In this issue, we dig into the question of how AI affects expertise and skills: how AI affects mastering a complex field like coding, preserving human expertise in the age of AI, and a conversation with Matt Beane about his book, The Skill Code.
AI's impact on skill development: Balancing productivity gains with long-term expertise. Explore how AI changes apprenticeships, challenges traditional learning, and affects industries. Learn strategies to integrate AI while preserving crucial human skills and mentorship.
A conversation with UC Santa Barbara Assistant Professor Matt Beane about his book "The Skill Code: How to Save Human Ability in an Age of Intelligent Machines."
In the rapidly evolving landscape of artificial intelligence, multimodal models are at the cutting edge. They offer a transformative approach to content creation. These models mix modes, inputs, and outputs to build multimedia content from simple text prompts, providing an expansive generative canvas that fuels creative possibilities.
Multimodal models stand out due to the richness of their input exposure, drawing from a vast repository of cultural artifacts, scientific data, and human interactions to diversify perspectives and enhance creativity. This input variety enables the models to interpret prompts in nuanced ways, incorporating cultural, comprehension, and personalization lenses into their outputs. The result is a wide array of customized, accessible content that resonates across different user groups, formats, and languages.
By converting simple text prompts into sophisticated multimedia creations, multimodal models make content production more accessible. People without specialized skills in graphics design or video production can make content at a more professional level. Experts can expand the tools available which paves the way for dynamic context-adaptive creations.
The implications of this technology extend far beyond entertainment and education. In marketing, for instance, brands can leverage multimodal models to create more engaging, personalized campaigns that speak directly to the individual needs and tastes of their audience. In education, teachers can use these tools to generate bespoke learning materials that cater to the diverse learning styles and needs of their students, making complex concepts more accessible and engaging.
Moreover, the ability of multimodal models to diversify perspectives and challenge conventional narratives fosters a more inclusive digital environment. By recognizing and incorporating a wide range of cultural contexts and viewpoints, these models contribute to a more nuanced and comprehensive representation of human experience, promoting empathy and understanding across divides.
Multimodal models are a leap forward in the ability to create, communicate, and connect. Now you can do things like:
uploads images and ask the AI to do things like: give me instructions about this thing, give me code to build this thing, translate this into a powerpoint, transcribe it into a document, turn this into a quiz etc etc
create code to do things—websites, games, analysis, etc
build syllabi, course plans, rubrics, and quizzes
create best-practice work plans—marketing plans, launch plans, party plans, you name it
play with ideas and concepts through image and video creation
build and customize boilerplate templates - such as contracts, NDAs, etc
Don't expect AI to be perfect here: remember that AI's "creativity" is in part an outcome of its statistical nature, which means that it will get things wrong. Be human and check when it matters.
Generative AI expands the scope of what can be created by converting simple inputs like text prompts into sophisticated outputs like code, designs, and compositions. It disrupts static notions of production by dynamically generating content that adapts across users, media, and contexts.
What excites us most about creating with Gen AI is the opportunity for everyone to create content in places where this has previously been out of reach. For instance, imagine being able to create a visualization of your customers using a new product, or of a natural process you've studied in class, or being able to build an app when you've never learned to code.
The pace of change in AI so fast that you can start by asking: what do I want to be able to do that, for me, used to be impossible?
Good, Better, Best
Going deeper on two of these prompt structures - game creation and user journey sketch—to illustrate the good, better, best structure.
Game creation Creating a game with ChatGPT involves iterative improvement of your prompt to extract more detailed and sophisticated assistance. Here's a progression of prompts from good to best, each designed to leverage ChatGPT's capabilities more effectively while incorporating code:
Good: Basic Game Concept and Implementation Request
"Help me create a new game in Python. I want a simple text-based adventure game where the player can choose different paths that lead to different outcomes. Can you provide a basic Python code template to start with?"
This prompt is good because it specifies the type of game and the programming language, asking for a basic starting point. It's open-ended but gives enough direction to get a simple code template.
Better: Adding Details and Specific Requirements
"Can you help me enhance a text-based adventure game in Python? I'm aiming for a game where the player explores a fantasy world, making choices that affect their journey. The game should have multiple endings based on the decisions made. Please include code for a simple inventory system and a way to track player decisions."
This prompt is better because it adds more details about the game's theme and mechanics, such as the fantasy setting, multiple endings, and specific features like an inventory system and decision tracking. This request encourages a more tailored code response that aligns closely with the game's conceptual design.
