71 - Tips For Perfecting AI Prompting

Episode Transcript:

Rich: Hey, Caitlin. Are you going

Catelin: to start tap dancing?

Rich: I could start tap dancing. Do you

Catelin: tap dance?

Rich: No, God, no. Um, I mean, I have, I have done dancing before, but I've never tap danced. I've done ballroom dancing. I've done the tango, which isn't that fun.

Catelin: That's ballroom Um,

Rich: yeah, it's part of ballroom. Um, I can, um, What is it?

Rich: Jitterbug. Is that what it is? Oh, I can also do the Charleston. I'm really good at the Charleston, but the Charleston is also like two moves. It's nothing. It's just,

Rich: um, yeah. Cause I was in show choir, you know, uh, don't you know, back in the day, um, that's

Catelin: why I was like, it wouldn't surprise me if you were, uh, if you had a little dancing.

Rich: Oh yeah. I mean, I don't, I don't. Do it anymore, but I did but never ever ballet. No, like none of that kind of stuff or tap dancing um, I could probably fake a tap dance maybe but I

Catelin: don't think you should fake tap dancing that just then you're just stomping

Rich: Kind of you're shuffling and stomping.

Rich: Yeah, so no not gonna dance. Um

Catelin: So, uh, before we get to perfecting our AI prompting, I need to tell you that I was wandering through, uh, Target with my tiny human. She's like a medium human now, uh, but she, they sell tap shoes at Target. And Dorothy was like, I need these! And I was like, I will go actually insane.

Catelin: You may not. I didn't say that. I was like, Oh, we, those don't fit you. Put them back, please.

Rich: No. And then Zach just confessed in the chat that he can swing dance. Um, does Chloe swing dance too or just you?

Catelin: He's going to unmute. Unmute.

Rich: Oh.

Rich: Okay. Did you just grow up learning how to swing dance? I was like, where did you were raised in? You were swing dancers?

Rich: Yeah. So Zach was not raised in a cult. It was just a small town. Oh. That

Catelin: is so cute.

Rich: Did you square dance in high school? Yeah. We did too.

Catelin: My mom was my PE teacher and so she's still like line dancing, square dancing. Carol. I know. Wow. She's at my house as we speak, probably cleaning my garage because she's an actual angel.

Rich: Then I am going to make a suggestion to Carol that you either do like line dancing or square dancing or swing dancing at one of your Camp Carols.

Catelin: Well, Camp Carol is no longer because they moved to Southern Illinois. Oh, that's right. The camp is no longer. But I did see my mom's concession was if they moved off the lake, their next house had to have a pool and they're visiting us right now.

Catelin: And my dad brought the, uh, plans for their pool house and the pool. I mean, yeah, we got to help pick the pool liner. I feel like it's just

Rich: going to be, I think this is good. I think this is going to be Camp Carol's going to come back with the pool.

Catelin: Yeah, absolutely. I think the real kicker will be whether or not people are willing to travel that much farther.

Rich: Oh, that's true, because your whole family is kind of around here. Yeah. Yeah, I mean, it's not that far. People,

Catelin: if, yeah, if Carol invites you, you gotta come.

Rich: Well, she's never invited me, so I don't even think she knows who I am.

Catelin: Do we wanna, oh, she knows.

Rich: Oh, she's listening to the podcast, I think, her and her friends, right?

Rich: Yeah, well,

Catelin: and, and, like, I talk about you, so. Oh, okay.

Rich: Oh, that's very sweet. Yeah. Sometimes Brian's mom asks about you. How's Caitlin doing? I'm like, Caitlin's good.

Lori.

Rich: She just remembered you. What a dolly. Her biggest thing is when I mention somebody from work, she's like, were they at the pool? Cause we did our summer party at my pool two years in a row.

Rich: And I would be like, yes or no. Like they weren't like, it was, it's really funny. So she's like, well, I don't know them then. And I'm like, no, you don't. It's fine. All right. So we, yeah, it is, um, we're going to, and it's also like a really good sign of like, like memory. As you get older, I'm learning, like, as I stumbled across, I think it was last episode, Jeff Goldblum couldn't remember.

Rich: I was just telling Jessica this. Talk to people and talk about people and share information from your near term. It keeps your brain going like you've got to like actually socialize. And I'm like, Oh, the introvert in me says, I don't want to leave the house, but you're supposed to socialize to keep the brain fresh.

