chatgpt It has exploded in popularity, and people are using it to write articles and essays, produce marketing copy and computer code, or simply as a teaching or research tool.
However, most people do not understand how it works or what it can do, so they are either not happy with its results or are not using it in a way that brings out the best of its capabilities.
I am a human factors engineer. The main principle of my field is to never blame the user.
Unfortunately, ChatGPT’s search-box interface elicits the wrong mental model and leads users to believe that entering a simple query should yield comprehensive results, but ChatGPT doesn’t work that way.
Unlike a search engine, with static and archived results, ChatGPT never copies, retrieves or searches information from anywhere.
Rather, it generates each word from scratch. You send it a signal, and based on its machine-learning training on massive amounts of text, it generates an original answer.
Most importantly, each chat preserves the context throughout the conversation, meaning that questions asked and answers given earlier in the conversation will inform responses that arise later.
Therefore, the answers are flexible, and require the user to participate in an iterative process to shape them into something useful.
Your mental model of a machine – how you think about it – is critical to using it effectively.
To understand how to shape up a productive session with ChatGPT, think of it as a glider that takes you on a journey through knowledge and possibilities.
dimensions of knowledge
You can start by thinking of a specific dimension or niche in a topic that interests you. For example, if the topic was chocolate, you could ask her to write a sad love story about Hershey’s Kisses.
The Glider has been trained on essentially everything Kiss has been written about, and so it “knows” how to glide through all kinds of story locations – so it will confidently take you on a flight through Hershey’s Kiss locations to create the desired story.
Instead you can ask him to list five ways chocolate is healthier and respond in Dr. Seuss style.
Your requests will launch the glider through different knowledge spaces – chocolate and health – towards a different destination – a story in a distinctive style.
To unlock ChatGPT’s full potential, you can learn to fly the glider through “transversal” realms – realms that cross multiple domains of knowledge.
By guiding you through these domains, ChatGPT will learn both the scope and angle of your interest and begin to adjust its response to provide better answers.
For example, consider this prompt: “Can you give me advice on how to stay healthy.” In that question, ChatGPT doesn’t know who “you” are, nor who “I” is, nor what you mean by “being healthy”. Instead, try this: “Pretend you’re a medical doctor, a nutritionist, and a personal trainer. Create a two-week meal and exercise plan for a 56-year-old man to enhance heart health. With this, you’ve given the glider a more specific flight plan spanning the areas of therapy, nutrition and motivation.
If you want something more precise, you can activate a few more dimensions. For example, add in: “And I want to lose some weight and build muscle, and I want to spend 20 minutes a day on exercise, and I can’t do pull-ups and I hate tofu.” ChatGPT will provide the output taking into account all your active dimensions. Each dimension can be presented together or in sequence.
flight plan
The dimensions you add through prompts can be informed by ChatGPT’s answers along the way. Here’s an example: “Pretend you’re an expert on cancer, nutrition, and behavior change. Propose 8 behavior-change interventions to reduce cancer rates in rural communities. ChatGPT will conscientiously submit eight interventions.
Let’s pick the three ideas that seem most promising. You could follow a prompt to encourage more details and start putting it into a format that could be used for a public message: “Combine the concepts of Ideas 4, 6, and 7 to create 4 new possibilities—give each a tagline, and outline the details.” Now suppose that intervention 2 looks promising. You can prompt ChatGPT to make it even better: “Present six criticisms of Intervention 2, and then redesign it to address the criticisms.” ChatGPT works better if you first focus on and highlight the dimensions that you consider particularly important.
For example, if you really care about the behavior-change aspect of the rural cancer rate landscape, you can force ChatGPT to get more granular and add more weight and depth to that dimension before going down the intervention path.
You can do this by first prompting: “Classify behavior-change techniques into 6 named categories. Within each, describe three approaches and name two important researchers in the category. This will better activate the behavior-change dimension, allowing ChatGPT to incorporate this knowledge into subsequent explorations.
There are several categories of prompt elements that you can include to activate dimensions of interest.
There is a domain, such as “machine learning approach.” The second is expertise, such as “responding as an economist with Marxist leanings.” and the second is the output style, such as “Write this as an essay for The Economist.” You can also specify audiences, such as “Create and describe 5 clusters of our customer-types and write product descriptions targeting each.” Explore, Not Answer By rejecting the search engine metaphor and instead adopting a transdimensional glider metaphor, you can better understand how ChatGPT works and navigate more effectively toward valuable insights.
Interaction with ChatGPT is not conducted as a simple or indirect question-and-answer session, but as an interactive conversation that progressively builds knowledge for both the user and the chatbot.
The more information you give it about your interests and the more feedback it gets on its responses, the better its answers and suggestions will be. The richer the journey, the richer the destination.
However, it is important to make proper use of the information provided. The facts, details and references presented by ChatGPT are not taken from the verified sources.
These are modeled on the basis of its training on a huge but non-curated set of data. ChatGPT will generate a medical diagnosis in the same way it writes a Harry Potter story, which means it’s a bit of a fixer.
You should always critically evaluate the specific information it provides and treat its outputs as explorations and suggestions rather than hard facts.
Treat its contents as hypothetical conjectures that require further verification, analysis, and filtering by you, the human pilot.