We already have many of the climate solutions we need. But scaling them is hard. The Green Blueprint is a show about the people who are architecting the clean economy. Every other week, host Lara Pierpoint profiles the founders, investors, and organizational leaders who are solving complex challenges in the quest to build climate technologies fast.
…
continue reading
תוכן מסופק על ידי Quiet. Please. כל תוכן הפודקאסטים כולל פרקים, גרפיקה ותיאורי פודקאסטים מועלים ומסופקים ישירות על ידי Quiet. Please או שותף פלטפורמת הפודקאסט שלהם. אם אתה מאמין שמישהו משתמש ביצירה שלך המוגנת בזכויות יוצרים ללא רשותך, אתה יכול לעקוב אחר התהליך המתואר כאן https://he.player.fm/legal.
Player FM - אפליקציית פודקאסט
התחל במצב לא מקוון עם האפליקציה Player FM !
התחל במצב לא מקוון עם האפליקציה Player FM !
Unlock AI Mastery: Powerful Prompting Technique Reveals Communication Secrets
MP3•בית הפרקים
Manage episode 502721876 series 3494377
תוכן מסופק על ידי Quiet. Please. כל תוכן הפודקאסטים כולל פרקים, גרפיקה ותיאורי פודקאסטים מועלים ומסופקים ישירות על ידי Quiet. Please או שותף פלטפורמת הפודקאסט שלהם. אם אתה מאמין שמישהו משתמש ביצירה שלך המוגנת בזכויות יוצרים ללא רשותך, אתה יכול לעקוב אחר התהליך המתואר כאן https://he.player.fm/legal.
Welcome to “I am GPTed,” the show where I, Mal—the Misfit Master of AI and formerly world-class tech skeptic—take you from “AI is probably selling my data” to “Hey, did I just automate my grocery list?” All without making you learn Klingon or memorize the difference between stochastic and existential crises. So, if you’re tired of jargon-laden sermons and want AI you can actually use, you’re in the right place.
Today, we’re demystifying one *specific prompting technique*: the mighty “few-shot prompting.” I know, it sounds like either a sports move or a cheap cocktail. Here’s what it means: **you give AI a few examples of what you want before unleashing it on your real task**. Picture teaching a dog to fetch by actually—brace yourself—throwing a stick a few times first. Revolutionary.
Let’s do a “before and after,” because nothing motivates like proof I used to be terrible at this:
- Before, I’d just type: “Write an email to my boss about needing a day off.”
- AI’s Response: “Hello Boss. Day off please. Kindly Regards.” Which, sure, screams professionalism if you’re a confused time traveler.
- After, using few-shot prompting:
- I prompt: “Here are two sample emails. [Example 1: Friendly, clear professional tone. Example 2: A bit formal, but positive.] Now, write one to my boss about needing Friday off.”
- AI’s Response: “Good morning, Pat. I’d appreciate Friday off to handle a family matter. Let me know if there’s coverage needed—I can coordinate. Thanks for understanding!” See? It’s alive, Jim!
That’s *few-shot prompting*: show, don’t just tell. If you’re like me and have flashbacks to middle school presentations where no one explained the assignment… let AI’s confusion be a lesson.
*Practical use case for real life, coming at you fast*: Automate your weekly shopping list, but level up. Give AI examples: “Each week, I buy these basic items: eggs, bread, bananas. If my calendar mentions ‘friends over’ or ‘party,’ add chips, guac, extra drinks.” Now, feed it your upcoming calendar and—bam—AI-generated shopping plans that adjust to your week. Who needs a butler when you have bits?
Confession corner—because what’s a show without public self-flagellation? My rookie mistake: I kept firing off one-line demands and then getting annoyed when my results were… let’s say, “minimalist.” Turns out, the AI is not a mind reader (my therapist’s job remains secure). **Biggest blunder?** Never giving examples or context. Solution: treat AI like a toddler meeting your in-laws for the first time. Be *painfully* specific. Fewer tantrums, more useful answers.
Let’s get to the hands-on bit—an exercise to flex your AI interaction muscle:
Tonight, pick a small writing task. Come up with two example outputs—good or bad, doesn’t matter. Toss them in with your real request. Compare the AI’s reply to your earlier attempts. Bask in the glory of incremental progress, or at least fewer existential emails.
Final tip for evaluating your AI-generated gems: Don’t just ask, “Does this make sense?” Instead, check: is the tone right for my audience, does the information actually answer my need, and could I show this to another human without crying? If not, go back and refine—give more details or tweak your examples.
That’s it for today’s episode of “I am GPTed.” If you got something useful—or even a new favorite way to phrase regret—smash that subscribe button. Thanks for hanging out with me, Mal, as we do our part to make AI advice just a little more human (with only a reasonable amount of sarcasm).
This has been a Quiet Please production—find out more at quietplease.ai. Now, go forth and prompt responsibly!
…
continue reading
Today, we’re demystifying one *specific prompting technique*: the mighty “few-shot prompting.” I know, it sounds like either a sports move or a cheap cocktail. Here’s what it means: **you give AI a few examples of what you want before unleashing it on your real task**. Picture teaching a dog to fetch by actually—brace yourself—throwing a stick a few times first. Revolutionary.