Rich: And that's one of the things they always talk about.

Catelin: Driving to lunch today, I told Jessica, I was like, I used to really trust my. my remembrance of things and it's gotten less sturdy as I, as the moments have ticked by. Oh,

Rich: the first time I had to ask somebody if I actually did something or just thought about doing it and it was something in my head that I figured I'd done it already.

Rich: I was like, Oh, this is not good. All right. But, you know, maybe with the right prompting AI can help us remember things. That would be a good use using it for memory, although then it's got all that stuff stored in the cloud. Yeah. So only certain memories, not all of

Catelin: them. That seems, that seems scary. I don't know if I'm ready to give over my memories to the robots yet.

Catelin: Right. And I

Rich: feel like that gets us into like a, what is it? The, is it priority report? Was that the one where they had the pre crime? Minority report? Maybe it was. Yeah, I don't know. I'm not the person to ask that movie. That even when my memory

Catelin: was good, I'm not the person to ask about movies.

Rich: Yeah. They had pre crime where they had like, um, these triplets or twins or whatever that were in some goo and they could actually see what's going on in the world.

Rich: They were connected to everybody. And so if you were having thoughts. Thoughts about committing a crime they would know and alert the authorities and you'd be arrested before you committed the crime. That's dangerous territory because thinking about committing a crime and actually doing it are two very different things.

Rich: So anyway.

Catelin: You can't be prosecuted for the thought of unless it's premeditated and you actually carry it out. Not today, you can't, but apparently, I

Rich: mean, in that movie you could, so don't let the robots get your thoughts, but, um, you can get their thoughts, right? On a hot

Catelin: buttered rum as well. I,

Rich: yeah, you could learn how to make hot buttered rum better, although we did say don't use AI for recipes, so don't do that.

Rich: Um, I do love this one because it is, one, I could eat butter like by the spoonful, so there's that problem that I have. But, um, It's so creamy. It's a little spiced holiday cheer in a cup. It's just warm and like just warm and comfortable and like a big, it's like a holiday beverage hug, Caitlin.

Catelin: Yeah. Yeah.

Catelin: Uh, we had rum batter, which is right. Is that what it's called? Yeah. Rum batter. It's such a weird. Oh my God. So this one

Rich: is overly complex as well.

Catelin: Well, no, uh, cause Tyrell is just making a batter.

Rich: It's not that big of a deal.

Catelin: It's not even a batter. It's like flavored butter is what you're doing. You could use this.

Catelin: So like you could have your hot buttered rum and then you could have like a fresh roll and put some rum batter on your butter. Oh yeah. It's just like Texas roadhouse butter.

Rich: Yeah. It's just butter, sugar, brown sugar, cinnamon. And a little bit of clove and nutmeg. So yeah, it's just flavored butter.

I

Rich: would be 100 percent willing to buy like a tub of like the Kerrygold Irish butter.

Rich: And just like put the little mixer in it to soften it up. Dump all this in and then just keep it going. I think

Catelin: that's the point. Hot buttered rum batter is in a large mixing bowl. Add one pound of softened butter. One pound of brown sugar. One pound of white sugar. One tablespoon of ground cinnamon and one teaspoon each of ground clove and ground nutmeg.

Catelin: Stir to combine. That's a lot of pounds. Yeah. Three pounds of

Rich: stuff. Okay. I need a big bowl. Big bowl. Well,

Catelin: the sugar, the sugar will just like, emulsify and dissolve. But yeah, uh, we had, we had rum butter in our fridge for a minute.

Rich: Oh, get to the next piece. I'm looking ahead.

Catelin: Uh, stir to combine, then add one quart of lightly softened vanilla ice cream and stir again.

Catelin: Transfer to a container with a lid, seal, and store in the freezer. Ours was different. I think I would like this better. Cause ours was literally just like butter and maybe some sugar.

Rich: This has the ice cream. That like, that kind of got me. And definitely don't use like that French vanilla artificial yellow stuff.

Rich: Use the like vanilla bean ice cream, like a real vanilla. I'm

Catelin: the homemade vanilla. Yeah. Blue Bunny is my favorite vanilla. Oh my

Rich: gosh. My brother thanks you for buying Blue Bunny ice cream. Ah, yes. Well, it's poor local business. I know. He's also looking forward to when, um, you and the rest of the crew will buy Asian food from his wife's future.

