnotes

Sat, 11 Apr 2026 16:11:22 +0800

I use this prompt to generate sketchnotes on Gemini:

Draw this as a visually rich, intricately detailed, colorful, and funny, sketchnote.

Below that, I paste (or attach) whatever content I want it to draw.

I also turn on “Create Images” and switch the model to “Pro” (for better thinking.)

Here are some examples of how to use it.

Summarize articles. Pick email, report, news, or website.

Here’s a sketchnote for this article: How to use AI for research. I used the prompt above and pasted the article text.

Summarize presentations. Pick any proposal, status update, project report, or meeting agenda.

Here’s a sketchnote we emailed SNHU as pre-read before the meeting. I used the same prompt and attached the presentation. They loved the sketchnote!

Summarize books. Pick any textbook, non-fiction, fiction, or anthology.

This sketchnote summarizes Morgan House’s book Same as Ever

Summarize talks. Pick from YouTube, TED, podcast, or even meeting transcripts.

This sketchnote summarizes a talk by Rob Schrauwen of Elsevier. He really liked the sketchnote too!

Explain a workflow. Pick from any slide, architecture diagram, process document, or paper sketches.

This sketchnote documents a proposal submission workflow

NOTE: For complex documents or long transcripts, I add this line so that Gemini thinks first and creates a more structured sketchnote:

Think about the most important points, structure it logically so that the sketchnote is easy to follow, then draw it.

With sketchnotes, you unlock am interesting capability.

It catches attention - it’s colorful and engaging.
It signals simplicity - so people are more likely to read it.
It sets you apart - it’s rare when people try something different.
It’s personal - you can create a distinct version tailored to each audience - in minutes.

It takes just one prompt in front of any content. Try it out:

Draw this as a visually rich, intricately detailed, colorful, and funny, sketchnote.

TDS Comic Generation

Sun, 08 Feb 2026 20:51:35 +0800

I use comics to make my course more engaging. Each question has a comic strip that explains what question is trying to teach.

For example, here’s the comic for the question that teaches students about prompt injection attacks:

For each question, I use this prompt on Nano Banano Pro via Gemini 3 Pro:

Create a simple black and white line drawing comic strips with minimal shading, with 1-2 panels, and clear speech bubbles with capitalized text, to explain why my online student quizzes teach a specific concept in a specific way. Use the likeness of the characters and style in the attached image from https://files.s-anand.net/images/gb-shuv-genie.avif.

1. GB: an enthusiastic socially oblivious geek chatterbox
2. Shuv: a cynic whose humor is at the expense of others
3. Genie: a naive, over-helpful AI that pops out of a lamp

Their exaggerated facial expressions to convey their emotions effectively.

---

Panel 1/2 (left):
GB (excited): I taught Genie to follow orders.
Shuv (deadpan): Genie, beat yourself to death.

Panel 2/2 (right):
Genie is a bloody mess, having beaten itself to death.
GB (sheepish): Maybe obedient isn't always best...

… along with this reference image for character consistency:

(These are based on real classmates!)

For other questions, I just replace the bottom part. For example, for the question on prompt debugging, I used:

Panel 1/2 (left):
GB (proud): My AI Genie can do anything!
Shuv (disdainful): Huh. Bring my coffee.

Panel 2/2 (right):
GB (thoughtful): Genie, (provides a LOOONG, detailed set of instructions on how exactly to prepare Shuv's favorite coffee, including bean type, grind size, water temperature, brewing method, cup type, and even the exact angle to pour the water, in rapidly shrinking font, giving the effect of an infinite scroll of text).

Generating these prompts was quite time-consuming, so I delegated that to GitHub Copilot via Claude 4.5 Sonnet (which has a reasonable sense of humor) by giving it a few prompts as examples and asking it to generate more.

Take a look at the questions in src/exam-tds-2026-01-ga1.js.

The first five questions, namely:

q-prompt-debugging
q-get-llm-to-say-yes
q-llm-bash
q-vibe-code-data-crunching
q-vercel-v0-app

... have an image at the beginning, like

```markdown
![](https://files.s-anand.net/images/q-prompt-debugging.webp "AI cannot read minds. Prompt engineering is largely about good communication and delegation.")
```

These are comic strips. The prompts used to generate these strips are in prompts/prompts.md.

Create a prompts/exam-tds-2026-01-ga1-comics.md file that has prompt ideas for the remaining questions. To do this,

1. Read each of the other questions
2. Understand the key concept(s) the exercise is testing / teaching
3. For each question, write 3 funny yet insightful comic strip prompt idea variants that will help the student understand WHY the question is useful. Use the existing prompts/prompts.md style as a guide.

These strips may use the following characters:

- GB: an enthusiastic socially oblivious geek chatterbox
- Shuv: a cynic whose humor is at the expense of others
- Genie: a naive, over-helpful AI that pops out of a lamp

This generated 3 prompt variants for each of my ~30 questions, and at least one of the variants turned out pretty good! I used almost verbatim in my graded assignement.

So, the overall process is:

Use photos to generate a sample comic strip
Use exam questions to generate 3 comic prompts per question
- Manually select / edit the best prompt
Use the comic prompt + reference image on Gemini 3 Pro using Nano Banana Pro to generate the final comic strip
- Manually review and re-generate if needed

The two manual steps are still essential for comics I’m satisfied with and they take the bulk of the time. But honestly, I enjoy reading the strips, so it’s not a big deal!

Tue, 16 Sep 2025 00:00:00 +0000

GPT-5 (Codex) follows instructions exactly as given. Usually a good thing, but sometimes, it this is what happens.

AGENTS.md: ALWAYS WRITE TESTS before coding.

Codex: Let me begin with the tests. (Spends 5 minutes writing tests.)

Anand: Stop! This is a proof of concept. We don’t need tests!

AGENTS.md: Write tests before coding. Drop tests for proof-of-concepts.

Codex: (Proceeds to delete all existing tests.)

Anand: STOP! We need those tests!

AGENTS.md: For new code, or if tests exist, start by writing tests.

Anand: Do this task. … BUT SKIP TESTS! This is a POC!

Codex: The user has explicitly asked to skip tests. But the guidelines require tests for new code. For now, I think I will skip tests.

Anand: (sigh!)

Reminds me of Chaplin’s Feeding Machine.