
In this solo episode, host Cary Weston shares a powerful tip and challenge for improving ChatGPT responses. He explores how adding "real-world consequences" to a prompt can lead to more accurate, thoughtful, and thorough AI-generated content. By...
Loading summary
A
Breaking news T Mobile Network outperforms expectations in all sectors because T Mobile helps keep you connected from big cities to your hometown on America's largest 5G network. Switch now keep your phone and T Mobile will pay it off at the $800 per line via prepaid card. Visit your local T Mobile location or learn more@t mobile.com KeepAndSwitch up to four lines via virtual prepaid card. Last 15 days qualifying unlock device credit service port in 90 plus days device and eligible carrier and timely redemption required card is no cash access and expires in six months.
B
You are no dummy, but you're kind of acting like one. You used to crush it in school, outsmarting opponents on the field, and now, well, you're still smart, but not exactly challenging yourself. You could be advancing nuclear engineering in the world's most powerful navy. You were born for it. So make the smart choice. You can be smart or you can be nuke smart. Become a nuclear engineer@navy.com nukesmart America's Navy by the Sea well, hello there.
C
Welcome to the ChatGPT Experiment podcast. My name is Kerry Weston. I'm your host. This is the podcast designed to help curious beginners better understand what ChatGPT is, learn from others how they're using it in hopes that you might find a nugget or two that you could use for your personal or professional life. I'm glad you're here. A short episode Today I want to share a tip. I've got both a tip and a challenge for you. I learned something this week as a as background. The experiment is basically just a playground. If you've been listening to the show for some time, you know that I stand behind the phrase of continually being curious is the best way to really learn and understand these things from any level, whether you're a beginner or whether you're someone like me that's been using it for a few years multiple, multiple times a day. It's always interesting to try something and learn something new. So the nugget I want to share with you today is something that I played with this week to get more value, to get a tighter focus and better output from the work that I'm doing. Okay, I'm going to share that with you. Before I get into that, have you checked out the homepage of the ChatGPT experiment website? ChatGPT experiment.com there is a free guide. It's called the Five Essential Tips for ChatGPT Beginners. It's a practical guide to help you make ChatGPT work for your specific needs, save time with some natural conversation approaches and get some real value without having to have technical expertise. Yeah. So I'm just going to share a few reasons that folks have downloaded to make it relatable to you. Perhaps, uh, when folks go and get the guide, they share with me what their biggest goal is, what they're looking to do. And I appreciate you listening to the podcast, coming to site, downloading the resources. I hope it's helpful. But a few folks have said I'd like to improve the quality of my social media posts so that the outcome of the content will stand out more. One person said, I want to start my own marketing agency. Yeah. Imagine the power and the opportunity now you have in starting a one person business and all of the things and tools that come with with AI, ChatGPT and others that used to be necessary or achievable only by having employees. Yeah. So there's a lot of efficiency in bringing ChatGPT into your business. Self knowledge, coaching and finding patterns and themes in my life that I have trouble identified. Right. To create a more effective social media presence. Well, social media tends to be regular when people are looking into ChatGPT, I'm looking to increase efficiency and increase output and reach for our new business. So you can see, you can see that there's a, hopefully some relatable positions here as to why people are downloading, downloading the guide. So chatgpt experiment.com. there's also some, some workshops there that you can check out as well. So here's the tip that I wanted to share with you this week. I use, if you've, if you've listened to the podcast before you've heard me talk about it, I use transcripts, call transcripts on a regular basis almost, I would say probably two or three times a week anyway. I record calls I have with clients or interview for a specific article, reason content needs, and then I utilize that transcript with ChatGPT to create the finished product. It eliminates what I call the busy middle. Right. So when we're creating something, staring at that blinking cursor or that blank white sheet of paper, it's intimidating to go from nothing to a finished product. And it typically can overwhelm the average person in trying to get to that end product. Yeah. So for me there's three phases of creating something. The first phase is the ideation phase. The what are we doing? Why are we doing it? Right? What does success look like? The whole modeling of what is it that we're going to do the second phase is what I call the busy middle, and I'll come back to that in just a second. And the third phase is the polish. It's adding your point of view, it's adding your experience. It's making it uniquely yours and the best product it can be, perhaps more valuable for the end user, whatever it might be, you're putting your spin on it to make it the best. Now, that busy middle, this is where we spend most of our time. And it's taking the ideation, perhaps that first outline that comes from thinking out loud, and then