Let’s do a “before and after,” because nothing motivates like proof I used to be terrible at this:
- Before, I’d just type: “Write an email to my boss about needing a day off.”
- AI’s Response: “Hello Boss. Day off please. Kindly Regards.” Which, sure, screams professionalism if you’re a confused time traveler.
- After, using few-shot prompting:
- I prompt: “Here are two sample emails. [Example 1: Friendly, clear professional tone. Example 2: A bit formal, but positive.] Now, write one to my boss about needing Friday off.”
- AI’s Response: “Good morning, Pat. I’d appreciate Friday off to handle a family matter. Let me know if there’s coverage needed—I can coordinate. Thanks for understanding!” See? It’s alive, Jim!
That’s *few-shot prompting*: show, don’t just tell. If you’re like me and have flashbacks to middle school presentations where no one explained the assignment… let AI’s confusion be a lesson.
*Practical use case for real life, coming at you fast*: Automate your weekly shopping list, but level up. Give AI examples: “Each week, I buy these basic items: eggs, bread, bananas. If my calendar mentions ‘friends over’ or ‘party,’ add chips, guac, extra drinks.” Now, feed it your upcoming calendar and—bam—AI-generated shopping plans that adjust to your week. Who needs a butler when you have bits?
Confession corner—because what’s a show without public self-flagellation? My rookie mistake: I kept firing off one-line demands and then getting annoyed when my results were… let’s say, “minimalist.” Turns out, the AI is not a mind reader (my therapist’s job remains secure). **Biggest blunder?** Never giving examples or context. Solution: treat AI like a toddler meeting your in-laws for the first time. Be *painfully* specific. Fewer tantrums, more useful answers.
Let’s get to the hands-on bit—an exercise to flex your AI interaction muscle:
Tonight, pick a small writing task. Come up with two example outputs—good or bad, doesn’t matter. Toss them in with your real request. Compare the AI’s reply to your earlier attempts. Bask in the glory of incremental progress, or at least fewer existential emails.
Final tip for evaluating your AI-generated gems: Don’t just ask, “Does this make sense?” Instead, check: is the tone right for my audience, does the information actually answer my need, and could I show this to another human without crying? If not, go back and refine—give more details or tweak your examples.
That’s it for today’s episode of “I am GPTed.” If you got something useful—or even a new favorite way to phrase regret—smash that subscribe button. Thanks for hanging out with me, Mal, as we do our part to make AI advice just a little more human (with only a reasonable amount of sarcasm).
This has been a Quiet Please production—find out more at quietplease.ai. Now, go forth and prompt responsibly!
106 פרקים
Unlock AI Mastery: Powerful Prompting Technique Reveals Communication Secrets
I am GPTed - what you need to know about Chat GPT, Bard, Llama, and Artificial Intelligence
MP3•בית הפרקים
Manage episode 502721876 series 3494377
תוכן מסופק על ידי Quiet. Please. כל תוכן הפודקאסטים כולל פרקים, גרפיקה ותיאורי פודקאסטים מועלים ומסופקים ישירות על ידי Quiet. Please או שותף פלטפורמת הפודקאסט שלהם. אם אתה מאמין שמישהו משתמש ביצירה שלך המוגנת בזכויות יוצרים ללא רשותך, אתה יכול לעקוב אחר התהליך המתואר כאן https://he.player.fm/legal.
Welcome to “I am GPTed,” the show where I, Mal—the Misfit Master of AI and formerly world-class tech skeptic—take you from “AI is probably selling my data” to “Hey, did I just automate my grocery list?” All without making you learn Klingon or memorize the difference between stochastic and existential crises. So, if you’re tired of jargon-laden sermons and want AI you can actually use, you’re in the right place.
Today, we’re demystifying one *specific prompting technique*: the mighty “few-shot prompting.” I know, it sounds like either a sports move or a cheap cocktail. Here’s what it means: **you give AI a few examples of what you want before unleashing it on your real task**. Picture teaching a dog to fetch by actually—brace yourself—throwing a stick a few times first. Revolutionary.
Let’s do a “before and after,” because nothing motivates like proof I used to be terrible at this:
- Before, I’d just type: “Write an email to my boss about needing a day off.”
- AI’s Response: “Hello Boss. Day off please. Kindly Regards.” Which, sure, screams professionalism if you’re a confused time traveler.
- After, using few-shot prompting:
- I prompt: “Here are two sample emails. [Example 1: Friendly, clear professional tone. Example 2: A bit formal, but positive.] Now, write one to my boss about needing Friday off.”
- AI’s Response: “Good morning, Pat. I’d appreciate Friday off to handle a family matter. Let me know if there’s coverage needed—I can coordinate. Thanks for understanding!” See? It’s alive, Jim!
That’s *few-shot prompting*: show, don’t just tell. If you’re like me and have flashbacks to middle school presentations where no one explained the assignment… let AI’s confusion be a lesson.