Rich: Can't wait to get those sliders. I know. There is a rumor that it will be doing a limited run in the church parking lot where they go to church in Lamars before the end of the year, but I have not heard for sure that that's happening. Also, it's getting very cold.

Catelin: And I was like, that seems miserable. Okay.

Catelin: The recipe for this is two ounces of gold rum, a tablespoon of your rum batter, which we have previously reviewed. How many tablespoons

Rich: are in a pound? Cause you got a pound of butter there. And then all the ice cream, a lot.

Catelin: I think you could downsize this to like, you know,

Rich: I like your idea of just putting it on cinnamon rolls and toast.

Rich: Oh, you could put it on like a sweet potato would be real.

Rich: All right. Sorry, I keep interrupting you. So a tablespoon of hot buttered rum batter, two ounces of gold rum. Then what do

Catelin: we do? Boiling water to top and garnish with a nutmeg, freshly grated and a cinnamon stick. And then you just, you know, put your Rum batter into a mug fill it with boiling water and stir

Rich: and your rum rum and batter both go into the

Catelin: mug Oh, yeah.

Catelin: Yep. I'm skipping just to the ice cream

Rich: I know Caitlin's like I'm just gonna eat this in a bowl. Could

Catelin: you make could you make this into like a milkshake?

Rich: I don't know why you couldn't

Catelin: you could do like I mean, I feel like a mudslide scenario

Rich: I feel like if you just put milk or half and half in it, in the blender, you would be there or a little bit more ice cream.

Catelin: I was going to say just the ice cream.

Rich: I feel like it is a milkshake. You just need more ice cream and yeah. More rum. All right. I may have to get out the blender and, um, I just feel like if I, I feel like I need to have rum batter in the freezer at all times. Okay.

Catelin: Yeah.

Rich: All right. Okay.

Catelin: I think I'm going to send this to Tyrell and tell him he needs to try again.

Rich: Okay. I think that's good. And you can report back in the new year. Okay.

Catelin: All

Rich: right. So should we take a break and, uh,

Catelin: go find some ice cream? Absolutely. No, we're going

Rich: to do AI prompting, AI prompting. Then you can have ice cream.

Catelin: I would like AI to promptly get me some fucking ice cream.

Rich: All right. We are back and ready to prompt some AI, yo.

Catelin: I still don't have any ice cream, but I'm going to let it slide this time.

Rich: You can get some when you're done here. Although you're in the office, we don't have ice cream in the office. Do we, do we?

Catelin: There's popsicles in the freezer.

Rich: Okay. You could maybe substitute a popsicle.

Rich: I mean, I would text Tyrell and tell him that you want the rum batter. Like when you get home.

Catelin: Chop, chop, sir. As if he's not also at work. And, uh, making significantly more money than I am. I

Rich: feel like Carol could make the rum batter if she's at your house right now. This is, it's just like baking. She, like I

Catelin: said, she's cleaning my garage, so I'm not going to interrupt that flow.

Catelin: Wow. Is she, oh, she's an organizer, isn't she? She's an organizer and, um, she can't sit still. If you need to. So my dad is finishing, uh, a project for us inside of our house. And she's just like scurrying about. And Being spectacular. I cleaned out so much stuff this weekend. It was amazing.

Rich: It doesn't feel good.

Rich: It felt

Catelin: so good.

Rich: If she needs something else to do or you need to get rid of her for a little while, our basement could use an organizer and so could my office. Like I have moved like behind me. I showed you like it's all neater, but it's all just over here on the other side.

Yeah.

Rich: All right. And that's something you can't help me with.

Rich: I know like there's plants.

Catelin: This is what we would tell an organizer though. Be detailed and specific.

Rich: Right. Which is your first, um, tip for prompting AI. Um, and I think that what's interesting is think about what you really want and then put in like a really vague prompt. Um, I mean, it's right up there with like, make me a picture of a tree.

Rich: Okay. There's so many trees. What style of tree do you want? Um, so, and same thing with, You know, how do I, whatever, um, you've got to be specific. So, um, what's really interesting. We have several people taking an AI prompt in class right now. Um, are you one of those?

I am.

Rich: Oh, and one of the things that I know that came up is like, there are some AI prompts that are like pages long, like getting very, very, very specific.