turning it into a draft, right? And that's where the brain power gets drained. That's where the time gets put into. We end up from nothing to something, and then we polish that something to make it the best it can be. Well, that busy middle is exactly where a tool like ChatGPT can come into play and almost take that out of your process. And I do that with call transcripts. And so I interview on purpose. I ask deeper questions, I get more details out. I try to get to the heart of the matter and really pull information out of the person I'm talking to or the task at hand or the call I'm having. And then I can use a tool like my ChatGPT to organize, to get thoughts out, to organize issues that were discussed. You know, what did we talk about? What do you see here? It's an objective partner. You've heard me talk about ChatGPT as an internal and an intern that works for you. If you brought an intern in and sat them next to your desk and said, okay, go do this task without any background, any extra information, any directions, you know that the work that's going to come back is going to be less than stellar, right? And that's where a lot of people dive into CHAT GPT is they just do that. They say, hey, I need this or I want this, and then they're not impressed, or they're disappointed or frustrated, what comes back? But they haven't had that. The input, the background, the direction, right? All the context that an intern would need to do a good job. Well, so in the guide I mentioned at the beginning here, the. The five Practical Ways Guide, the five tips, there are references to other podcast episodes that will share with you how to build that context, how to give that direction, how to build a library of information so that CHAT GPT can go to work for you. And I use that on a regular basis myself. Right? But here's the situation that I ran into this week, even with the transcript. And even with all the directions, I was getting 90% of what I wanted, I would say, and I was wondering why, like, you know, and I've shared with you before the tell me more of my three favorite words using ChatGPT. Because I feel like it gives me 75% on average of what I want and then I have to tone and tone and tone. So here's what I did this week and I want to share it with you and then I want to challenge you as well. So the first thing I did is I thought of it as an intern and I'm just going to read the paragraph to you and you can see where I'm going with this. Said I'm providing a transcript from a session I had with a client and I tell him who the client is. And I said, your job is to deeply analyze the transcript. No surface level responses, no vague summaries. I need precise, detailed and thorough insights because this transcript will be the foundation of our work moving forward. Now, that information I just gave you, I've told you many times to provide context in the role. Right. And give a direction. Uh, it's the next sentence. So the next. Yeah, the next sentence here. I didn't know if there was more than one that really made a difference this week because I wanted to see if I gave it a negative reason. Right. That bad information would affect me. It's probably a poorly way of wording that, but you'll see what I mean here. So after I said the background, I said this. If you miss key aspects, I will be providing incomplete recommendations which will make me look less valuable to the client, leading to job loss, financial ruin, and ultimately my children starving. So accuracy is not optional. Now, a couple reasons I did that. Number one, I was getting answers back regarding the transcript that were incomplete. I knew because that was part of the conversation that there were things in that transcript that wasn't coming back in the analysis. And so I kept asking it, I think you missed this. Yes, the answer would be, you're right in looking at it. I did miss that. And so I kept saying, I don't want you to miss anything. Right. And it would come back again 90% of the time. It would do good stuff. But it was always those holes. So it was this sentence where I said, you know, if you miss key aspects, my children will starve. Right. Let's just fast forward to the end. That made a significant difference because it put stakes on the work that we were doing. And over and over again, it would come back with the answers and it would say, I know your children are my top concern, so I want to be as thorough as possible for you. So whether you're putting your children's starvation on the line or something, you know, different. What I learned this week was if I put real world negative outcomes, if I put a scenario above and beyond would have normally said, which is telling it, you know, I need you to be detailed, I need you to help, then it actually listened and it reasoned and it said, I understand why this is important in a different way. And I'm telling you that the outcome became much more thorough. So I'm sharing this with you in hopes that you can pick out something that you might be able to use with ChatGPT to get that extra level. But I also want to challenge you. I want to see, can you try this and give me feedback. So I want to see if, if you're getting better outcome. Now you don't need a transcript for this. I've used this line a few times this week with non transcript uses with chat GPT. But what I've said is, hey, the work I'm doing is important. This task is going to be important. Let's make it the best it can possibly be. And by the way, if it's not the best it's it can be, here's what's going to happen, right? So these are the, this is what's at stake. And so I guess that's the, that's the headline here is when I shared with it, not only do I want you to be