Jordan [00:00:17]:
Can you use ChatGPT as a second or a third employee? Well, stick around, we're gonna talk about it. My name is Jordan Wilson, I'm the host of Everyday AI. This is your daily podcast, live stream and newsletter. We have 1 of those 2. Talking about the latest in AI developments, using tools like we're going to talk about today, like ChatGPT, Mid Journey for Images, everything. Keeping up with what's going on in the world. Helping us do that today is Sandra West. Sandra is the CEO of Bellastrega. Sandra, thanks for joining us.
Sandra [00:00:55]:
Hi, glad to be here.
Jordan [00:00:58]:
All right, let's get to it. So as a reminder, If you are watching us live, please leave a comment for Sandra or myself. Like I said, we're gonna be talking today about how you can use complex ChatGPT prompts for simple business solutions.
Apple releases new XR headset
But before we get into that, let's talk about what's going on in the world of AI news. It was an exciting one yesterday, Sandra. So I know a lot of people were probably seeing this. It was all over the news. Apple finally, after many years of rumors, they finally released their new AR, VR headset, technically XR is what they're calling it. So the Vision, it's called the Vision Pro.
So it looks cool, but what's important to note is, right before the announcement, Apple stock had just reached an all time high, probably in anticipation. And then in the minutes and hours following, once they finally unveiled it, it started to dip pretty drastically too. So Sandra, aside from the stock news announcement, what's your takeaway from this new Apple Vision Pro headset device?
Sandra [00:02:04]:
Well, I didn't get a chance to actually see any of the news release on what the features are going to be. But I know that, you know, people that are Apple fans are pretty diehard fans. So I assume that there will be the normal lining up to make sure that they get to be the first person and announce on social media that they've acquired 1. And it's a really tight knit kind of cultish community.
And I'll admit I'm a recovering Apple cultist. I just started comparing, you know, what the actual parameters and qualifications are of products a few years ago and making some changes. But I assume it's gonna make a really big splash. And I'm actually a little surprised about the stock drop. Like I said, I didn't get to see anything about the specs, but I would say that either means that it was a little disappointing and not as impressive as people expected it to be, or the price point is outside of the range that people were expecting.
Jordan [00:02:59]:
Yeah, That $3,500 price point probably took a lot of people by surprise. So we'll have more about that in the newsletter. But let's move on because I think that Europe is doing something that a lot of people here in the United States have talked about.
But a recent news story today talking about how Europe is moving closer to labeling all AI content to essentially tell people on the other end consuming the content. Is this made by a human or is this AI generated? Sandra, what are your thoughts on this? Is this something that you think could help, number 1, could it help misinformation and disinformation? And do you think it would ever, you know, the majority of our listeners here in the US, do you think this would ever happen here in the US?
Sandra [00:03:44]:
I think that it's eventually going to happen everywhere because I don't think AI is going away and I think it's gonna be just pervasive. But I will say that when it comes to digital regulation, we've notoriously been behind in the US. We're constantly playing catch up on how to even handle legal matters that involve anything digital, anything that's happened online. And so I think that it'll be a slower rollout.
I think that it'll probably have to be handled by case law and precedent where there's been some massive issue that someone's able to take into court, because that's just how things usually make it through the system here. But I do feel like it's necessary and I feel like it will be helpful. I think it's really 1 of the things that will start to alleviate some of the major concerns about AI, both for people that are celebrities and concerned about their content, or people that are concerned about their jobs and whether there's just going to be this rampant use of raw AI material instead of using any skill to actually design or write something.
Europe may pass regulations impacting US technology
Jordan [00:04:40]:
Yeah, it should be interesting to see and especially in this use case, you know, when it comes to technology, normally, you know, the US is kind of ahead in creating those boundaries or those regulations. But this time it is Europe. So it's going to be interesting to see, you know, what they finally pass and how it affects things here in the US. So speaking of things here in the US, ChatGPT has obviously taken the, probably the world by storm, I'd say, at least if you are living and working in front of a computer like many of us.
But a new story today in the Washington Post that we'll talk more about in our daily newsletter is talking about how there's kind of a problem that no one's talking about with all of these AI chats. So your open AI's, ChatGPT, Google BARD, Microsoft Bing chat. The problem is every time we use them, the company supplying them is losing money. So without getting into too many of the details, these chats require a lot of computing power running on some very expensive GPU chips and you need a lot of them. So Sandra, you know, ask someone that's using, you know ChatGPT on an ongoing basis at Bellastrega. Like, does this concern you at all? Are you, or is this something more of like, Hey the companies know there's so much money in this. So they don't really care right now if they're losing money.