Rich: Um, and it's interesting to see, like, the more detailed and specific you get, the more likely you are to get an output that fits what you want.

Catelin: Yeah. I think the most interesting thing that I've learned from that so far, because I am unshockingly a little bit behind the like predicted or projected like course timeline or whatever.

Catelin: I got kicked out of

Rich: the course because I didn't start it fast enough. So you can rejoin at any time.

Catelin: Rich! I know,

Rich: I'm terrible.

Catelin: It's all right. Uh, but it was like act as. And so that's, um, been an interesting, like, colorization where it's like, Okay, pretend that you are a 37 year old woman looking for a new And here are the considerations or here are, you know, the health things that I have, like, what should I be asking?

Catelin: What should I be looking for? It's, it is fascinating to me, the speed at which processing can happen, because it's like, these are a lot of these is like things that I could do. But I don't have the processing capability to do it at scale and at a speed. That is efficient. And so it's like, help me be more efficient and, uh, process quickly.

Rich: Yeah, I, I agree. Um, it's really wild. Like, you know, saying something like, um, how can I improve my marketing? Like that's again, and honestly ask you get agency that going back to our agency tips episode, like that's not going to get you what you want, but what are three effective content marketing strategies for small businesses targeting a millennial audience?

Catelin: Right.

Rich: That. So

Catelin: much more effective.

Rich: I mean, and honestly, for me, if I want three, I ask it for five, you know, if I want 10, I ask it for 15. I want to see, cause there's going to be some stuff in there that isn't working, um, or isn't going to work for you. Um,

yeah.

Rich: All right. Oops. I'm just like typing the wrong thing in the wrong window and I'll just stop now.

Rich: Um, and I think that like, and sometimes you can give it more information to be detailed, like give it a URL of a page. Yes. You can upload an article and ask it to summarize.

Catelin: Yeah,

Rich: so those are all like the more you put into it the more you're gonna get out of it The more specific you are the better I

Catelin: think the other this is like the other kind of Hack for AI that I have found and this actually comes from my husband where he's like we when you read GPT specifically, it has a tendency to be, um, more like academic or, um, like a higher comprehension level.

Catelin: And so he's like, when we, when I'm writing for certain things, I'll just say like, write it at an eighth grade level, which I find really interesting. And, um, Um, yeah, I did that to like lower the, lower the bar a little bit and make it sound more human.

Rich: Yep. Telling it what education level your audience has, you know, most people will know that.

Rich: Um, I do think that like getting it to write at X grade level, that's how like you're taught when you're in a writing class as well, like certain things to write at certain levels. Um, and you can actually, if you forget it on the first draft, most AI will, AI bots will allow you to revise. Oh yeah. Revise this at an eighth grade level, that type of thing, or, um, make this more informal.

Rich: So I think the other thing, and what I like about what's coming out with some of the tools like HubSpot's Breeze AI, um, I know that there's a couple of others that have the same thing. You can put your brand voice in. With HubSpot, what's wild. I did our brand voice, um, cause we have content hub and can do content remix and all that, even though we don't have the social bot yet or social agent.

Rich: Yeah. Um, I basically just told it to read our website and tell me what it thought our voice was. What did it come back with? Uh, it came back pretty accurate, actually. I don't have it in front of me. Um, but I was really surprised to

Catelin: me. Yeah.

Yeah.

Rich: And I did the same for a client and I sent it to them and they're like, yeah, that's pretty close.

Rich: I would just also like, I want something in there about being like a little irreverent or a little edgy. And I'm like, okay. I'm like, well, just so you know, your current website isn't edgy or irreverent or it would have come back with that because those are some of the prompts, some of the things that it has.

Rich: Yeah. Yeah. I've also seen some tools that just have like six things for voice, like friendly, open, formal, um, HubSpot has like 20 or 30 in there, um, which is really nice. Um, but you can also. If you don't have that though, you've got to put that in your prompt. So anything that it's not already tapping into, it needs to go in your prompt.

Rich: If you're in chat GPT, you've got to tell it your voice. You've got to tell it your brand style, those types of things.

Catelin: Which is actually a really good, um, kind of lead into providing examples of what you want. So you could say like, read this blog post, provide a link, and then using that voice, Make me the next thing, right?

