Ep 91 - Revisiting the Basics: Getting More Out of ChatGPT

Summary

The ChatGPT Experiment – Ep 91: Revisiting the Basics: Getting More Out of ChatGPT

Host: Cary Weston
Date: October 21, 2025

---

Episode Overview

Theme:
Cary Weston takes listeners on a reflective walk to revisit the fundamental principles of using ChatGPT, aimed at both beginners and regular users. He emphasizes that deepening one’s understanding of the “basics” unlocks greater value from AI, regardless of experience level. The episode explores evergreen strategies to get more out of ChatGPT, focusing on mindset, conversational engagement, and feedback techniques applicable across various AI platforms.

---

Key Discussion Points & Insights

1. Revisiting the Value of Basics

• Cary starts with a metaphor: Like song lyrics that change meaning over time, the basics of ChatGPT reveal new value as users gain experience. ([00:01])
• Even advanced users benefit from “seeing things through a new lens” by returning to foundational ideas.

2. AI as an “Amazing Intern”

• Cary urges users to think of ChatGPT as a “highly capable tool ready to serve you, but it needs your guidance, it needs context.” ([03:24])
• Instead of treating it as an autonomous expert, give it clear instructions and context as you would to a skilled assistant.

3. The Mindset Matters—Not Just the Platform

• Evergreen best practices apply across platforms like Claude, Gemini, Copilot, and Google’s AI—not just ChatGPT. What matters most is how you approach and interact with these tools. ([02:14])
• “It’s more of a mindset... It’s the way in which you see the opportunities and are able to talk to it and work with it.” ([02:26])

4. The Four-Part Framework for Stronger Prompts

• Cary shares his tried-and-true method for delegating tasks to ChatGPT:
  1. What are we doing?
  2. Why are we doing it?
  3. What does success look like?
  4. Do you have any questions for me? ([04:38])
• This framework works not just at the beginning, but throughout the conversation—mid-task feedback should reference these principles.

Quote:

“Those four principles are helpful if you’re delegating a task. And that’s really what you’re doing with ChatGPT—using software to delegate a task.” ([05:08])

5. Emphasizing Natural Conversation

• Cary notes that ChatGPT thrives on natural language:
  • More helpful than overthinking prompts or using code-like language.
  • “Give it as much detail as you can, as if you were talking to a friend in a natural conversation.” ([05:44])
  • It’s okay if instructions are “not in order, not sequential, just real.” ([06:00])

6. Treating Output as a Work in Progress

• Output from ChatGPT is rarely a final answer—think of it as brainstorming, not gospel.
• Cary encourages users to iteratively coach ChatGPT: Take what works, give feedback, suggest specific changes, and see improvement with each round.
  • Example: Building a restaurant menu. Don’t just accept the list; explain what you like/dislike about individual items to refine results. ([07:47])

Quote:

“Anytime you get something back, it’s not right or wrong. Don’t think of what you’re getting from ChatGPT as right or wrong. Think of it always as in process.” ([07:25])

7. The Importance of Feedback and Ongoing Guidance

• Cary shares a story about a counselor who uses ChatGPT for writing but gets frustrated by lack of exactness.
  • Cary explains that, like with people, ChatGPT requires ongoing guidance and context—simply telling it to "stop" isn’t enough. ([09:25])
  • The more context and feedback you give, the better the output becomes.
• Analogy: When a client tells an agency, “I don’t like it,” it’s impossible to make improvements without specifics. The same goes for ChatGPT. ([12:19])

Quote:

“So talking to it, reasoning... it needs to be involved in the conversation. It’s not a one-way act.” (14:08)

8. Growing with Experience and Context

• Recur to the song lyric metaphor: The “basics” will reveal new learning as you interact more deeply with ChatGPT.
• For power users: “Go deeper in conversations... see if you get more value, more clarity, more detail.” ([15:47])
• For beginners: “The more that you engage with it and enable it to engage with you, you’re going to find a far more useful tool in your toolbox.” ([16:32])

9. Curiosity as the Most Powerful Tool

• Cary wraps up by championing curiosity as the key to getting the most out of ChatGPT:

“It’s your curiosity that will serve you the best. It’s the most important element of you getting any meaning out of what you’re doing with ChatGPT.” ([16:52])

