Harnessing AI's Power: Michael Venidis on Transforming Business Strategies and Leadership in a Digital World
Send us a text Michael Venidis makes a triumphant return to the Successful Life Podcast, sharing his journey since the acquisition of RYNO Strategic Solutions and providing fresh insights on the revolutionary impact of artificial intelligence in business. Get ready to learn how AI tools like Claude, ChatGPT, and Perplexity are set to transform the way industries like home services process data and make decisions. We dive into the critical role of quality prompts and the balance between AI and...
Michael Venidis makes a triumphant return to the Successful Life Podcast, sharing his journey since the acquisition of RYNO Strategic Solutions and providing fresh insights on the revolutionary impact of artificial intelligence in business. Get ready to learn how AI tools like Claude, ChatGPT, and Perplexity are set to transform the way industries like home services process data and make decisions. We dive into the critical role of quality prompts and the balance between AI and traditional search engines in optimizing business processes for unparalleled efficiency.
Next, we explore the evolving landscape of AI and digital advertising, focusing on how these technologies are reshaping strategies, especially for sectors like HVAC. With kids growing up in a digital-first world, the trust and use of AI are changing rapidly, pushing search engines towards obsolescence. We discuss the potential for ads on platforms like ChatGPT and the need for businesses to craft compelling value propositions that resonate with consumers’ values. Discover how AI can help businesses establish a more robust online presence by accurately representing their services and meeting customer needs.
Finally, our conversation turns to the broader implications of AI on human roles and leadership. As AI takes over certain jobs, it simultaneously opens new avenues for opportunity, pushing us to retain the irreplaceable essence of human interaction. We share personal stories highlighting the importance of passion, empathy, and authenticity in this AI-driven world. Through leadership and digital marketing, I aim to inspire and motivate others to embrace AI innovations, drawing parallels to past technological shifts and underscoring the importance of trust in integrating AI into our lives. Join us for a thought-provoking discussion on the future of AI and its potential to change the business landscape.
00:00:01.239 --> 00:00:02.943 Welcome to the Successful Life Podcast.
00:00:02.943 --> 00:00:06.352 I'm your host, Corey Barrier, and I'm here with my man, Michael Vanitas.
00:00:06.352 --> 00:00:06.913 What's up, brother?
00:00:07.480 --> 00:00:08.282 What's up, man?
00:00:08.282 --> 00:00:09.384 It's good to be back.
00:00:09.384 --> 00:00:10.990 When's the last time we did this?
00:00:10.990 --> 00:00:11.752 It's been a few years.
00:00:12.140 --> 00:00:14.122 I think it's been a couple of years, dude.
00:00:14.122 --> 00:00:22.955 It's been a couple of years I am super excited about just to talk to you and then I get to come visit you in a couple of days, which I'm excited about as well.
00:00:22.955 --> 00:00:26.804 Get to come visit you in a couple of days, which I'm excited about as well, bro.
00:00:26.804 --> 00:00:27.225 So what's been going?
00:00:27.225 --> 00:00:29.292 On Well really quick for the folks that may not know who you are.
00:00:29.292 --> 00:00:31.021 You could just give us a quick rundown.
00:00:32.024 --> 00:00:32.746 Yeah, right on.
00:00:32.746 --> 00:00:35.713 Well, one thank you for having me back, man, seriously, genuinely.
00:00:35.713 --> 00:00:39.284 And for those of you that are not familiar with me, my name is Mike Venitas.
00:00:39.284 --> 00:00:43.030 I'm the CXO of Rhino Strategic Solutions.
00:00:43.030 --> 00:00:47.555 It's a digital marketing agency that exists solely for the home services, started in 08.
00:00:47.555 --> 00:00:59.368 And 13 years later, after we built this thing, I've just been falling in love with leadership and building people, and when you build people, you meet people.
00:00:59.368 --> 00:01:00.479 And here we are.
00:01:00.479 --> 00:01:01.100 Here we are.
00:01:01.140 --> 00:01:02.500 Here we are here we are.