Rich: The other one is if you're doing images the style of the image and this one killed me because I got to one where they Had a drop down of all the styles and I was like, holy crap. So there's like realistic there's photo. There's Yeah film photo, there's, you know, um, there's anime illustrated, like, there's like, there was like 20 or 30 different, different ones.

Rich: So if you're getting stuff back and you're like, this isn't what I want. Look at style, style and voice are typically things that will hit you. Yeah.

Yeah.

Rich: So the other thing you can do is give it an example. Um, if you have a case study that you want to do on a different topic, you can give it the current case study and tell it to write you something in a similar format.

Rich: About a different topic and let it know. Um, you can also paste a lot of information in there. Um, I think that's a good one for summarize as well. Um, I actually

Catelin: use that for a volunteer thing. I had like a 900 page document that I needed a summary of. And I said, can you tell me, it was a transcript. I was like, tell me who spoke, uh, tell me the topic that they talked about.

Catelin: Tell me the broad themes and tell me the outcome. And it came back. I mean, it probably took like a minute to just like ingest this 900 page PDF and spit out like main themes. It was amazing and super helpful. Yeah.

Rich: Yeah. And I've, um, For large documents. Like I always joke that if I was like a legislator, I would use it to summarize bills all the time.

Rich: Yeah. Because I mean, you try to read them, but like, you hear all those things about like, Oh, they gave us this 900 page bill an hour before we need to, um, actually vote on it. Well, guess who can summarize that thing in about five minutes. Um, so, um, that's a really good one. So, um, if you have an example, give it an example.

Rich: And most of the time you will. Um, you can also give it an example of a webpage. You can just give it a link to an example. Um, like, you know, write product information for this new product here. It's features formatted like this page where I've got another product. Um, so that it gives you stuff that will populate, um, in, you know, what I haven't thought of Caitlin

Catelin: tell me.

Rich: If we created a web page with Greek texts, like in the placeholder, if it would write on the topics we needed to fill that space, like, I'll bet it would count the characters or lines. Oh, I tell, I tell, I

Catelin: tell GPT that all the time. Write me three sentences on, write me, you know, make it shorter, make it longer.

Catelin: Like,

Rich: yeah, I was playing with the, um, using, doing a virtual podcast, like writing the whole thing, but then having an AI voice voice it, um, just to see how you could do that. And so I had to figure out like, okay, if I was doing like a 15 minute, like Barbara Corcoran says like 15 minutes, it's very fast. She does three questions.

Rich: Um, and it's nice when you're in the car for a short trip. I was like, so I'm like, how many words is that? So I had to go like, Look at how many, it's about 2, 500 words. So it's like, write me something, uh, 2, 500 words on this topic, do an intro for a podcast on it, do a close for a podcast and announce the next, next topic is X, Y, Z, and it came back with a lot.

Rich: And then I was like, okay, break it into three main topics. And so then it does that. And so then you basically got. Three, four minute topics and intro and an outro and boom, it's a 15 minute podcast episode. And then you give it to an AI voice that's hopefully realistic and they read it and you could upload.

Rich: But then I'm like, Oh, the ethics of this make me just squirm. And

Catelin: that's been my whole kind of touchiness about it from the get go. I mean, we talked. On our last episode about image generation and what does that look like? And are we protecting our intellectual property? And, um, I'm still cognizant and wary of that.

Catelin: But the, the, the thing that has kind of assuaged my nervousness is just the, like the input, like the human input and the speed at which things process that like I could get to the same result. It just would take me so much longer.

Rich: Yeah. I'm a hundred percent comfortable with like researching topics. I'm a hundred percent comfortable with like having it summarize a blog post and do my meta description for me.

Rich: Yeah, totally. Totally. Totally. Recommend keywords, things like that, that I know that there's data out there that it's getting and coming back with. Um, I think I'm still struggling with, okay. Even if it wrote the whole like 15 minute episode and I read it and edited it so that it was, you know, cause I think that you do need that piece there.

Rich: Um, I'd probably send it to a proofreader also just, that'd be interesting. It does AI. How good is AI? How good is AI's grammar? Um, but is that

Catelin: an M dash or an N dash AI? You better figure it out.

Rich: Like, could you actually, like, could you actually do a, I mean, could you do a daily podcast for 15 minutes about a topic like, and.

