50 Days of LIT Prompts
David Colaursso
Originally Published 2024-01-19 through 2024-03-29. Update Aug. 2024 to replace refernces to deprecated models.
As an experiential educator, I firmly believe experience is an excellent teacher. So, let's build some stuff. In the process we might just gain some perspective on this whole generative AI craze.
I teach law students about technology, and recently I created the LIT Prompts browser extension to help them explore the use of Large Language Models (LLMs) and prompt engineering. You can download the extension for Firefox or for Chrome.
Every weekday for the next 10 weeks, I'll update this page with new prompt patterns and invite folks to play with them inside LIT Prompts. We'll be plugging into the technology that runs tools like ChatGPT, giving you a level of control that most people don't see. You won't have to do any "coding," but we will think like coders. I'm still suspicious of the term prompt engineering, esp. when used as a job title. The idea that an entire job could be built around "talking" to an AI seems a bit too loosey goosey. So, we'll try to introduce some rigor. We'll aim to think about workflows and systems where writing prompts is just part of a larger endeavor. You'll be supported by the fact that LIT Prompts lets you produce reusable interconnecting prompt templates. This will let us build upon what we learn and improve our practice over time. We'll explore the promise and perils of this new technology while you build custom interactions/workflows. Plus, I'm pretty sure we'll have fun along the way.
Questions or comments? I'm on Mastodon @Colarusso@mastodon.social
Week 1
Micro-Lessons (Tuesday-Thrusday): An explination of how to build an artificial neural network
- Live DemoUsing AI to Distill and Question Texts: Summarize and question the contents of a webpage from within the browser
- Live DemoA Rose by Any Other Name: Define a selected word, phrase, idiom, or initialism on a webpage
- I'm Sorry, Dave Can't Do That: Use the text of an email to draft a polite reply declining any request(s)
- You Can't Be Neutral on a Moving Train: Define a selected word, phrase, idiom, or initialism given the context of the webpage on which it is found
- Flip a Poem; Roll an "App": Turn the outcome of a coin flip into a poem and package this as an "app"
Week 2
Micro-Lessons (Monday-Thrusday): Learn how to turn words into numbers with word2vec
- Live DemoI Turned My Scholarly Papers Into Chatbots so People Don't Have To Read Them 🤞: Turn scholarly papers into chatbots so people don't have to read every word
- Give Me the Recipe, Not an SEO-Motivated Short Essay: For an online "recipe page," return just the recipe absent extraneous verbiage
- Add Email Asks to My To-Do List: Find the "asks" in an email; produce a bulleted list of same; add these to a to-do list; save to a file
- Live DemoFollow This One Trick to Write Great Headlines: Produce a collection of possible
clickbaitcompelling headlines based on the text of an article - A Pome for the Moment: Turn the time of day into a poem
Week 3
- Live DemoWhat if Members of Talking Professions Could Log Time in Simulators Like Pilots?: Craft Simple Training Simulations with "AI actors"
- Taming Texts or: How To Turn Unstructured Prose Into Structured Data - Diagram selected sentence, output JSON
- The Dream of Universal Translation: Translate text into English then translate your reply into the original language
- Old-School Automation: Complete a "form letter" without engaging an LLM
- This Day in History: Generate fun "facts" for this day in history
Week 4
- Live DemoIf a 3L Can Automate Law Students, I Can Automate Law Profs—Robo Socrates Lives! A law school instructor shows you how to get AI to brief cases, explain them to a 5yo, play the gunner, and conduct a Socratic dialogue all for FREE!
- All of the Above: Use AI to draft multiple choice questions based on a reading assignment
- True or False? Use AI to draft true or false questions based on a reading assignment
- A Hawk from a Handsaw: Have an AI "guess" the right answers for multiple choice and true or false questions
- 2.B, or Not 2.B? Have an AI answer multiple choice and true or false questions based on a reading assignment
Week 5
- Live DemoNarrative Quanta: LLMs as RPG Building Blocks. Generate the outcome of a scene for a role-playing game based on a single sentence
- TL;DR LLM: Blurb a webpage for social media
- Live DemoFlag Logical Fallacies with a Browser Extension: Have AI read a webpage and flag logical fallacies
- Is This Good? Extract sentiment from social media posts, label each, and structure output as JSON
- Your Own Personal Anger Translator: Translate angry comments into thoughtful replies in context
Week 6
- Live DemoShould the USPTO ask AI for help with Pop Culture References? How LLMs could help screen for bad-faith trademark applications
- Live Demo500 Characters? I'll Make It Fit! Use AI to shorten text
- Magnifying Ideas and Expanding Text with AI: In which I show you how to have AI write something both personal and original, bucking the assumptions that AI writing can never be personal or original
- Live DemoTranslate Legalese: An AI tool for rewriting texts in plain language
- Live DemoThe Elements of Interactive Style Guides: Have AI provide writing feedback based on a house style guide
Week 7
- Live DemoDecodable Books On-Demand: Use AI to create custom books for beginning readers
- Live DemoAnticipating Reader Questions: Have AI/LLMs suggest questions readers might have about your writing
- Live DemoSummon the Demon: Strengthen your arguments with an AI-powered devil's advocate
- Live DemoAlgorithmic BS: Shoot the breeze with your base LLM
- Three Summaries—Three Levels of Difficulty: Produce three summaries at different levels of complexity (i.e., for a child, general audience, and expert)
Week 8
- Live DemoMy Students Have Been Using an Interactive Tool for Reflective Journaling & I Love It! Avoid the "blank page" problem and engage your day with computer-assisted reflection
- Wait, Have I Read This Book Already? In which I extract and clean data from 10 years of emails and online receipts to reconstruct a list of 400+ books because I don't want to start a "new" book only to realize I've read it before
- Magic Copy-and-Paste: Use AI to pluck out and copy only what you want AKA entity extraction
- AI-assisted LIT Reviews: Create a sphreadsheet of citations found in a paper with only a few keystrokes
- Not Quite Dictation: Draft emails based on preset scenarios and an "AI"-mediated Q&A
Week 9
- Live DemoBuild an AI-Augmented Word Processor: Include an interactive version of your style guide; predict reader questions; argue with a devil's advocate; shorten or simplify text... Move beyond spelling and grammar checks—evaluate your logic!
- Auto-Generate Code Comments: In which "AI" suggests comments and explains a function that approximates pi (Ď€)
- What is Retrieval-Augmented Generation (RAG)? In which we approximate RAG using only LLMs and no vector database (for the Roman Empire's Wikipedia page)
- Produce Simulations Based on Random Scenarios: Produce and run AI-drive text simulations that are different every time
- Rubric-Based Machine Evaluation of Assignments: Have students check their work against your rubric before they turn things in (unit testing for written work)
Week 10
- Live DemoThe Library of Unwritten Books: Create novel novellas on-demand, become a reader-author
- Live DemoAI Magic + Algorithmically-Driven News Feeds Minus the Surveillance Capitalism: Using AI to build open-source tools that work for their users, not their builders
- Live DemoThe AI Ouroboros: Can 2024's AI Build 2014's AI?: Using LLMs to suggest features for linear and logistic regression models
- Constructed Criticism: Simulated feedback on how to improve your simulation performance
- The AI-Powered Everything Extension: Curate a Mathemagician's spellbook w/o doing any arithmetic (LLM "functions" that run in your browser)
TL;DR References (A.I. & Large Language Models)
Browse blubs for a selection of works I referenced in the posts above. Functionally, it's a reading list for folks who want to get up to speed on large language models. Updated regularly.