Sandra [00:06:02]:
But I think that's actually a question that I can answer from a marketing standpoint. I think that when you're launching something new that is gonna be something you need early adopters for, launching it for free is always the best answer. And it has certainly created a stir and a massive following. So they're well on their way to being in a position where they can start to say that they're going to limit or remove free options. And I think that that was probably the eventual plan all along.
It's just not sustainable for the full plans to be free and people to have access to everything, particularly with the amount of attention it's generated and people using it just to play and test all day long. So I think that that will settle down some, the volume will, but I think that they're going to have to move to, you know, paid tiered plans once it's become just commonplace.
Jordan [00:06:52]:
Yeah, Sandra, that's, that's a great, a great response. Yeah. A lot of times in marketing, you need to bring as many people as possible to your platform. So you really have to gather and garner that attention, bring it to your platform and then market and start to charge for it. So, I'm interested. So tell me a little bit about what you're doing at Bellastrega, you know, what your team's working on, and also just, you know, let's start to talk about how you're even using ChatGPT on a, you know, daily or weekly basis.
How Sandra Uses ChatGPT
Sandra [00:07:26]:
Absolutely. So we do business design and brand management. So we can basically design and build anything for the online footprint from websites, graphics, social media, the whole 9 yards, and then manage it as well. So we do ongoing content, advertising, social media, web services for all of our clients. So we have lots of use cases for bringing in the AI to be able to kind of extend what we can do. Really just adds bandwidth to the whole team. So let's say for branding, we're working with a new client and we wanna get some quick ideation done, put some things in front of the client, kind of brainstorm together.
We can use it for things like doing some logo concepts just to get a rough idea of what we're aiming for. We can use it for producing a whole list of possible taglines so that we can start to see what resonates with someone. We use it for social media, often within very specific parameters of the audience and voice, et cetera, to produce content ideas and just give us a list of social post topics. So we really use it for just a lot of ideation and that initial phase of working into something just to give us those ideas and some guidelines to work with and it's also amazing for market research. So it can point us in the right direction is something that I would usually have a team member spending hours on for market research. We can have lists of top companies, top influencers, like really fast.
Jordan [00:08:57]:
So that's super helpful too. Yeah, it's and and I do want to get back to kind of how we started the show about, Hey, is this almost like having a second and third employee? But before I get there, you know, I do want to get your reaction. So kind of when, um, ChatGPT first announced this, um, this, this kind of newer technology, Obviously the GPT, the technology that it's based on had been out for a year or 2 and a lot of marketing agencies were using it before then. But kind of when ChatGPT was announced, I think in late November, early December of this past year. What was your reaction? As someone who's working in marketing and branding every day?
Sandra [00:09:41]:
I'll admit I had a little bit of a mixed reaction. I immediately recognized that it could be a tool, but I also recognized that this could be an industry issue. And it's for me just kind of triggered some memories of when the freelance stuff kicked off really huge a few years back. And we had just kind of the market flooded by people who were claiming they could freelance and they weren't really sure what they were doing. And it just caused a huge issue for those of us who were already doing contract work and had been for years.
Um, I kind of had like a flashback to that moment. We're going to have another, uh, you know, resurgence of people who are using AI to just spit out things that they can sell to people and it's going to cause kind of a market disruption for those of us who are in these industries that are going to be heavily affected by AI.
Jordan [00:10:28]:
You know, so once I guess. Do you still have those feelings Or do you think now after it's been out for, you know, 5, 6 months, do you think that's not gonna be the case?
AI tools: great but needs editing and branding
Sandra [00:10:42]:
I, it's definitely abated my fears at this point. I think that it is a tool, but I've tried just playing with even amazing prompts. And what I get as an end result, if I ask for a blog article with an extensive detailed prompt is not something I would publish. It's always something that needs to be heavily edited, and it needs to be brought in line properly with a company's branding and tone of voice. And so for me, I don't see most of the things that are being produced as being an appropriate end product.
I think there's some exception there for some of the images, but there are also some limitations for that as well. So for me, I've kind of moved into this is a great tool, and it's 1 of those things that you need to learn and harness to be able to stay ahead. I don't think it's an option at this point to say, like, I refuse to use AI on principle. I feel like that would cause anyone to fall behind because it's not going away. Yeah, that's that's a great question. I think, you know, especially early on, you know, and when AI development is tied to job loss, specifically,
Jordan [00:11:50]:
you know, people started to view, whether it's ChatGPT or just AI in general, as a villain, and not someone that can be on your team, not someone that can be an unsung hero on your team. So, you know, Sandra, getting back to how we opened the show, can, you know, using ChachiPT as a researcher, as, you know, an ideation partner, can it actually be like a second or third employee to, you know, after yourself?