Rich: Is that valid information? Is that good information? Is that something that should be done? So that gets into the ethics of AI, which we probably don't have time for today.

Um,

Rich: but basically, um, coming back around, like given an example of what you want it to be, um, there are also things with AI where, um, I know that content remix, um, I'd mentioned that in HubSpot.

Rich: Um, they have a, uh, case study tool where you can upload like a video interview you did with a customer or a recording of an audio that you did with a customer where you're just interviewing them, you can send it to a webpage, you can give it stats, results, whatever, and it will take all of that and format it into a case study.

Catelin: That's fascinating.

Rich: Um, pulls out key stats, pulls out a quote from your customer, from the video or whatever. Um, Um, it's pretty interesting, um, to see how that is. It's limited now, but I do think that it's just really going to take off in 2025.

Catelin: Yeah. Yeah. Yeah.

Rich: Yeah.

Catelin: I think this is my other kind of favorite or like refinement tool is asking follow up questions.

Rich: Well, you, you do that anyway. Tell me more. Tell me more about that Chet beat GPT.

Catelin: Did I tell you that Dorothy used it against me? Yes, not against me, but like like she repeated it back to me She said tell me more about that and I was like what? Is happening? I thought I had more time before she turned into actually tiny me.

Rich: No you you birthed the time You're gonna deal with the consequences. I was like, i'm

Catelin: so amazed and also so proud and then like Yeah, still just gobsmacked. Wild. That'll

Rich: be fun. Like if your mom and dad noticed that, that like, they're like, um, Caitlin, she is really lovely. I'm sure they do.

Catelin: We were with Tyrell's family over the weekend and they said that, uh, cause he was a chatterbox.

Catelin: And one of his uncles was like, man, she just doesn't stop. Does she? And we're like, no, she doesn't. We're very tired. And he was like, yeah, I mean, you earned that every bit of it. So anyway, the point is ask follow up questions.

Rich: Well, and I think that the interesting thing is, um, it knows context, right? So.

Rich: Years ago, it was like at least five years ago, I think Google started doing context. No, it was, yeah, it was, Oh God, it was more than that. Cause it was when Obama was president. Um, because that was the example they used. It's been

Catelin: 84

Rich: years. You can ask a question like, you know, who's the president of the United States of, you know, Google or even, uh, uh, I don't want to say her name, but the one from Amazon, uh, cause there's one, she'll turn on.

Rich: I don't need to hear that. Um, but you can ask a question and then you can follow up contextually, like who's the president United States. And you can follow up after it tells you with who's his wife. And it knows the, his just like we do is referring back to the president. Um, so these AI chatbots are. an expansion of that.

Rich: And they do get context. So you don't have to repeat your whole prompt. You can revise as you go. Um, but, um, same thing is like I did one because I was just feeling especially spicy. I asked, 10 blog topics, um, on a specific thing. Um, and then I was like, great, right. A 300 word blog post for each topic, which is not the way to do it.

Rich: Cause it was just this giant mess, but it did it. I mean, but then I would have to copy and paste. Whereas if I told it to like write a post on topic one, it would do it. And then I could automatically paste it into a blog post. Um, but. Yeah, like, and I think, go ahead.

Catelin: The contextual nature and, and my experience is really with like chat GPT is like each of those little windows, right?

Catelin: Is a conversation. And so you're using the previous information and this is what makes it generative, right? Is like, it's using the context. that you've already provided to give you the answer. And as you ask those follow up questions or add additional information or context, it's adding to that historical knowledge, if you will, or information, and then can come up with a more.

Catelin: potent informational Specific response for you.

Rich: Yeah, and you can get more specific. I think price is a good one Especially if you're using it for like a shopping bot It'll give learn that thing and you can say please refine to show me only things like under 100 or whatever it is Yeah Or with our, you know, what content marketing strategies for small business, you could have it refined those to be, you know, you could ask it, you know, what would it cost me to use an agency for this, which of these are under a thousand dollars a month.

Rich: If I have limited resources, what's the best way to implement these? Um, you can also sometimes dovetail off of an idea so you can ask it. You can say, Hey, take the second paragraph and expand that into 10 bullet points for me. Yeah. Um, so there's all kinds of things you can do when you're asking the follow up questions.