00:01:02.540 --> 00:01:44.343 So we were just chatting about you, were giving me a rundown about how your job has shifted a little bit since Rhino was acquired, and my question was that I think a lot of people really just don't know how to use artificial intelligence, and I think a lot of people you mentioned this previously it's only as good as the prompt that you give it, but the possibilities here are endless.
00:01:44.343 --> 00:01:55.153 You can use this tool to learn about things that you have no idea about, or, in your case well, you could talk about your case.
00:01:55.153 --> 00:02:14.932 In our case at WhoHire, we use it to basically crunch massive amounts of real data and then match that with performance data psychographic traits and then match it with performance data, which gives us the ability to have a blueprint for contractors to hire by.
00:02:14.932 --> 00:02:16.962 And so you were telling me.
00:02:16.962 --> 00:02:23.211 Actually, I'll just let you dive in and talk about how you guys have been using this.
00:02:24.080 --> 00:02:26.248 I mean, we use it in so many different ways.
00:02:26.248 --> 00:02:29.087 It would take me forever to go through them all.
00:02:29.087 --> 00:02:41.108 I think it depends on what you're trying to accomplish, right, because I'm a fan of Claude when it comes to creativity and the quality of wording.
00:02:41.108 --> 00:02:54.121 I go back to chat GBT if it's analytical, if I'm trying to crunch data sets to your point, like and I'll even use perplexity, which is another popular one, right, for Q&A formats.
00:02:54.121 --> 00:03:00.600 That has a little bit more of, I think, advances in between the realm of AI and search engines, right.
00:03:00.600 --> 00:03:19.324 But I think we're about to enter a whole new world, similar to how we would go to different directories or websites for different things, right, who here goes to yelp to leave negative reviews and complain about their restaurant dining experience?
00:03:19.324 --> 00:03:37.393 Like, somehow and somehow, these brands, they develop this presence and people utilize them a certain way, and I think what we are about to do is enter the forefront of an era where you've got AI for so many different things.
00:03:37.393 --> 00:03:42.932 Right, it's just who is going to brand it correctly and win the consumer.
00:03:43.199 --> 00:03:44.343 We all say Google it.
00:03:44.343 --> 00:03:45.769 Google won trust.
00:03:45.769 --> 00:03:47.125 They won search engines.
00:03:47.125 --> 00:04:04.626 And I think, similar, a lot of people want to talk Is ChatGPT going to kill Google and while that's very much a possibility, I still think that OpenAI and ChatGPT is branding themselves more of like efficiency, convenience, simplicity.
00:04:04.626 --> 00:04:12.808 It's whether or not they continue to push consumers in that direction, to use it like a search engine, that really is going to determine the outcome of that whole game.
00:04:12.808 --> 00:04:14.717 But that's a whole left turn man.
00:04:14.717 --> 00:04:16.482 That's a whole different conversation.
00:04:16.482 --> 00:04:26.533 We use it for a lot and we use it for a number of things, but to your point, I think the prompt and how well you use it really does determine the outcome and the result you're going to gain.
00:04:27.661 --> 00:04:35.475 So when you go to look up something now, do you first go to Google or do you first go to chat?
00:04:36.519 --> 00:04:36.721 Yeah.
00:04:36.721 --> 00:04:39.819 So for me personally, obviously I'm a digital marketer Right.
00:04:39.819 --> 00:05:04.194 So, knowing what I know, I'm still very much using Google, you know, in instances where it's trust oriented, like the quality of the result, in my willingness to trust it, I'm still very much stuck in that mindset Whereas I shift and I use chat, gpt and AI tools for cutting timeframes in half.
00:05:04.194 --> 00:05:06.908 So I use it more of an efficiency tool.
00:05:06.908 --> 00:05:09.002 How do I get things faster?
00:05:09.002 --> 00:05:16.021 How do I cut corners, simplify things, how do I teach it a process, these types of things?
00:05:16.802 --> 00:05:26.636 My mind doesn't look at chat GPT as hey, this is where I'm going to go to get the best plumber, because AI is still so new.
