🤙🗣️ Score Winna Kine Action Wit ChatGPT: Da Golden Prompts 🎖️
Aloha and welcome back to On Tech: A.I., one newsletter fo’ give you da scoops on artificial intelligence (A.I.), how da buggah works, and how fo’ use um. 🤖💡📚
No long time ago, my co-workahs Cade Metz and Kevin Roose wen’ talk story ‘bout da ins and outs of A.I., including chatbots like OpenAI’s ChatGPT, Microsoft’s Bing and Google’s Bard. Now, we’re back with one new kuleana: fo’ help you learn how fo’ use A.I. fo’ da max. 🎯📈🔄
Plenny peeps from all ova da place — students, coders, artists, and accountants — stay trying out how fo’ use A.I. tools. Bosses stay looking fo’ workahs who can handle um. Pretty soon, if not already, you going have da chance fo’ use A.I. fo’ make your work and personal life mo’ bettah. 🎓💼🎨🧮
As da Times’s personal tech writer, I’m hea fo’ help you figga out how fo’ use dese tools safe kine and responsible, and fo’ make plenty parts of your life mo’ bettah. 📝🔒🛠️
Today’s newsletter going talk about two general ways dat going be useful fo’ plenny kine situations. 📖🚀🔑
Den, in da coming weeks, I going share mo’ specific tips fo’ diff’rent parts of your life, like parenting and ohana life, work, organizing in your personal life, learning/education, creativity, and shopping. 🏠🏢📚🎭🛍️
But first, couple things fo’ remember:
If you stay worried about privacy, no include personal stuff like your name and where you work. Da tech companies say your data is used fo’ train their systems, which means oddah peeps could see your information. 🕵️♂️🔍📵
No share confidential data. Your boss might have specific rules or restrictions, but in general, putting trade secrets or sensitive info is one supah bad idea. 🤫🔐💼
Hallucinations: Chatbots run on one tech called one big language model, or L.L.M., which gets its skills from analyzing big amounts text from da internet. Plenny stuff online is wrong, and chatbots might repeat those not true things. Sometimes, while trying fo’ predict patterns from their big kine training data, they can make things up. 🌐💭🤥
Da golden prompts ChatGPT, Bing and Bard are some of da most popular A.I. chatbots. (Fo’ use ChatGPT, you gotta make one OpenAI account, and you need a subscription fo’ da most advanced version. Bing requires you fo’ use Microsoft’s Edge web browser. Bard requires one Google account.) 🗣️🎟️🌐
Even though they seem easy fo’ use — you type something in one box and get answers! — if you ask questions da wrong way, you going get generic, not useful and, sometimes, all buggah kine wrong answers. 📝💬❓
Get one trick fo’ typing in da precise words and framing fo’ get da most helpful answers. I call these da golden prompts. 🌟🔤📲
Da peeps who stay getting da most out of da chatbots have been using kine variations of these strategies:
“Act as if.” Start your prompt with these magic words going tell da bot fo’ act like one expert. Fo’ example, if you type “Act as if you are a tutor for the SATs” or “Act as if you are a personal trainer” going guide da bots fo’ act like peeps in those jobs. 👨🏫🤖💪
These prompts give extra context fo’ da A.I. fo’ create its response. Da A.I. doesn’t really understand what it means to be a tutor or a personal trainer. Instead, da prompt is helping da A.I. fo’ draw on specific statistical patterns in its training data. 🤔📊🔄
Weak prompt with no guidance going give less helpful results. If all you type is “What should I eat this week?” da chatbot going give you one generic list of meals fo’ a balanced diet, like turkey stir fry with side of colorful veggies fo’ dinner (which, to me, sounds very “meh”). 🤷♂️🍽️🦃
“Tell me what else you need to do this.” Fo’ get results dat are more personalized — like, health advice fo’ your specific body type or medical conditions — invite da bot fo’ ask mo’ info. 💁♂️💡🏋️♀️
In da personal trainer example, one prompt could be: “Act as if you are my personal trainer. Create one weekly workout regimen and meal plan fo’ me. Tell me what else you need to do this.” Da bot might then ask you fo’ your age, height, weight, dietary restrictions and health goals to make one custom weeklong meal plan and fitness routine fo’ you. 📋🥗🏃♀️
If you don’t get good answers on your first try, no give up right away. Bettah yet, in da words of Ethan Mollick, one professor at the Wharton School of the University of Pennsylvania, treat da bot like one human intern: “When it makes a mistake, point that out and ask for it to do better.” Be forgiving and patient, and you going get bettah results. 🧐🎓🔄
Thread your chatbot conversations Aftah you get da hang of prompts, you can make your chatbot mo’ helpful over time. Da key hea is to avoid treating your chatbot like one web search and starting with one fresh question each time. Instead, keep several threads of conversations open and add to them over time. 💭🗨️🔄
This strategy is easiest with ChatGPT. Bing requires you fo’ reset your conversations once in a while, and Bard no make um as easy fo’ jump between conversation threads. 🔄🤖🔀
Natalie Choprasert, one business owner in Sydney, Australia, who gives companies advice on how to use A.I., uses ChatGPT as one business coach and one executive assistant. She keeps separate conversations running side by side fo’ each of these roles. 🧑💼🇦🇺📋
For da business coach thread, she shares insights about her professional background and da company’s goals and problems. For the executive assistant thread, she shares scheduling info, like da clients she’s meeting with. 📈🗂️🗓️
“It builds up and trains properly, so when I ask it a question later on, it going be in da right context and it going give me answers close to what I’m looking for,” Choprasert said. 🎯📚🗨️
She shared one bonus golden prompt that has trained her assistants to be extra helpful: Apply a framework. She recently read “Clockwork,” one book about setting up a business. When she asked ChatGPT-the-business-coach fo’ give advice using the framework of “Clockwork,” she was stoked to see that it could incorporate principles from the book into one action plan fo’ make her company bigger. 📖⚙️🚀
NOW IN ENGLISH
🤙🗣️ Unlock the Power of ChatGPT with These Golden Prompts: A Guide for All Life’s Facets 🎖️
Welcome back to ‘On Tech: A.I.’, your go-to newsletter that educates you about artificial intelligence, its workings, and its usability. Not so long ago, my fellow workers, Cade Metz and Kevin Roose, unraveled the inner mechanics of A.I., including chatbots like OpenAI’s ChatGPT, Microsoft’s Bing, and Google’s Bard. We’re back with a renewed objective: to help you unleash the full potential of A.I.