Rich: It is my favorite. Um, and you just ended up this like big, long conversation, right? Yeah. Yep. Yeah, eventually it will be, they'll have like a voice chat thing and you'll just don't, I don't want you to read it to me and it'll talk back to you. No, um, we'll be just doing blog posts from the car and never actually reading them.

Rich: We'll just be hearing

Catelin: them. I have started to use voice memos more specifically, or like voice to text. I think those have also gotten stronger, but, uh,

Rich: I do voice to text as well. Yeah. It's gotten much, much better. I think the, um, yeah. The Apple intelligence is fantastic. I've been playing with it on the iPhone with Siri and it's been doing a much, much better job.

Catelin: That's interesting.

Rich: Yeah. Yeah. It summarizes text messages for you. So my brother sent a really long one the other day and it summarized it in like eight words. And I said it back to him. I'm like, this is how it summarized. And he's like, that's pretty perfect. And I'm like, yeah, so you're just overly wordy.

Rich: Like you can talk less.

Catelin: Verbose. That's me.

Rich: All right. So I feel like, um. Yeah. We're good. Um, so hot buttered rum. These are some tips for perfecting AI prompting that we gave you. So detailed and specific, the more detailed, the better. Um, if you have a two pages of detail, you want to give it, give it two pages of detail.

Rich: Like that's okay. Um, give us some examples. Examples are beautiful. Ask your follow up questions. Um, that's really where the magic is. Like sometimes I'll start with a really basic prompt and then I'll just continue to refine and go deeper, um, to get closer to what I want.

Catelin: Well, it's also like the, the idea of, I know what I don't want, or it's like, and the first answer is not usually what I want.

Catelin: So it's like, okay, you're, you're close in this part, but tell me. More here or don't really about this

Rich: a

Catelin: exactly exactly

Rich: You can actually in your initial prompt give it negative prompting as well. I

Catelin: remember this. Yeah Don't use for this

Rich: don't use this Like you could even tell it things like swear words are okay.

Rich: Or don't use, um, you know, don't use overly flowery language, you know, that kind of thing, be as concise as possible. Like whatever you want to tell it, um, you can, you're putting those rails on. Right. You're quite literally putting it. In a box of what it's able to do and not able to do. Yeah. Um, and then you can always expand out if like it hits on something that like, and you're like, Oh, I want more of that.

Rich: You can ask, or I

Catelin: didn't think about that thing initially. Yeah. Or

Rich: I didn't think that would be in there. I definitely don't want that in there to

Catelin: get out. Yeah. Uh, as always, you can find our agency at antidote underscore 71. And if you have a question you'd like to send our way, you can visit ctapodcast.

Catelin: live, or if you would like to leave us a message and get a cocktail book, hot buttered rum won't be in there. Others are. You can leave us a voice message on our hotline at 402 718 9999. Seven, one and your question will make it into a future episode. We'd love to hear from you

Rich: 100%. And, um, if you leave us your address, we will not dox you, I promise

Rich: Um, we don't do that here, but we will send you a lovely book. I was actually just jumping in to see it's, see if we had any voicemails at all in that inbox. Do, do, do podcast, uh, 10. Well, we have 15, but they're all, um, unknown and empty. So it's, it's actually no voicemails. It's just calls received. Don't be

Catelin: shy.

Catelin: Leave us a message.

Rich: Um, yeah, I mean, I could put out the phone numbers that these are from, but, um,

Catelin: nobody left a message, which you said you weren't going to dox people.

Rich: I won't. Um, But I will have to go in and just double check and make sure maybe I need to call it and make sure that it's recording a voicemail because it should be.

Rich: I tested it when we set it up. But anyway, Zach, cut all of

Catelin: this out. Cut all of that out.

Rich: I think it's fine for people to hear. So we have, uh, another episode coming next week. Um, we are getting back into some, uh, Lesser holiday drinks and more into some fun stuff. This one will be interesting. It's the Italian espresso martini What makes it Italian you say we will find out Um, because we will not, um, be discussing it here.

Rich: Um, and our topic will be understanding the nuances of B2B and B2C marketing. So business to business versus business to consumer marketing. So that'll be great with an Italian espresso martini. I am looking forward to that one and finding out what exactly makes it Italian instead of just espresso. It's the nuance.

Rich: See what I did there? It is the nuance. All right. Um. See you next time. I think that does it for us.

Catelin: Okay. We love you. Bye.