00:05:26.636 --> 00:05:35.444 So for me and I think some other consumers might agree, like AI is a new world, we're still learning to trust it.
00:05:35.444 --> 00:05:38.531 It's convenient and we like the simplicity and the outcome.
00:05:38.531 --> 00:05:44.209 But at what point do you get a result that is very much incorrect.
00:05:44.209 --> 00:05:46.853 It's inaccurate because it happens.
00:05:46.853 --> 00:05:53.249 I have analyzed the results of chat GPT and it makes suggestions all the time that are inaccurate and they're bad answers.
00:05:53.249 --> 00:05:55.127 So still learning how to trust it.
00:05:55.860 --> 00:05:58.168 Yeah, and that's true, and I think that there.
00:05:58.168 --> 00:06:00.627 So I think two things here.
00:06:00.627 --> 00:06:12.925 If you're looking for the top 10 plumbers in phoenix, all right, yeah, likely chat may be close, but I don't know if it's all the way there.
00:06:12.925 --> 00:06:26.901 However, if we took all the data, let's say, from a plumbing company, we took all the data out of service type, we took all the data from internally, right, and we pulled all that into chat.
00:06:26.901 --> 00:06:39.110 And now this is a bit different perspective and now you're using that plumbing company as an example, just using that internally for processes or for even SOPs, whatever.
00:06:39.110 --> 00:06:50.874 That it would be completely accurate if that was the case, right, but going out looking for the top 10 plumbers, maybe not quite accurate yet, would you agree?
00:06:51.495 --> 00:06:52.175 Yeah, for sure.
00:06:52.175 --> 00:06:58.487 I think when you have internal data sets, that you feed it and you have it analyze it, I feel a lot better about it.
00:06:58.487 --> 00:07:00.125 Let me give you an example.
00:07:00.125 --> 00:07:17.687 I've asked ChatGPT to literally run through a list of 1000 websites and collect simple information for me, and then when I go back and I check what it's collecting, it's not entirely accurate, and I'll even give you the actual example.
00:07:17.687 --> 00:07:26.701 Like I had it run through a list of websites and tell me how many of them offered electric, electrical plumbing, plumbing and air conditioning all three.
00:07:26.701 --> 00:07:41.024 And then I wanted it to spit back a list, give me a chart and a format that helped me understand, out of all of these domains and all of these contractors, how many of them did x, y and z, and it had all kinds of errors in it.
00:07:43.810 --> 00:07:46.341 Because you're, because, I guess, because you're dependent.
00:07:46.341 --> 00:07:47.423 Yeah, that makes sense.
00:07:48.065 --> 00:07:48.507 That's different.
00:07:48.507 --> 00:07:58.410 I'm asking it to crawl sites for me and it's different than when you give it a formalized data set that you know is accurate and then you import it into the tool.
00:07:58.410 --> 00:08:08.894 See, I know that's what you and I were talking about a little bit before we started going here is, I think a lot of consumers haven't fully mastered the art of how to even use it Like chat GPT alone.
00:08:08.894 --> 00:08:11.141 Are you using a free version?
00:08:11.141 --> 00:08:14.209 Are you using a licensed version, a team version, an enterprise version?
00:08:14.209 --> 00:08:17.004 Because all of these different things give it different abilities.
00:08:18.146 --> 00:08:30.680 When training and teaching the AI, I think, is when you start to enter a different arena, Like a lot of your entry-level consumers that are just becoming used to this are just asking it simple questions and looking for answers.
00:08:30.680 --> 00:08:42.005 But when you start to train a virtual assistant to be able to think like you, to talk like you, to write like you, it's pretty impressive.
00:08:42.005 --> 00:08:46.354 I've got friends that are literally building answering services.
00:08:46.354 --> 00:08:50.168 Yeah, and these voices have emotional intelligence.
00:08:50.168 --> 00:09:05.014 Man, Like I could say hey, I'm having a bad day and I will literally listen to this AI that is built on these systems change its tone and its pitch just to create an empathetic feel.
00:09:05.014 --> 00:09:08.365 Mind boggling.
