Loading summary
Transcript1 lines
- [00:00]
Morgan Debon
Hey, everybody, it's Morgan Debon. Welcome back to the Journey podcast, and I'm here to give you some really quick tips to help you use AI smarter, not harder. And this isn't going to be a long episode. But I just noticed in conversations with people that I've been having, with entrepreneurs that I've been advising small business owners, creators, people who are trying to get to the next level and mostly are solopreneurs, or maybe I have like one or two kind of co founders or business partners, but have like a string of freelancers or hourly employees that they're using AI, but they're very much using it at the surface level. So this is an episode just for you, so that you can figure out how to get to the next level of usage. And it doesn't have to be super complicated. We can make it very, very simple. So thanks for joining and let's get to it. So the first thing that's really important is audio prompting. Like, y' all type way too much. It's like, appalling to me. I'm like, bro, this is taking forever for you to get your thoughts out on paper. You can just use the voice button on ChatGPT and you can just audio, like, just say the things and it transcribes it for you. It also makes it less intimidating because you don't have to try to figure out the exact wording. You can just Talk. You know, ChatGPT and all these other tools, these large language models are already trained on, like, human, like, responses and how to kind of navigate human nuances. So you can just have a train of thought, like, hey, Chad, like, I was thinking about this and I want to do this, and I want to do this, but I don't know what I should do about this. And boom, boom, boom, boom, bam. Okay, so audio transcribe your conversations in chat and then just press go and let chat do the hard work for you. I think the other thing that's really important about using these tools, and I'm just going to focus on ChatGPT for the. For the sake of today's conversation, although I highly recommend that you explore other tools outside of ChatGPT is ChatGPT is smarter at prompting than you are. So there's this kind of narrative right now where everyone's like, you should be trained on how to be a good prompt engineer and all these different things. And I think that that was probably relevant, like, 18 months ago. It's not as relevant now. I find a lot of people spending a lot of Time writing and perfecting a prompt. And I'm like, why didn't you just ask Chat to write a prompt for you? And people are like, oh yeah. So I'm telling you now, just have Chat write your prompt for you because chat's going to be better at prompt engineering than you are. So you can literally ask Chat and say, hey, I'm trying to do this, I'm trying to do this, I'm trying to do this. Can you write a incredible prompt for me that helps me get this done that I can use in insert whatever tool you're trying to build a prompt for? You know, you can build prompts for mid journey, for Claude, for, for lovable, for Replit, for any of these other AI tools that require in depth prompts to generate new things. Chat is a great place to start because it is pretty good at ideation and just getting your ideas out on paper. So that's my next tip. Now let's talk about context engineering. When I say context engineering, I mean that you need to be better at providing tools, ChatGPT and all these tools with all of the files and information to actually be able to make a decision about what information is relevant for you. So have you ever been in a conversation with someone and you're like, remember that thing with that guy from a couple years ago and your friend's just like, girl, no. Like, I need more information than that. Like which guy? Oh, the guy that we met at that bar. Oh yeah. And he said that crazy thing to me. Well, remember? Well, he just got married. Okay? So you got to give it more context. Your friend's like, how would I just know what you're talking about? Like, you have to give me just a little bit more context. It doesn't have to be perfect, but I need a little direction here, okay? That is basically what is happening now with these LLMs. They've got so much information that you need to provide more context. The when, the where, the how, and ideally the files and the uploads and the PDFs. This is one of the reasons why AI is moving to having browsers. And Perplexity just released a browser. It's called comet. It's about $200 a month. Not about. It is $200 a month. Your girl's on the waitlist. So if you all know anybody at Perplexity, I mean, I know people, Perplexity, but can you just take me off the waitlist, please? Thank you. But yeah, like, the reason why these browsers are coming is because inevitably a lot of our context lives in the things that we're searching and reviewing and being on every day. Think about it this way, when I'm on Chrome and I'm like, right now I have a script in front of me of all the different topics that I wanted to cover in today's episode. And technically I would have to open up another Chrome tab to get into ChatGPT to be like, oh, can you improve this? Can you improve that? Can you improve this? Can you improve that? But if I'm in the Perplexity browser or any AI browser, I don't have to open up another tab. It's built into the system so it's able to dynamically look at everything that's on my screen. And then we can have a conversation about what's happening on my screen or my emails, or when it has Agentic webflows, it can then move things into other folders or files for me and drag it over. So having a browser powered by AI is a natural next step for all of us. What's weird about this is that Google isn't a first mover when it comes to innovation. Now, Google's not always a first mover, just like Facebook is shameless about stealing features from TikTok or Snapchat or whomever. But I feel like Google is still that bad bitch, like, who is like still dominating our world. Like, I look at my tabs open right now, half of them are like Gmail, Google Calendar, Google Doc, Google Spreadsheets. You know what? From a if you ecosystem perspective, as a person who works Google dominates my life. And the idea that in 18 months I might not be looking at a Google Doc, then I might be looking at like a Perplexity Doc or a chatgpt Doc is crazy anyways, so I digress. Let's go back to context engineering. So context engineering, you want to give chat, you want to give chat or whomever, all of the context so files, previous conversations. You know when you go back to a certain chat because it has all of this context, because you've been chatting with it about how to buy your house as an entrepreneur and it has all your financials in it or your health ailment that you've been trying to figure out, it has all your medical records in that one session. Well, you should be moving those things into projects and you should be adding your prompt and your custom instructions and your files, all that context into your project section in Claude, if you're using Claude or In of course ChatGPT, that then just allows