---

Memorable Quotes & Moments

• “I want you to be more confident, to be curious, to try things...” ([02:55])
• “If you think less about prompts and codes and more about natural conversation... it works better.” ([05:30])
• “Don’t accept any output from ChatGPT as either its best answer or its only answer. There’s always more—more depth, more detail, more explanation, a different perspective.” ([07:09])
• “ChatGPT is the same way [as a person]. It thrives off natural conversation.” ([14:37])

---

Key Timestamps

| Timestamp | Segment | |---|---| | 00:01 | Cary’s introduction and song lyric analogy | | 02:14 | Applying basics across different AI platforms | | 03:24 | AI as a capable intern; the need for context | | 04:38 | Four-part framework for effective prompting | | 05:44 | The value of natural conversation with ChatGPT | | 07:25 | Treating responses as iterative, not final | | 09:25 | Counselors and coders: the need for detailed guidance | | 12:19 | Feedback analogy: client-agency communication | | 14:08 | Two-way conversational engagement | | 15:47 | Encouragement for experienced users to deepen conversations | | 16:32 | Advice for beginners on engagement | | 16:52 | The importance of curiosity |

---

Conclusion

Cary Weston delivers a reflective yet practical refresher on the basics that create confident, effective ChatGPT users. He reminds listeners that their willingness to converse, give feedback, and experiment—powered by curiosity—is more valuable than mastering technical prompts. By using a conversational, context-rich approach rooted in real-world analogies, Cary ensures listeners of all backgrounds walk away empowered to get more out of their AI tools—today and in the future.

Final thought:

“Stay curious. It’s the most important element of you getting any meaning out of what you’re doing with ChatGPT.” ([16:52])

Transcript

[A]

Hey, it's Kerry.

[B]

Hey.

[A]

Like those song lyrics that you listen to through the years and take on different meaning to you through life experiences, the new lenses that you have. I think ChatGPT revisiting the basics is very, very helpful. I'm going to take you on a walk with me this week. So my apology for my breath changing, but this is right after a presentation to a couple hundred people where I realized that the basics do matter, regardless of whether you are new to this or use it every day. So I'm going to walk you through some of the basics that will keep you getting more value out of ChatGPT. That sounds valuable to you? I'll see you on the other side of the music.

[B]