00:09:08.365 --> 00:09:09.464 I'm like man.
00:09:09.464 --> 00:09:09.905 Here we are.
00:09:09.905 --> 00:09:11.443 We're going to be here for it.
00:09:12.167 --> 00:09:14.325 Yeah, no, I'm, yeah, I'm with you.
00:09:14.325 --> 00:09:26.491 It's interesting I wonder how many people I got into a little bit of that space, training the voice models and answering.
00:09:26.491 --> 00:09:57.157 And I think, at least from my perspective, our industry it's very clear, our industry is a decade behind most industries with technology, and so I think better helping the industry as a whole better understand the way to use this stuff and the benefits of it, I think it would be huge, and I don't know how many people out there are really focused on that.
00:09:58.398 --> 00:10:09.423 Yeah, I can't remember who I was talking to a couple weeks ago, but I took a stern approach that I believed the end result was inevitable.
00:10:09.423 --> 00:10:26.273 It's very much like if I was doing a presentation in a room full of contractors depending on the demographic, the age you still see hands go up that are using the phone books on the doorstep.
00:10:26.273 --> 00:10:27.974 I look at it that way.
00:10:27.974 --> 00:10:30.495 I don't think you're ever going to see the end of Google.
00:10:30.495 --> 00:10:46.644 As long as we say Google, it has some value, people go there and as a marketer, I have to make sure that my clientele exists in every single potential avenue in which somebody might find a business.
00:10:46.644 --> 00:10:47.479 So as long as the world says they're Googling it, it matters to me.
00:10:47.479 --> 00:10:55.673 But I do think that it will be the next cassette tape.
00:10:55.673 --> 00:10:57.923 Just how long is the question?
00:10:58.645 --> 00:11:02.801 Because, think about it, corey, I don't have kids, but I got a lot of friends that do.
00:11:02.801 --> 00:11:05.322 And when we talk about that, I don't have kids, but I got a lot of friends that do.
00:11:05.322 --> 00:11:12.649 And when we talk about that, how many of them are sitting in front of screens, phones, ipads and that's a conversation for another day.
00:11:12.649 --> 00:11:18.014 But the quicker you put these devices in their hands as they grow up with it and they interact with it.
00:11:18.014 --> 00:11:20.677 It's what they know, it's what they trust, it's what they use.
00:11:20.677 --> 00:11:26.748 So, regardless of how inaccurate it might be, that's the future.
00:11:26.748 --> 00:11:29.220 If children grow up with it, that's how you develop this trust and this relationship.
00:11:29.220 --> 00:11:37.408 And next thing, we're all using AI to do searches and Google over time, becomes the next phone book.
00:11:37.408 --> 00:11:38.410 It's very much possible.
00:11:38.410 --> 00:11:40.847 It all depends on how Google plays the game right.
00:11:40.847 --> 00:11:51.514 They've got Gemini and they are trying to compete and push back, and I think that this year will be an interesting year and we'll start to see how some of it's going to play out.
00:11:53.321 --> 00:12:11.100 So let me ask you this If I'm forward thinking, and HVAC company as an example, how do I get ahead of potentially an ad play on chat GPT, which is not the case now, to my knowledge.
00:12:11.100 --> 00:12:20.841 There's no ads, which is really nice, but it's got to be coming, and so how, as a contractor, how do you start to get ahead of that?
00:12:21.923 --> 00:12:25.691 Yeah, that's a fantastic question, Goodness.
00:12:25.691 --> 00:12:30.625 I think there's a couple of things that I'll say to this is one I can't remember where.
00:12:30.666 --> 00:12:39.812 I was reading it, ceo of OpenAI, I think, hired whatever specialist guru over, and I think it might've been from Meta.
00:12:39.812 --> 00:12:44.936 Like we knew they were hiring somebody that focused on ads, so right.
00:12:44.936 --> 00:12:50.731 So the signs indicate that they're going in that direction, but we also know that CEO is not a fan of ads.
00:12:50.731 --> 00:13:00.674 Now I think that in the beginning it was very much hey, let's do this $20 subscription-based model and we can generate a business going through this direction.