Sandra [00:12:17]:
I would say in some respects. So for me, what it means is whichever team member I have doing market research would now be able to complete the market research foundations that we start with for 2 to 3 clients in the time that we usually would do it for 1. So we have a spreadsheet for market research that we fill out, and it just increases how quickly they can get those answers and get them into a spreadsheet for me to come review them and say, okay, here's the direction we're going.
So in that respect, then yes, I would say it's like having 2 or 3 people on the research team instead of 1. And same with our social media team. They're able to create content much faster because they're able to generate these ideas, decide which ones they think are best, send them over to me for approval and then get those posts drafted a lot faster than they would if they were spending hours looking at competitors' social media pages to come up with the best post ideas. They're able to just generate them and say, these look great, Let's go with this.
Jordan [00:13:16]:
Yeah, yeah, that's a that's a great question. Mohammed, thank you for the thank you for the compliment saying Sanchez doing a great job, you know, diving into everything here. If you do have a question, please, please feel free to to drop it in the show here. So, Sandra, I have a question. So when we talk about prompts, right, it seems like there's so much misconception and people always saying like, oh, you know, these 15 prompts are going to change, change your life, change your business, make, make you a millionaire overnight. What are some, some best practices through your experience that you can give to people who are maybe newer to working with GPT? What are those 2 or 3 things that you can say, hey, do this, and you're going to get a lot better result from your prompts?
Specific parameters are crucial when creating prompts
Sandra [00:14:01]:
So a couple of things. I've tried out some of these, like the 1 prompt to rule them all formats. And some of them are kind of cool, but they're a little overcomplicated. I think the most important thing is to make sure that you are providing specifics and that you're providing some sort of context and parameters for what you're trying to get. So, for example, if you were to say, write a blog article on spring fashion for 2023, you would get an article, but it's not going to be tailored to anything specific. If you go in and you tell ChatGPT to act as a fashion influencer in a particular niche and then you say with an audience of and you give the parameters for your audience, and then you ask for a blog article on the Spring Fashion and you give them a length and a tone. If you provide all of those parameters, then you're going to get a much better result.
So that's really the basis of creating prompts. They can be done a lot of different ways. I've seen them done where they look like complex coding. I've seen them done really simplistically. And to be honest, you'll get a result if you just type it in as a paragraph, like you were talking to someone. It may not be as good as the more coded result, but you're going to get something that's closer than if you just said, write me a blog article.
Jordan [00:15:23]:
Yeah, exactly. Actually, our guest from yesterday, G, saying, tell Jett TPT, please and thank you. In your experience, I've actually tried this. Sandra, have you tried basic manners? Do you think it helps at all?
Sandra [00:15:39]:
I think that wording the prompts in a manner that makes them really clear, kind of that clear as kind mentality is what seems to be the best rule for me. So for example, if you give it a prompt and you get something that's way longer than what you wanted because you didn't give it a specific length, then instead of saying, make that shorter, if you say like, please condense this, then that seems to get a really good result. So I think that I do kind of by default put in some manners in the way that I do my prompts, especially when I'm progressively adjusting something. Yeah, that's that's great. Yeah, kindness clear. I love that. And Gee, thanks for the comment. So a good question. I'll let you feel this 1, Sanjur. This is a tough 1 from Arturo.
Jordan [00:16:25]:
So Arturo is asking, if you generate a piece of text and then you have to edit it, isn't that double work? Yeah, that's a great question. Yeah, if you're using these tools and if you have to put so much on the front end and then if you have to edit it or, you know re-prompt it on the backend, is it worth it?
Sandra [00:16:43]:
So I would say, yes, that is some double work But just to add some context there, so we don't ever use AI-created full content. We use it for ideation, creating lists. We've tested out doing some blog articles, and I really felt like they needed a lot of editing. But what I will say is for someone who's working on a low budget, who maybe has a team of freelancers who are producing something that they still have to edit, then I would say if the person that is writing your blog articles sends you things that you then have to edit, then doing it in ChatGPT and then having to edit it wouldn't be double work. Yeah, that's a great point.
Jordan [00:17:23]:
This 1, I'm very excited for this question. PJ, thank you. You're stopping by every week with great questions. So PJ is asking, Sandra, what are your top 3 AI tools beyond ChatGPT?
Tools used: BARD, MidJourney, WriteSonic. Beneficial for ideation, research, and teamwork
Sandra [00:17:36]:
Absolutely. So I do like BARD. I think that it has some specific use cases, and I really do like it for doing the things that are more research-related. I also love MidJourney. We haven't actually put any of the images into play yet, but just being a designer myself, I absolutely love using it for ideation. And for my team, I actually use WriteSonic because WriteSonic combines the GPT technology with access to the current internet technologies. They're able to actually access Google and get you current answers on something that may be a news story from yesterday.