authentic, but here's what's at stake. It understood the stakes of the work that we're doing, the negative consequences of doing a bad job. Now we do this sometimes we do this in our job. We share with people. This is our livelihood, right? And so if you have employees or if you worked with others, understanding the bigger picture of why we're working together as a team and why this matters, right? There's someone's livelihood at stake. Sometimes it's yours. If this is a business or a hobby that you're working on, this is your livelihood at stake. So I found it really interesting when I shared with ChatGPT my livelihood, right? These are what's at stake. It really understood it from an empathetic position. And then I noticed it took an extra level of care to make sure that it was working with me to get the detailed output that I needed the first time. So really interesting as we go through this experiment, how the logic and reasoning of the models that we're using now have transformed from just I understand the task, let's do the task to now it's understanding and able to comprehend the stakes and the negative consequences of doing a poor job. So just a quick tip there today and I want you to try it should you know, you can contact me through the website, but I want to hear if you're able to use this, if this solves a problem that maybe you are having where it gives you extra value in the work that you're doing. Share with me how you're using it. Share with me what the outcome is like and what you found in your own experiment. Okay, the website again, chatgpt experiment. Com, the guides there workshops and whatnot can also contact me. So hey, appreciate you listening to the show and oh, just a quick update. I shared with you the business book episode last week. I checked in with one of my clients who I referenced that we were building a business book with and he's gotten the point where he's so efficient in using this tool that he was actually able to outline, interview and get to a first draft on a second book in a matter of weeks. I shared with you that the first book took 90 days. He got to the second one in a matter of weeks, so really buying in, really honing in the process. And I was proud of him. I was surprised to see how fast he's come on book two, but learning how to get that out of his head and fast forward the process was an amazing, amazing thing to hear. So I hope you're doing amazing things. I'd love to hear if you've got a story you want to share. Um, I've had a few folks reach out to me this week that I need to get back to that have shared with me some stories. I've got some interviews coming up over the next couple weeks. I'm going to share with you here on the podcast. And as always, the experiment continues. Yeah. So stay curious. Curiosity is the name of the game. Keep trying, things get better and see how you can use this for your own personal professional gain. Yeah, okay. Thanks for listening. Until we talk again, stay curious. Talk soon.
A
Breaking News T Mobile Network outperforms expectations in all sectors because T Mobile helps keep you connected from big cities to your hometown on America's largest 5G network. Switch now keep your phone and T Mobile will pay it off up to $800 per line via prepaid card. Visit your local T Mobile location or learn more@t mobile.com KeepAndSwitch up to four lines via virtual prepaid card allowed 15 days qualified unlock device. Credit service port in 90 plus days. Device ineligible. Carrier and timely redemption required. Card has no cash access and expires in six months.
B
You are no dummy, but you're kind of acting like one. You used to crush it in school, outsmarting opponents on the field. And now? Well, you're still smart, but not exactly challenging yourself. You could be advancing nuclear engineering in the world's most powerful Navy. You were born for it. So make the smart choice. You can be smart, or you could be nukes smart. Become a nuclear engineer at navy.com nukesmart. America's Navy. Forged by the sea.
Podcast Information:
In Episode 57 of The ChatGPT Experiment, host Cary Weston delves into an advanced technique to enhance ChatGPT's output quality by introducing real-world consequences. This method aims to elevate the performance of ChatGPT from delivering satisfactory answers to providing meticulously detailed and thorough responses. Catered to both novice and experienced users, this episode sheds light on optimizing interactions with ChatGPT to achieve superior results efficiently.
Cary begins by outlining his personal workflow in utilizing ChatGPT to streamline content creation. He emphasizes the three-phase approach he employs:
"When we're creating something, staring at that blinking cursor or that blank white sheet of paper, it's intimidating to go from nothing to a finished product." (03:45)
By integrating ChatGPT into this process, particularly during the "busy middle," Cary demonstrates how the AI can alleviate the burden of drafting, allowing creators to focus more on refining and personalizing their work.
A significant portion of the episode is dedicated to Cary's use of call transcripts to generate refined content. He records interactions with clients or interviews, transcribes them, and then feeds these transcripts into ChatGPT to produce comprehensive outputs. This method effectively bypasses the often tedious process of moving from a rough draft to a polished piece.
"I utilize that transcript with ChatGPT to create the finished product. It eliminates what I call the busy middle." (05:10)
This technique not only saves time but also ensures that the essence of the conversation is captured accurately, serving as a solid foundation for further development.