00:13:00.674 --> 00:13:05.893 But advertising it's way more scalable than a set subscription model.
00:13:05.893 --> 00:13:09.259 You can spend more, you make more, right?
00:13:09.259 --> 00:13:16.360 There's, I think, a lot of controlling factors in why I think advertising is inevitable and I agree with that and why I think advertising is inevitable, and I agree with that.
00:13:16.360 --> 00:13:29.433 And I'd even read somewhere that through their subscription model Now it's the company isn't doing very well, even though it's well-known and it's being utilized Open, open AI.
00:13:29.433 --> 00:13:52.703 I was reading somewhere online like that the financial model through their subscription platform is is not the best route for them to continue surviving, meaning a lot of their consumers like to just use the free version they're not paying for it, but I will say, to get ahead of it.
00:13:53.446 --> 00:13:57.390 The ads I think it's no different.
00:13:57.390 --> 00:13:59.674 The ads, I think are the same game.
00:13:59.674 --> 00:14:07.022 Right, like I think every single search model has always been developed trust and, in some way or form, by contributing value.
00:14:07.022 --> 00:14:13.182 Right, google's value proposition and trust was when you don't have the answer, I will find you one.
00:14:13.182 --> 00:14:21.586 So people kept coming back right and then over the years, they added the ads on top to disrupt that Meaning.
00:14:21.586 --> 00:14:34.547 As I scroll down and I look for trust through the SEO and the organic sections of the search engine, I'm going to disrupt that process by showing them an ad that is compelling and what I call that as a value proposition.
00:14:34.547 --> 00:14:36.892 What are the things that contribute value?
00:14:36.892 --> 00:14:37.833 Right?
00:14:37.833 --> 00:14:40.695 You either save money, you save time.
00:14:40.695 --> 00:14:48.614 Right, there's convenience, like different consumers value different things in their buying cycle.
00:14:48.614 --> 00:14:54.303 So to me, there's no difference If you want to be ahead of the game.
00:14:54.664 --> 00:15:09.652 As open AI starts to implement advertising into their platform, I would imagine it's going to be very similar in the sense that the art of advertising is that I need you to see that before you come in for whatever your original value was.
00:15:09.652 --> 00:15:20.083 So if I'm coming into chat GPT for efficiency hacks and to save time, we know the advertising in some way, shape or form has to be on the top, whether it be.
00:15:20.083 --> 00:15:23.832 I pray they don't do a banner where I've got to X out.
00:15:23.832 --> 00:15:26.388 They would just kill themselves, right?
00:15:26.388 --> 00:15:36.405 There's so many different ways you can do it, but ultimately it's how do I get eyes in front of this advertisement piece before I actually get into the user experience?
00:15:36.405 --> 00:15:38.350 We know that's going to be the game.
00:15:38.350 --> 00:15:49.951 So for me, at the end of the day, if you're asking me what I think is the way you prep for that, it's what is the value proposition, which is no different than how it is today.
00:15:50.779 --> 00:16:00.505 Like any single contractor that watches this webinar, I beg you, if you're going to play advertising in any way, shape or form on digital, what is the value prop?
00:16:00.505 --> 00:16:01.769 Is it same day service?
00:16:01.769 --> 00:16:02.850 Can you guarantee that?
00:16:02.850 --> 00:16:03.994 Is it a discount?
00:16:03.994 --> 00:16:05.022 Is it a coupon?
00:16:05.022 --> 00:16:10.842 And for the love of God, don't tell me it's $20 off if I don't know what the actual value of the service is.
00:16:10.842 --> 00:16:15.630 $20 means nothing to me if I don't realize that it's 25% of the value.
00:16:15.630 --> 00:16:17.754 How you word it is essential.
00:16:17.754 --> 00:16:24.312 Sure, $20 off of $100 value is different than just telling me $20 off in a coupon ad.
00:16:24.312 --> 00:16:27.985 These are the types of things that we look for.
00:16:27.985 --> 00:16:31.373 Are you actually going to be there at 11pm.