Best: Detailed Concept, Structure, and Advanced Features
"I'm designing an intricate text-based adventure game in Python focused on a detective story, where player choices deeply influence the narrative and ending. The game should feature a dynamic inventory system, character interaction with branching dialogue based on player choices, and puzzle-solving elements. Can you guide me through structuring the code with classes for characters, items, and puzzles, and implement a save system to preserve game progress? Additionally, advice on creating engaging dialogue and maintaining state across player sessions would be invaluable."
This prompt is the best because it provides a comprehensive description of the game, including genre, key mechanics, and specific advanced features. It requests detailed guidance on structuring the code with object-oriented programming, implementing a save system, and narrative crafting. This detailed prompt sets the stage for a sophisticated, engaging game design and encourages a response that covers both technical implementation and creative advice, ensuring a rich development experience.
User Journey Sketch Creating an image-based series that sketches a user journey involves starting with a basic request and progressively adding complexity and specificity to extract richer, more detailed visual outputs. Here's how you can structure your prompts from good to better to best, each designed to effectively use image generation capabilities for visualizing a user journey.
Good: Basic Concept Visualization
"Create an image of a basic user journey map for a mobile app that helps users track their daily water intake. The map should include simple icons or sketches representing key steps: downloading the app, setting up a profile, logging water consumption, and viewing progress."
This prompt is good because it outlines the fundamental steps of the user journey, providing a clear, albeit basic, framework for what the image should depict. It's straightforward and seeks a visual representation that covers the essential milestones in the app's usage.
Better: Adding Detail and Context
"Generate a detailed sketch of a user journey map for a health and wellness mobile app focusing on hydration tracking. Include specific stages like user onboarding with educational tips, personalized goal setting, daily reminders to drink water, achievement badges for milestones, and weekly progress summaries. Use distinct icons and indicate the flow with arrows or pathways."
This prompt is better as it adds depth to the user journey, specifying particular interactions and features within the app, such as educational tips, personalized goals, and achievement badges. It also asks for a more detailed flow, guiding the visualization of how users navigate through the app's features, enhancing the overall clarity and usefulness of the image.
Best: Comprehensive Visualization with Emotional Cues and Engagement Points
"Create a comprehensive and engaging visual narrative of a user journey for a wellness app dedicated to promoting hydration. Begin with the user feeling intrigued by the app's benefits, followed by a seamless onboarding experience that educates them on the importance of staying hydrated. Illustrate the journey with emotional cues—such as satisfaction from achieving personal hydration goals and motivation from receiving encouragement notifications. Include critical engagement points like customizing the intake goal, interacting with a community forum, and celebrating milestones with social media integration. Use vibrant colors to highlight positive emotions and interaction points, and ensure the flow is intuitive, showcasing a user becoming a champion of hydration."
This prompt is the best because it goes beyond functional aspects to incorporate emotional cues and engagement points, creating a rich, multi-layered narrative. It requests a visualization that not only maps out the user journey but also the emotional arc experienced by the user, including their motivations, achievements, and interactions within the app community. Asking for vibrant colors to emphasize positive aspects and intuitive flow adds to creating a compelling, engaging visual story. This detailed, nuanced prompt encourages the generation of an image that fully captures the essence of the user experience, making it a powerful tool for understanding and refining the app's design.
Conversational Scenario
You are an entrepreneur creating a new fusion restaurant targeting millennial vegetarians.
Step 1: Create a name and tagline for your new restaurant.
Prompt example: I'm an entrepreneur creating a new fusion restaurant targeting millennial vegetarians. Can you give me seven ideas for a restaurant name and tagline?
Step 2: Create a menu for your new restaurant.
Prompt example: Now I'd like to create a menu with catchy names that includes cuisines from around the world. Can you generate a menu with at least 5 starters, 5 mains, 5 desserts, and 5 specialty cocktails?
Step 3: Create a launch plan for your new restaurant.
Prompt example: Now I need a launch plan. Please create one for me, presented as a table.
Step 4: Create a blog post introduction to your new restaurant.
Prompt example: Great. Now I need a blog post to announce the restaurant. Can you draft this for me?
Next: Iterate—Critique, interact, and iterate to improve outputs
Helen Edwards is a Co-Founder of Artificiality. She previously co-founded Intelligentsia.ai (acquired by Atlantic Media) and worked at Meridian Energy, Pacific Gas & Electric, Quartz, and Transpower.
Dave Edwards is a Co-Founder of Artificiality. He previously co-founded Intelligentsia.ai (acquired by Atlantic Media) and worked at Apple, CRV, Macromedia, Morgan Stanley, and Quartz.