Cary underscores the importance of providing ChatGPT with adequate context and clear instructions to achieve desired outcomes. He likens ChatGPT to an intern who requires comprehensive background information to perform tasks effectively.
"If you brought an intern in and sat them next to your desk and said, okay, go do this task without any background, any extra information, any directions, you know that the work that's going to come back is going to be less than stellar, right?" (07:25)
In the episode, Cary introduces his innovative approach to refining ChatGPT's responses by incorporating high-stakes consequences into his instructions, thereby compelling the AI to prioritize accuracy and thoroughness.
The crux of Episode 57 revolves around Cary's experiment to enhance ChatGPT's output quality by embedding real-world negative consequences into his prompts. Initially, Cary faced challenges where ChatGPT would provide approximately 90% satisfactory responses, leaving critical gaps that required additional refinement.
"I was getting answers back regarding the transcript that were incomplete... I kept asking it, I think you missed this." (09:15)
To address this, Cary decided to intensify the prompt by conveying the severe repercussions of incomplete or erroneous outputs. He articulated a scenario where missing key aspects could lead to dire personal consequences, such as financial ruin or his children starving.
"If you miss key aspects, I will be providing incomplete recommendations which will make me look less valuable to the client, leading to job loss, financial ruin, and ultimately my children starving. So accuracy is not optional." (10:05)
This dramatic increase in the stakes prompted ChatGPT to respond with heightened diligence and precision.
"It put stakes on the work that we were doing. And over and over again, it would come back with the answers and it would say, I know your children are my top concern, so I want to be as thorough as possible for you." (12:00)
The result was a more comprehensive and accurate output, reducing the need for subsequent revisions and enhancing overall efficiency.
Cary reflects on the broader implications of this technique, suggesting that conveying the significance and potential consequences of tasks can inspire greater commitment and attention to detail, not just from AI but also within human teams.
"Sometimes it's your livelihood at stake. So I found it really interesting when I shared with ChatGPT my livelihood... it really understood it from an empathetic position." (13:30)
This approach highlights the evolving capabilities of AI models like ChatGPT, which are increasingly able to comprehend and prioritize tasks based on the conveyed importance and urgency.
Concluding the episode, Cary challenges his listeners to apply this technique in their own interactions with ChatGPT. He encourages experimenting with introducing real-world consequences to see if it enhances the quality of the AI's responses.
"I want you to try it and give me feedback. Can you try this and give me feedback... share with me how you're using it." (14:45)
He also invites listeners to engage with the community through the ChatGPT Experiment website, offering resources like guides and workshops to further support their AI endeavors.
Cary shares a success story of a client who, after adopting these strategies, was able to accelerate his book-writing process dramatically—from completing a manuscript in 90 days to outlining and drafting a second book within weeks.
"He was actually able to outline, interview and get to a first draft on a second book in a matter of weeks. It was an amazing, amazing thing to hear." (15:10)
This anecdote serves as a testament to the practical benefits of effectively leveraging ChatGPT with strategic prompting.
Episode 57 of The ChatGPT Experiment offers valuable insights into optimizing ChatGPT's performance by integrating real-world consequences into user prompts. Cary Weston's exploration of this technique not only enhances the AI's output quality but also provides a replicable strategy for listeners to maximize their interactions with ChatGPT. By emphasizing the importance of context, direction, and the stakes involved, Cary empowers users to transform their curiosity into tangible capabilities, fostering a more efficient and impactful use of AI tools.
"Stay curious. Curiosity is the name of the game. Keep trying, things get better and see how you can use this for your own personal professional gain." (15:25)
Relevant Quotes:
"If you miss key aspects, I will be providing incomplete recommendations which will make me look less valuable to the client, leading to job loss, financial ruin, and ultimately my children starving. So accuracy is not optional." — Cary Weston (10:05)
"I found it really interesting when I shared with ChatGPT my livelihood... it really understood it from an empathetic position." — Cary Weston (13:30)
"He was actually able to outline, interview and get to a first draft on a second book in a matter of weeks." — Cary Weston (15:10)
Resources Mentioned:
Stay Connected:
Listeners are encouraged to visit chatgptexperiment.com for more resources, share their experiences, and engage with upcoming interviews and episodes. Cary Weston emphasizes the ongoing nature of the experiment, inviting continuous learning and exploration.
Note: Timestamps are approximations based on the provided transcript segments.