00:16:31.373 --> 00:16:35.711 When you say emergency service, what is the value proposition?
00:16:35.711 --> 00:16:37.417 I believe that's the bigger game.
00:16:37.417 --> 00:16:39.282 How do you compete with your competition?
00:16:39.282 --> 00:16:40.924 How it advertises?
00:16:40.924 --> 00:16:43.951 I think is it won't be different.
00:16:44.653 --> 00:16:47.907 All right, so let's move away from the advertising portion.
00:16:47.907 --> 00:17:03.812 So you've used it enough and you've put enough information about yourself in chat that if I go into chat and I look up your name, it's likely going to give me a pretty good description of you because you've used it, you've put your information in it.
00:17:03.812 --> 00:17:19.538 So I was thinking more along the lines of should a business do that in preparation for this coming for people using chat for search even though there's not paid advertising?
00:17:19.538 --> 00:17:20.320 Is that possible?
00:17:21.143 --> 00:17:21.344 yeah.
00:17:21.344 --> 00:17:24.309 So large language models don't necessarily work like that.
00:17:24.309 --> 00:17:31.176 Like I can put in information on my end, it doesn't mean it's going to mirror or reflect on what on your end.
00:17:31.176 --> 00:17:36.088 I could have 10 people in 10 different states ask chat gpt who who is Mike Finitas?
00:17:36.088 --> 00:17:39.082 And the answer is going to vary dramatically.
00:17:39.082 --> 00:17:43.796 The one commonality it's using right is an index.
00:17:44.277 --> 00:18:02.005 You see a large language model like OpenAI, cloud, all of these tools, the answers they're pulling they're pulling from an index, and an index is what Bing and Google and these tools use, meaning these large language models and AI.
00:18:02.005 --> 00:18:06.153 They still use search engines to feed it and make it smarter, right?
00:18:06.153 --> 00:18:14.334 So, in theory, me building a website about me putting that out into the internet for search engines to find.
00:18:14.334 --> 00:18:20.925 That's how I'm going to influence commonality across all of these different searches through artificial intelligence.
00:18:20.925 --> 00:18:23.893 It's not what I type into the AI.
00:18:23.893 --> 00:18:28.971 It doesn't necessarily always remember because you have so many different platforms.
00:18:28.971 --> 00:18:36.071 Like I said, you got paid, you got free, and what it retains and remembers varies based upon how you have it set up.
00:18:36.071 --> 00:18:37.654 Does that make sense?
00:18:37.654 --> 00:18:39.522 Yeah, yeah it does.
00:18:40.965 --> 00:18:42.749 It does so, all right.
00:18:42.749 --> 00:18:51.221 So you're saying that just because I've put in information about myself or whatever, that doesn't necessarily mean it's going to reflect for somebody else.
00:18:51.682 --> 00:18:52.103 Correct.
00:18:52.103 --> 00:18:59.154 However, if I was building a website, I gotta give you the perfect example.
00:18:59.154 --> 00:19:01.082 We do this podcast.
00:19:01.082 --> 00:19:07.733 You put it on your website where search engines can find it and it's indexed.
00:19:08.961 --> 00:19:21.728 There's very much a chance that when I ask the artificial intelligence the question, then it may start to see enough information about me over and over that it might start to have an opinion right.
00:19:21.728 --> 00:19:35.049 So the more popular, the more brand equity, the more awareness that is created out in the worldwide web, the more you're going to start to see, I think, consistency in some of the trends of what the large language model says.
00:19:35.049 --> 00:19:38.115 But the AI doesn't just look at an index.
00:19:38.115 --> 00:20:00.001 It looks at a lot more than that, right, it might be looking at how you've communicated with it personally, historically, and then modify what it finds on the index based upon previous historical searches that just you privately have engaged with, searches that just you privately have engaged with.
00:20:00.001 --> 00:20:12.601 And all of that again still varies based upon whether or not you're using a paid version and it's retaining everything that you pour into it, because you only have so many GPTs as well, like you, can hit a limit and it only remembers so much, depending on certain settings.