chat to keep like basically this giant memory of this category of information in there. Just like you wouldn't ask your mom about the guy you met like two years ago at the club. You'd ask your girlfriend. Your mom's a different project. Your girlfriend has all the context related to this guy. So. So that's a different project. So just consider you're putting a lot of information in it. And to get better outputs, it's better if you have better inputs. That's my third tip. First tip, audio. Second tip, let chat be your prompt engineer. And third tip, projects. Please put things in projects and organize them. Now, you could use a custom GPT, but I don't really view those as projects. I view those as types of assistants that aren't smart yet. But basically I can tag a GPT in a conversation with chat and say, hey, run this through this GPT because it's trained on how to do a specific thing. So, for example, I have a GPT design just for my newsletters. It has my voice, my tone, how I like to format my newsletters. I probably could have also done this in a project, but I like doing it in a GPT because then I can tag a GPT whenever I'm in any other conversation. I can say at this GPT and it'll run it through that process. So that's the difference between a project and a GPT when it comes to me. All right, next thing, if you are a. What are we called? White collar workers. We. White collar workers. Kate? Yeah. So if you're a white collar worker like me, which apparently I just found out I was a white collar worker last week. But if you're a white collar worker like me, you're in a lot of meetings all day, whether you're in person meetings or in virtual meetings. And there used to be a time when people would like, have assistants join the meeting for them to take notes and then like delegate the TO dos, right? You don't need to do that anymore. You can have AI be your assistant. So if you're in Google, make sure you turn on Google Gemini when you're doing a Google hangout, because it will transcribe everything and take all the notes for you and then email you and everybody else on that call all the notes from that meeting and all the TO dos. This is really helpful, especially at a company like ours where we have a lot of internal and external meetings. We have hundreds of clients. We have entire customer services team, customer success and customer service teams that are meeting with vendors that are meeting with agencies that are meeting with direct brand leads. And it's really inefficient when seven people have to be on a call to get the information. One thing that's been really helpful is if everybody turns on Gemini notes, then that meeting note just goes into a folder. So if somebody's in a different time zone, like one of our senior leaders lives in Hawaii, which doesn't have to join every meeting, she can just go ahead and read the transcript and that's super helpful. Or if I'm like, oh, whatever happened to that one client that you guys met with last week? Well, they can just send me the transcript and that's really helpful for me from a context perspective. Just consider having your meeting notes automated if you don't want to use Google Hangouts. A lot of people use Otter at our company as well. I like Otter a lot because Otter has integrations with other tools. So I can have an external meeting, have Otter record it, have Otter record it, and then send those meeting notes to my chief of staff so that that's a more automated process for us. And it also lists out all the to do's. So Otter is another tool that a lot of people use. I'm sure there's other tools, but those are the two that I see the most frequently that I've seen be the best. All right, so the last thing that I'm going to tell you because I feel like this is a lot. If you're just trying to go from AI101 to AI102, the next thing I would suggest is that you always tell ChatGPT or whatever you're prompting to be a critic or be an expert. Because one of the things that I've noticed recently, and they haven't fixed this yet, that it's really agreeable, it'll very much affirm and say, you know what, that's a great suggestion, let me look into it and really just kind of like egg you on and make you feel great, like the smartest person in the room. When you're having these conversations, when you tell chat and these other tools, hey, be an expert, then you're able to kind of remove some of that fluff and just get straight to the point. So if you say, hey, be an expert, that's one way to do it. You also, if you have, I think the Pro version, the teams version, you can tell chat like a set for your workspace how you like it to communicate with you, like the tone. And in that conversation I have a custom instructions that say, like you are always an expert. So now I don't have to type at every prompt. Like you are always an expert, you are always a critic and you're always open to contrarian thinking and not afraid to push me forward if I'm missing something. And that just allows me to move faster when I'm ideating and when I'm having conversations with chat. And it also just removes like that first paragraph and they're like, wow, what a great question. Da da da da. I'm like, I don't need any of that. Just give me the information. I hope this was really helpful for you. You know, my goal with these videos and the episodes is to help you just work faster and be able to not feel as intimidated by how quickly AI is changing. If you want to learn more about this, you can also join our Work Smart community. We meet monthly and I always give people like live demos of whatever I'm making and then also an accountability group in a lot of ways. So we do like live conversations with each other and we'll like live execute one of the ideas or prompts. So last week we did a conversation and I had everybody make a website and code a website and kind of basically vibe code. I didn't use the word vibe code because I literally hate that framing, but that's what it was. And everybody then like shared their screen and talked about what they built and why they built it. And it was really fun because I think for many of them they never considered a world where they could not have to pay someone $10,000 for a wireframe or some sort of high fidelity prototype. They're actually able to do it really easily on Cloud Pro. And you know, over time we can move into things like lovable and other tools. But if you are looking for a working group to help you stay accountable, this is a very low cost option. All the money that we make from this group just goes back into the community. So check it out on my website, morgandobond.com Other than that, I will see you soon. Hopefully I will see you on August 8th at AI Edge, which is Afro Tech's full day of free training. We have people from Perplexity, Zapier. Hey Jen. So many other tools coming to teach you guys. Live conversations and completely free. So make sure you sign up for that AI Edge. I think you can just Google it or look anywhere in my profile for afritech AI Edge Summit and I will see you there as well. Until next time, bye.