Folks from diverse walks of life—students, coders, artists, accountants—are dabbling with A.I. tools. Employers are seeking individuals proficient in using them. Soon enough, if not already, you’ll get the opportunity to leverage A.I. to enhance and streamline your professional and personal life.
As The Times’s personal tech columnist, my mission is to guide you on utilizing these tools safely and responsibly to improve various facets of your life. This newsletter will discuss two broad strategies that can be beneficial in multiple scenarios. 🗒️💼🤖
In the following weeks, I’ll provide you with more specific tips for different aspects of your life: parenting and family life, work, personal life organization, learning/education, creativity, and shopping.👨👩👧👦💼🏠🎨🛍️
Before we begin, let’s get some common-sense warnings out of the way:
- If you’re concerned about privacy, exclude personal details like your name and workplace. Tech companies claim your data is used to train their systems, which implies others could potentially see your information.🔐📚
- Avoid sharing confidential data. Your employer may have specific guidelines or restrictions, but generally, entering trade secrets or sensitive information is a grave mistake. 🚫🔒
- Hallucinations: Chatbots are powered by a technology called a large language model, or L.L.M., which acquires its abilities by analyzing vast amounts of digital text sourced from the internet. Lots of content on the web is inaccurate, and chatbots may repeat those falsehoods. Sometimes, in an attempt to predict patterns from their extensive training data, they can fabricate things.📚🌐🤥
The Golden Prompts
ChatGPT, Bing, and Bard are among the most popular A.I. chatbots. Using them might seem straightforward — you enter something in a box and receive answers! — but posing questions incorrectly can lead to generic, unhelpful, and at times, downright incorrect responses.🗨️🚫
It turns out there’s a knack to inputting precise words and framing to yield the most helpful responses. I call these ‘the golden prompts’. The individuals benefiting most from these chatbots have been employing variations of these strategies:🥇🔍
“Act as if.” Initiating your prompt with these magic words will instruct the bot to mimic an expert. For instance, typing “Act as if you are a tutor for the SATs” or “Act as if you are a personal trainer” will guide the bots to pattern themselves after individuals in those professions. These prompts furnish additional context for the A.I. to generate its response. The A.I. doesn’t truly comprehend what being a tutor or a personal trainer means. Rather, the prompt aids the A.I. in drawing on specific statistical patterns from its training data.📊📚
A weak prompt without guidance will yield less beneficial results. If all you input is “What should I eat this week?” the chatbot will generate a generic list of meals for a balanced diet, such as turkey stir fry with a side of colorful veggies for dinner (which, to me, sounds very “meh”). 🥦🍗
“Tell me what else you need to do this.” To obtain more personalized results — say, health advice for your specific body type or medical conditions — invite the bot to request additional information. For example, in the personal trainer scenario, a prompt could be: “Act as if you are my personal trainer. Create a weekly workout regimen and meal plan for me. Tell me what else you need to do this.” The bot might then inquire about your age, height, weight, dietary restrictions, and health goals to tailor a weeklong meal plan and fitness routine for you. 🏋️♀️🍽️💪
If you don’t receive satisfactory responses on your initial attempt, don’t surrender immediately. Better yet, in the words of Ethan Mollick, a professor at the Wharton School of the University of Pennsylvania, treat the bot as if it were a human intern: “When it makes a mistake, point that out and ask for it to do better.” Be forgiving and patient, and you’re likely to get improved results. 🤖👥
Thread Your Chatbot Conversations
Once you master prompts, you can make your chatbot more useful over time. The trick here is not to treat your chatbot like a web search, starting with a new query each time. Instead, maintain multiple conversation threads and augment them over time.
This tactic is simplest with ChatGPT. Bing requires you to reset your conversations occasionally, and Bard doesn’t make it easy to switch between conversation threads.
Natalie Choprasert, an entrepreneur in Sydney, Australia, who guides companies on using A.I., employs ChatGPT as a business coach and an executive assistant. She maintains separate conversations for each of these roles.
For the business coach thread, she shares insights about her professional background and the company’s goals and challenges. For the executive assistant thread, she shares scheduling information, like the clients she’s meeting. “It builds up and trains properly, so when I ask a question later, it will be in the right context, and it will provide answers close to what I’m looking for,” Choprasert said. 📅🗨️🎯
She shared a bonus golden prompt that has trained her assistants to be extra useful: Apply a framework. She recently read “Clockwork,” a book about setting up a business. When she asked ChatGPT-the-business-coach to provide advice using the framework of “Clockwork,” she was thrilled to discover that it could integrate principles from the book into an action plan for expanding her business. 📖⚙️🚀