So while most people, myself included, listen to podcasts while they're out for a walk, I thought with a few things in my head I would spend my time during the walk recording a podcast. So if you hear me breathe a little differently, that's why. So today I thought I would revisit some of the basics. More conversations I'm having with more people, even those that are listening to the podcast and those that tell me they're comfortable and use ChatGPT on a regular basis. There are some evergreen principles that once we start talking about them and really explaining them in more detail, people see new opportunities. So today, if you are a regular user of ChatGPT, there might be one thing here that you can look at through a new lens to help you get a little bit more out of it. And if you're not that familiar with ChatGPT, then there are some things here that you can try and always my only goal with sharing things that I do and know and learn is to help you be more confident in being curious. That's it. Just to help you be more confident in being curious on exploring this ChatGPT thing. Now, here's the kicker. Some of you listening, and maybe many of you listening are working inside an environment that has other aspects of AI in it. So what's very popular is you might be in a company that has a Microsoft infrastructure. So maybe you use Outlook and maybe you use Word and spreadsheets, right? SharePoint, all that. Well, Microsoft has their own AI, as does Google, and as does other places. And their AI is being developed to help you do more with that software. But the reason I say I want you to be more confident, to be curious, to try things are many of the things that we talk about on the show from an evergreen best practice point of view applies to those other softwares as well, because it's more of a mindset. It's not really how to use this feature or how to use this menu or what does this button do. It's the way in which you see the opportunities and are able to talk to it and work with it, you know, and get more out of it. So the things that work well in ChatGPT work well in Claude, they work well in Gemini, they work well in Copilot. So it's really understanding that these are evergreen principles they would talk about and not just features and functions and benefits of one particular software. Right? So the first thing from a basics point of view that I always come back to is the ability of looking at AI as the amazing intern. Meaning this is a highly capable tool ready to serve you, but it needs your guidance, it needs context. So think of an assistant, an intern, that is dedicated to work for you, that's highly capable. It can give you directions on fixing an auto radiator or building a tree house or recipe for an apple pie, or how to make an article or post to social media. But when you talk to it, because it has all of that capability, what works well is giving it context. So it forgets about the treehouse and the radiator and the apple pie. It can only focus on the task that you've told it. Okay? So it's highly capable and it wants to help. And the more context and direction and clarity, the better. So my four part framework that I use over and again, and there's an episode again on this if you want to review it, but what are we doing? Why are we doing it? What does success look like? And do you have any questions for me? If you think about that for a second, working with a person or a model like ChatGPT, those four principles are helpful if you're delegating a task. And that's really what you're doing with ChatGPT is you're using software to delegate a task. And that works better if you think less about prompts and codes and more about natural conversation. And a lot of folks that I talk to that have any kind of coding or programming or computer background, they don't understand what I mean when I say have a conversation with, it literally does not compute, right? And so the best you can do is just have natural language conversations. So give it as much detail as you can, as if you were talking to a friend in a natural conversation. It doesn't have to be in order, it doesn't even have to be sequential, it just has to be real so as you tell it what it's doing, talk to it like you would that amazing intern. Use natural conversational language. The next thing I would share is don't accept any output from ChatGPT as either its best answer or its only answer. There's always more. There are more depth, more detail, more explanation, a different perspective. It's programmed to give you information and as much as you want, as much detail as you want. So anytime you get something back, it's not right or wrong. Don't think of what you're getting from ChatGPT as right or wrong. Think of it always as in process. So let's do it. For instance, if I were to ask it to create a menu for me for my restaurant, and I want five items and I need creative names and ingredients and descriptions based on whatever ingredients I'm going to give it and maybe even give it a theme. One of the presentations I remember giving to a food service company was just that if you're working with a restaurant that wants to do a main themed menu for summer tourist traffic and wants to reuse of some main ingredients in creative ways so it can get more yield out of its core ingredients in the refrigerators and some profit and have some fun doing it, I would tell it that and then I would share with it that I wanted to come up with the titles of the items, the descriptions of the items and the ingredients so I can make the items right. And it might come back and give me those five food items and maybe I like them and maybe I don't, but the answer that it's going to give me is not right or wrong. Think of it as a brainstorming. These are great. This is a great start. Maybe you like 2 of the 5. So tell it why you like 2 of the 5, and tell it why you don't like 3 of the 5, and then hint at it as to what would make the others better, the ones that you don't like. Just last week I had a conversation with a counselor who deals in, you know, pretty heavy stuff in people's life on a daily basis. And he works with his clients to help them deal with and cope with trauma in their life. And to escape that after work. He prefers to work in coding, programming computers, because it's as far away from the world that he lives in during the day as possible. It's absolute, it's fixed, it's resolute, it's black and white. And he shared with me that's his escape. And he shared with me that when he does things with ChatGPT, especially writing. He told me, I'm frustrated because I'll tell it not to do this thing and then it doesn't listen. And so I asked him, well, what other direction did you give it? What other suggestions or follow ups or recommendations did you give it? Did you guide it at all? He said, no, that's for the computer to figure out. I said, okay, let's take this out of the programming world. So if you had a female client who was having troubles at home and was feeling sad and was overwhelmed by sadness because of the situation, would you simply tell her to stop being sad? And he said, no, of course I wouldn't. And I said, why? Because it doesn't work that way. I said, so what would you do? And he said, I would work her through it and give her some support and give her some information and talk it out. You want this tool to stop doing something. Sometimes you can say stop. But if you don't guide it in another direction and give it more data, more feedback, more context as to what you want it to do more of, or what success looks like, or what you'd be happy with, or what you'd like it to model more, it doesn't know. It continues to guess and essentially flounder, much like anyone you're working with would. So that four part framework of what are you doing, why are we doing it, what does success look like? And do you have any questions for me plays a role in, you know, following up in the middle of a engagement with ChatGPT as much as it does at the beginning. Right? So in this case he could have said, listen, you know, we're doing this thing, we're on the right path, but you keep doing this act, or using this term or whatever it was. I forget the specifics of what he was trying to get Chat TBD to stop doing. But in other words, it keeps doing this. So you could literally have a conversation and say, you keep doing this, I don't like it, I don't want it. So help me guide you to do something different. What questions do you have for me? And then ChatGPT would come back and say, well, what don't you like about it? Right? And then follow up and you can kind of go from there. I ran a marketing agency for 25 years and we do a lot of creative work with clients. And honestly, the hardest conversation to have, or not even conversation, that's the point. The hardest feedback to get from clients was I don't like it. Because without any Other context. I'm just continuing to guess as to what it is they might like. Right. And in the same spirit, you know, when a business works with an agency, you think you're working with creative people who will figure it out. And so if I say I don't like that, you know, some people expect another round that's completely different that they will like. Almost like mind reading. Right? Well, what would have made those situations better? And what does make those situations better? His context reasons. You know, I'm not digging this. I don't like blue, right. I don't like sand. I, I don't like the way the font looks, I don't like the colors. But I do think that if we used more red or had sunset or whatever, it's the additional context of shedding light when one success might look like, you know, or even opening the door. Like I've shared with you in my prompts, do you have any questions? Maybe. I'm not even sure what I don't like. So let's talk it out. ChatGPT is the same way. It thrives off natural conversation. So talking to IT reasoning. Sounds funny to say that when we're talking about computer software. I understand, but it's the mindset of working with something that needs fluent and two way conversation. It needs context, it needs background, it needs details, and probably more importantly, it needs to be involved in the conversation. It's not a one way act. Right? So some evergreen tips that I find helpful when I work with groups or individuals that I want to share with you. So again, if you are comfortable with ChatGPT, take a look. If you know this is almost like you ever had a song that you liked 20 years ago and you hear it again 10 years later and 10 years again. And because your life experiences are different, the lyrics mean something new, right? You almost see the same song in a different way. So these basic principles are kind of the same way here. You know, the more you practice them through different experiments and tasks, the more that they take on a different meaning when you hear things like this. So if you have been working with ChatGPT for some time, try to go deeper in the conversations, feedback, color, context, two way exchanges and see if you get more, more value, more clarity, more detail. And if you're just starting, just keep in mind that the more that you engage with it and enable it to engage with you, you're going to find a far more useful tool in your toolbox.