*Practical use case for real life, coming at you fast*: Automate your weekly shopping list, but level up. Give AI examples: “Each week, I buy these basic items: eggs, bread, bananas. If my calendar mentions ‘friends over’ or ‘party,’ add chips, guac, extra drinks.” Now, feed it your upcoming calendar and—bam—AI-generated shopping plans that adjust to your week. Who needs a butler when you have bits?
Confession corner—because what’s a show without public self-flagellation? My rookie mistake: I kept firing off one-line demands and then getting annoyed when my results were… let’s say, “minimalist.” Turns out, the AI is not a mind reader (my therapist’s job remains secure). **Biggest blunder?** Never giving examples or context. Solution: treat AI like a toddler meeting your in-laws for the first time. Be *painfully* specific. Fewer tantrums, more useful answers.
Let’s get to the hands-on bit—an exercise to flex your AI interaction muscle:
Tonight, pick a small writing task. Come up with two example outputs—good or bad, doesn’t matter. Toss them in with your real request. Compare the AI’s reply to your earlier attempts. Bask in the glory of incremental progress, or at least fewer existential emails.
Final tip for evaluating your AI-generated gems: Don’t just ask, “Does this make sense?” Instead, check: is the tone right for my audience, does the information actually answer my need, and could I show this to another human without crying? If not, go back and refine—give more details or tweak your examples.
That’s it for today’s episode of “I am GPTed.” If you got something useful—or even a new favorite way to phrase regret—smash that subscribe button. Thanks for hanging out with me, Mal, as we do our part to make AI advice just a little more human (with only a reasonable amount of sarcasm).
This has been a Quiet Please production—find out more at quietplease.ai. Now, go forth and prompt responsibly!
…
continue reading
Today, we’re demystifying one *specific prompting technique*: the mighty “few-shot prompting.” I know, it sounds like either a sports move or a cheap cocktail. Here’s what it means: **you give AI a few examples of what you want before unleashing it on your real task**. Picture teaching a dog to fetch by actually—brace yourself—throwing a stick a few times first. Revolutionary.
Let’s do a “before and after,” because nothing motivates like proof I used to be terrible at this:
- Before, I’d just type: “Write an email to my boss about needing a day off.”
- AI’s Response: “Hello Boss. Day off please. Kindly Regards.” Which, sure, screams professionalism if you’re a confused time traveler.
- After, using few-shot prompting:
- I prompt: “Here are two sample emails. [Example 1: Friendly, clear professional tone. Example 2: A bit formal, but positive.] Now, write one to my boss about needing Friday off.”
- AI’s Response: “Good morning, Pat. I’d appreciate Friday off to handle a family matter. Let me know if there’s coverage needed—I can coordinate. Thanks for understanding!” See? It’s alive, Jim!
That’s *few-shot prompting*: show, don’t just tell. If you’re like me and have flashbacks to middle school presentations where no one explained the assignment… let AI’s confusion be a lesson.
*Practical use case for real life, coming at you fast*: Automate your weekly shopping list, but level up. Give AI examples: “Each week, I buy these basic items: eggs, bread, bananas. If my calendar mentions ‘friends over’ or ‘party,’ add chips, guac, extra drinks.” Now, feed it your upcoming calendar and—bam—AI-generated shopping plans that adjust to your week. Who needs a butler when you have bits?
Confession corner—because what’s a show without public self-flagellation? My rookie mistake: I kept firing off one-line demands and then getting annoyed when my results were… let’s say, “minimalist.” Turns out, the AI is not a mind reader (my therapist’s job remains secure). **Biggest blunder?** Never giving examples or context. Solution: treat AI like a toddler meeting your in-laws for the first time. Be *painfully* specific. Fewer tantrums, more useful answers.
Let’s get to the hands-on bit—an exercise to flex your AI interaction muscle:
Tonight, pick a small writing task. Come up with two example outputs—good or bad, doesn’t matter. Toss them in with your real request. Compare the AI’s reply to your earlier attempts. Bask in the glory of incremental progress, or at least fewer existential emails.
Final tip for evaluating your AI-generated gems: Don’t just ask, “Does this make sense?” Instead, check: is the tone right for my audience, does the information actually answer my need, and could I show this to another human without crying? If not, go back and refine—give more details or tweak your examples.
That’s it for today’s episode of “I am GPTed.” If you got something useful—or even a new favorite way to phrase regret—smash that subscribe button. Thanks for hanging out with me, Mal, as we do our part to make AI advice just a little more human (with only a reasonable amount of sarcasm).
This has been a Quiet Please production—find out more at quietplease.ai. Now, go forth and prompt responsibly!
106 פרקים
כל הפרקים
×ברוכים הבאים אל Player FM!
Player FM סורק את האינטרנט עבור פודקאסטים באיכות גבוהה בשבילכם כדי שתהנו מהם כרגע. זה יישום הפודקאסט הטוב ביותר והוא עובד על אנדרואיד, iPhone ואינטרנט. הירשמו לסנכרון מנויים במכשירים שונים.