00:20:14.281 --> 00:20:15.221 Yeah, I didn't know that.
00:20:15.221 --> 00:20:34.173 I didn't really think about there being a limit, because I don't, I guess I don't, I guess I've never really, I guess, when you say limit, maybe, dive into that Questions and conversations.
00:20:34.292 --> 00:20:35.693 You go back and forth with it.
00:20:35.693 --> 00:20:55.521 The best example I can give you right is write a list of 100 questions and just start pasting each question in After you go back and forth with it enough times, you'll start to see the differences that occur, but not a limit, like you can't continue on.
00:20:55.541 --> 00:20:56.983 I think it gives you like a notification.
00:20:56.983 --> 00:20:58.726 It'll never tell you that it can't continue on.
00:20:58.726 --> 00:20:59.547 I think it gives you like a notification.
00:20:59.547 --> 00:21:00.448 It'll never tell you that it can't continue.
00:21:00.448 --> 00:21:07.655 But there's and I can't remember what it says word for word, but I've managed to get it to populate a trigger and a notification about a limitation.
00:21:07.655 --> 00:21:25.670 And I should also preface this with this, corey, because in market intelligence, like I'm feeding this thing Excel spreadsheets oh yeah over 200 000 data sets, like I'm truly testing this thing's capability.
00:21:26.372 --> 00:21:39.984 It got to a point where I fed it mass amounts of data but 10 pm the night before it was still populating the data the next morning and I would ask it, how is that project coming along?
00:21:39.984 --> 00:21:42.291 And it would respond and say I'm still working on it.
00:21:42.291 --> 00:21:49.321 Would you like me to give you what I've got now, or would you like me to continue and then respond when I'm done?
00:21:49.321 --> 00:21:51.507 And we got the system to trip up.
00:21:51.507 --> 00:21:55.723 It was so big that eventually it glitched, broke and it never finished the project.
00:21:55.723 --> 00:22:01.027 So there's for sure limitations in what it's capable of, depending on the size of what you put into it.
00:22:02.190 --> 00:22:02.672 Interesting.
00:22:02.672 --> 00:22:12.710 Yeah, well, I could see that with 200,000 different whatever you said, yeah, okay, but for the average person they're not going to get to the limit typically.
00:22:13.839 --> 00:22:15.667 Yeah, it depends how they use it right.
00:22:15.667 --> 00:22:21.621 If it's the average person and you're just doing simple Q and a, I don't think they're really going to have any concerns about a limit of any sort.
00:22:21.621 --> 00:22:40.548 But I think it's businesses that are really feeding it mass amounts of data and if they're using it to your point with very complex and thoughtful prompts to get quality out of it like I think there is a point in which you start to run into speed bumps a little bit- have you found that you've been able to teach?
00:22:42.840 --> 00:22:44.565 I'll tell you, at least for myself.
00:22:44.565 --> 00:23:10.336 I've found that if I don't know about something and I want to know let's say, for example, I wanted to know the top 10 questions that an HVAC service technician struggles with I could find that out really quickly and then speak on it pretty fluently because the answers are pretty accurate, right.
00:23:10.336 --> 00:23:27.112 So my point is I often wonder and I know there's not enough trust in the platform yet but customers eventually are going to be able to answer their own questions without calling up somebody, right?
00:23:28.299 --> 00:23:30.709 Yeah, I think it's going to be perceived that way for sure.
00:23:30.709 --> 00:23:43.990 So basically, what happens is you have chat GPT, which is what really has the crowd right, that's what everybody's using, and you'll notice the new little globe icon at the bottom, which is search GPT.
00:23:43.990 --> 00:23:49.385 It's just a form of chat GPT and they work differently.
00:23:49.385 --> 00:23:51.030 I can't remember, Don't quote me.
00:23:51.030 --> 00:23:58.273 I think chat GPT looks at just the index of Bing, whereas search GPT, I think, expands beyond it.
00:23:58.273 --> 00:24:02.029 I know that OpenAI has deals going.
00:24:02.029 --> 00:24:04.807 They had written deals with Reddit.