[A]

Right?

[B]

I appreciate you listening. Again, my goal is always to have you be more confident in being curious. So these little things I hope help. And as always, it's your curiosity that will serve you the best. It's the most important element of you getting any meaning out of what you're doing with ChatGPT. So till we talk again, do stay curious. Okay? Talk soon.

[C]

Morning, Zoe. Got donuts.

[D]

Jeff Bridges, why are you still living above our garage?

[C]

Well, I dig the mattress and I want to be in a T mobile commercial like you. Teach me. So Dana.

[D]

Oh no, I'm not really prepared. I couldn't possibly at t mobile get the new iPhone 17 Pro on them. It's design to be the most powerful iPhone yet and has the ultimate pro camera system.

[C]

Wow, impressive. Let me try. T mobile is the best place to get iPhone 17 Pro because they've got the best network.

[B]

Nice.

[D]

Jeffrey, you heard them.

[C]

T mobile is the best place to get the new iPhone 17 Pro on us with eligible trade in in any condition. So what are we having for launch?

[D]

Dude, my work here is done with.

[E]

24 monthly bill credit is on experience beyond for well qualified customers. Plus tax and $35 device connection charge credits ended balance due if you pay off earlier. Cancel Finance agreement. IPhone 17 Pro 256 Gig $21,099.99 and new line minimum $100 plus a month plan with auto PayPal taxes and fees required. Best mobile network in the US based on analysis by Ooklip Speedtest Intelligence data 1H 2025 visit t mobile.com hey guys.

[F]

Finding the perfect gift for the food lover in your life is easy thanks to Goldbelly. Goldbelly ships America's most iconic foods straight from world famous restaurants right to your door. So if you want to treat someone to Joe's Stone Crab from Miami, Franklin Barbecue from Texas or desserts from famous foodies like Ina Garten or Martha Stewart, Gold Belly has you covered. Just go to goldbelly.com and for a limited time get 20% off your first order with promo code gift. That's promo code gift.