And it also works well for my team because it has a breakdown of all of the different formats. So basically some of the prompting is built in. You can go straight in and make a Twitter thread or you can go straight in and make a LinkedIn post and it can actually format everything properly so that it comes out the way that you would want it to be for a post. And So those are really great tools when you're working with a team. Yeah, great, great. Yeah, I use all 3 of those as well.
Jordan [00:18:38]:
Fantastic. I love that those are your top 3 because they probably would have made my top 5 or top 10 as well. So As a reminder, if you're listening to this and you can't keep notes fast enough for all these great recommendations that Sandra's talking about, make sure to go to youreverydayai.com, sign up for the newsletter we have. We'll be sending out links to all these things that she's referencing. So 1 other question, kind of going back to what Arturo was talking about, like doing this double work.
Sandra, what's some best practices maybe for getting more or the most out of your prompts? You know, are you able to tweak them and use them for different audiences. So kind of after you put so much work into something and you've tweaked it and you have it just right, how can you reuse that for other purposes?
Sandra [00:19:26]:
I can give you a perfect answer because I have a document saved on my desktop with the prompts that have worked the best for me. And they allow me to just go through and very quickly fill out the act as the audience parameters, the link or the context. And it's really just kind of fill in the blank. So I copy it from my document on my desktop. I drop it into the program that I'm using. I add all the context to it and what I actually want to have the output look like. And it works pretty quickly. And then I'm able to kind of do some follow up prompts if I need to make it look a little different or if I wanted to continue generating more content beyond that. Yeah, that's great.
Jordan [00:20:08]:
So we'll try to tackle this 1, Sandra, if you can. If not, I might have something. So PJ asking, what do you recommend for video creation using avatars? It's a specific 1.
Sandra [00:20:19]:
So it's a very specific 1 and I actually have just started looking into that. So I don't have a recommendation yet. I honestly have not gotten to test out very many of the platforms to be able to give a recommendation on what I feel is best.
AI tools evolving fast, software DID leading
Jordan [00:20:32]:
Yeah, so PJ, there's a lot of newer tools. These AI tools literally are coming out hundreds, sometimes thousands a day. Don't go on product hunt, you'll get lost at looking at all these new possibilities. But I'd say PJ, the 1 that's kind of in the lead, I think, is a software called DID. So that is literally taking, creating an avatar, a custom avatar, and then you can have it match your, the words on the screen, and it's pretty incredible. So we'll make sure to link that in the newsletter.
So Sandra, I think we have one last question here. So from, I think Noelle, sorry, Noelle, if I didn't get your name right, but she's just saying, true, love this about using ChatGPT with the prompts. I think we're using them. So I'll end with this, Sandra, what would you say is your 1 most used prompt that you're, you know, whether it's on an everyday or every project, what's that go-to that you just keep using over and over again?
Sandra [00:21:32]:
It's a really simple prompt that just puts in the act as the audience and then the output. So that's my simple go to is just act as this expert for an audience of whoever your audience is in a concise manner and create or write this thing is the very the simplest prompt that you can use. That's going to give you a result that's really close to what you need.
Jordan [00:22:02]:
That's great. Yeah, I think definitely breaking it into pieces and the act as is great. So, kind of internally, internally, that's just kind of what we say. We say prime, prompt and polish. That's just our kind of internal piece. FYI, if you do sign up for the newsletter today, hit me up, tell me you want to learn more about that. I'll send you some information. So, uh, Sandra, you literally in a short 20 ish minute period here, you unleash so many great topics, so much great advice. Uh, Thank you so much for coming on the show and sharing your expertise with everyone.
Sandra [00:22:37]:
Thanks for having me. It's been fun.
Jordan [00:22:39]:
All right. So as a reminder, if you are listening to this on the podcast later in the day, please join us. We do this every weekday, Monday through Friday, 7.30am Central Standard Time on LinkedIn and other platforms, I think. So drop us a comment, participate, listen on the podcast, Spotify, Apple, all that stuff. And go to youreverydayai.com, sign up for the newsletter. We're going to have links to all of these different things that Sandra was talking about. So we hope to see you back tomorrow and every day on everyday AI. Thanks.
Sandra [00:23:13]:
And that's a wrap for today's edition of Everyday AI. Thanks for joining us. If you enjoyed this episode, please subscribe and leave us a rating. It helps keep us going. For a little more AI magic, visit YourEverydayAI.com and sign up to our daily newsletter so you don't get left behind. Go break some barriers and we'll see you next time.