00:24:04.807 --> 00:24:09.791 They had written deals, I want to say, with Google as well, Apple.
00:24:09.791 --> 00:24:22.134 They're having all these connections because really what they want is access to the data sets, because the language model pulls all of it from their data sets and then formulates a response.
00:24:23.181 --> 00:24:25.828 But the point that we get at right is in Search GPT.
00:24:25.828 --> 00:24:26.711 You can see it.
00:24:26.711 --> 00:24:40.342 It assesses seven to eight different domains, cross-references all of its choices and then formats an answer choices and then formats and answer.
00:24:40.342 --> 00:24:56.654 But let me ask you this, Corey could I go on the internet right now, build, write a blog, article that I know everything and anything there is about chat GPT, get an inexperienced person who sold right, Listen to this podcast and then I'm wrong.
00:24:56.654 --> 00:25:04.489 It's possible right Like I have the access to go to the internet and write what I want to write and be wrong.
00:25:04.489 --> 00:25:05.471 Sure, you have to understand.
00:25:05.471 --> 00:25:18.249 Like it's just looking at the internet and it's assessing that and if it wants to take a risk and believe I'm right now it's preaching wrong information to the general society.
00:25:18.249 --> 00:25:31.944 It doesn't necessarily understand what's right and wrong unless there's enough information out there for it to generate its own assessment right and back to how we started, like it's only as good as the prompt you give it 100.
00:25:32.384 --> 00:25:35.009 Yeah, you know it's scary stuff, man.
00:25:35.009 --> 00:25:36.251 I'm having a lot of fun with it.
00:25:36.251 --> 00:25:40.320 Yeah, I love it.
00:25:40.320 --> 00:25:41.983 We have an innovation team over here.
00:25:41.983 --> 00:25:47.734 I know we're very intentional about being on the forefront of this and making sure the contractors are protected.
00:25:47.734 --> 00:25:55.314 So we have different groups that are running all kinds of proactive A-B tests on Gemini.
00:25:55.314 --> 00:25:58.789 We have teams running proactive tests on OpenAI.
00:25:58.789 --> 00:26:07.106 They track what they type in, they look at the results that come back, they see the inconsistency and then they see the consistency.
00:26:07.420 --> 00:26:11.171 It all depends on what our hypothesis is like, what we're testing.
00:26:11.171 --> 00:26:13.247 But that's the beauty of innovation.
00:26:13.247 --> 00:26:16.846 Right, it's not so much that the data is right, it's almost having a track record of when.
00:26:16.846 --> 00:26:26.587 It's not so much that the data is right, it's almost having a track record of when it's wrong, so that you have data that supports some of the theories you know that we believe in and you have to understand like ai is a it's new world.
00:26:26.587 --> 00:26:30.273 It is even what I say here.
00:26:30.273 --> 00:26:32.625 I say it with confidence, but we're still learning.
00:26:32.625 --> 00:26:39.505 There are definitely some things here that you you might have me on the record right now and next year when we get together again.
00:26:39.505 --> 00:26:41.391 I say I was wrong about that.
00:26:42.060 --> 00:26:57.269 Yeah Well, and it's happened so fast that likely you are wrong, right, because it's evolving to the point of really it's the fastest evolving thing that I think we've ever seen.
00:26:57.269 --> 00:27:01.153 It's hard to even comprehend how fast it's moving.
00:27:02.075 --> 00:27:03.855 It really is, man, it really is.
00:27:03.855 --> 00:27:13.556 But what I've learned in my journey is that it's better to have a stance and be wrong than to never make a decision and have a stance.
00:27:13.556 --> 00:27:25.566 I learned this many years ago 13 years at Rhino, I think I was maybe four or five in and I went to an event called SMX and I think the biggest takeaway I ever got.
00:27:25.566 --> 00:27:29.403 I quoted a man named David Meem when he made a prediction.
00:27:29.403 --> 00:27:31.289 It was his prediction, but I believed in it.
+(1400+x+1400+px)+(1400+x+1400+px)+-+1.PNG)