EP 600: 6 AI myths, 10 AI systems you must learn and 10 AI trends

Topics Covered in This Episode:

1. Six Common AI Myths Debunked
2. AI as Competitive Advantage Myth
3. Productivity Gains from AI Tools
4. AI Copilot vs Autonomous AI Agents
5. Empathy and Creativity in AI Models
6. AI Job Creation vs Job Losses
7. Human in the Loop Limitations
8. Ten Must-Learn AI Systems Overview
9. ChatGPT Usage for Business Leaders
10. Google AI Studio and Gemini Applications
11. Importance of Agentic Browsers and Copilot
12. Open Source AI Model Adoption
13. AI Video Platform Skill Development
14. AI Coding Tools for Non-Developers
15. Evaluating and Benchmarking AI Models
16. Ten Key AI Trends for 2025
17. Digital Evidence and AI-Generated Content
18. Third-Party AI Chat Platform Decline
19. Impact of AI on Social Media Ads
20. Changing Landscape of Web Browsing
21. Surge in Open Source AI Solutions
22. World Models as Next AI Frontier
23. Rise of AI-Native Consulting Firms
24. Explainable AI and Agentic Traceability
25. AI’s Influence on US 2026 Elections
26. Generative AI Impact on Remote Work

Keywords:

AI myths, generative AI, AI systems, AI trends, AI fact vs fiction, AI competitive advantage, AI productivity, AI tool deployment, Copilot, Microsoft Copilot, ChatGPT, OpenAI, Google Gemini, agentic AI, agentic browsers, AI automation, workplace AI adoption, AI training, AI business strategy, AI model benchmarking, model evaluation, modular AI solutions, Hugging Face, LLM Arena, Google AI Studio, prompt engineering, context engineering, NotebookLM, AI video platform, AI creative tools, AI empathy, AI creativity studies, full-time jobs and AI, AI-driven employment changes, human in the loop, expert-driven loops, AI hallucinations, explainable AI, traceable AI, AI social ads, digital evidence and AI, deepfake misinformation, AI in elections, open source AI models, GPT-OSS, Ollama, AI coding tools, Cursor, Gemini CLI, AI for business leaders, AI return on investment, work from home and AI, ROI on GenAI, AI in consulting, new wave consulting groups, world models, Genie 3, real world data training, modular AI framework, AI agent SDK, AI business operating system, AI content creation, AI job automation, AI regulation.

Podcast Transcript

Jordan Wilson [00:00:16]:
I think that so much of a company's success when it comes to implementing generative AI can be boiled down to a handful of things. And three of those, we're gonna be going over today. What AI systems should we be using? How can we sort AI fact from fiction and what the heck is coming next? So one of the things that I've realized after doing the everyday AI show for two and a half years is I've been able to talk to a lot of smart people and I've spent thousands of hours of researching topics and trends on my own. So today to celebrate everyday AI's six hundredth episode, we're going to be going over six AI myths you should stop believing, 10 AI systems you must learn, and 10 AI trends you can't afford to ignore. I'm excited for this one. I hope you are too. Let's get into it. What's going on, y'all? Welcome to Everyday AI.

Jordan Wilson [00:01:20]:
My name is Jordan Wilson, and this thing, it's for you. It's your daily live stream podcast and free daily newsletter helping everyday business leaders not just keep up with AI, but how we can use it, cut the reel from the BS, and grow our companies and our careers. If that's what you're trying to do, welcome. It starts here with the unedited, unscripted livestream podcast. But if you wanna take it to the next level, you've gotta go to our website at youreverydayai.com. There are couple things you can do, but number one, sign up for the free daily newsletter. We're gonna be recapping the highlights from this very episode in there as well as keeping you up to date with all of the other AI news. But also, it's a free generative AI university.

Jordan Wilson [00:02:00]:
And as of today, there are 600 episodes. You can go watch the videos. You can go read our recaps. You can go listen even to the actual podcast, all on our website. It's all for free, sorted by category. No matter what you're trying to learn, we've probably already gotten all that info from the world's leading experts. So, yeah, if you want the AI news, make sure to go check out today's newsletter. Alright.

Jordan Wilson [00:02:24]:
We made it to 600 episodes. Can you believe it? Seems like yesterday, but also two decades ago that I started doing this everyday AI thing. And, you know, I was looking back at, you know, how we kind of, you know, celebrated or, you know, what shows we did at, you know, a hundred, two hundred, you know, 300, etcetera. And earlier on, it was a little easier. Right? Like, oh, here's a 100 facts you need to know about AI or something like that. At 600, we can't do that anymore. I'm not gonna give you, you know, 600, hot takes or anything like that. You all would fall asleep.

Jordan Wilson [00:02:59]:
So instead, we're doing a little math. Six times 10 times 10. Six AI math myths you should stop believing. 10 AI systems you must learn and 10 AI trends you can't afford to ignore. Alright. Livestream audience, good to see you. I'm wondering, when did anyone start listening to the podcast? Was anyone around for some of the first episodes? They were pretty bad. But livestream audience, good to see you.

Jordan Wilson [00:03:28]:
Thank you, Ruth, saying good morning from Boston. Jackie saying congrats. Appreciate it. Peter, thank you for joining Jay live from Grand Rapids, Michigan. Monica saying congrats on six hundred. Jay as well. Thank you all. Appreciate your support.

Jordan Wilson [00:03:46]:
So, let's get into it. So like I said, we're gonna tell you the six myths. You gotta stop bleeding. There's so much bad information out there. People are listening to these common AI myths. We're gonna, debunk some of the most common ones. Then I'm gonna tell you the 10 AI systems that you need to be learning and using, And this might come as a surprise to some, but we're gonna get into it. What makes my top 10 list there.

Jordan Wilson [00:04:15]:
And then I'm gonna tell you, some top trends that you need to pay attention to. And they might not all be relevant to your business lives, but these trends, I think, really are gonna help shape how the world works in the future. Alright. Let's oh, and one other thing. This was so hard. This was so hard to come up with this exact list. Y'all, like, I was thinking the other day. I'm like, I waste, quote, unquote, waste 80% of my research and 80% of my content.

Jordan Wilson [00:04:48]:
You all think, like, these shows are long as is. Y'all, like, I only give you the top 10 to 20% of information that I, you know, research on, compile, etcetera. So I think I'm gonna start doing more of these, like, hey. Go repost the show, and I'll send you all my, you know, not all my notes, but I'll send you the next tier that just missed the cut. Right? So, I did put together seven more AI myths debunked, seven more AI systems to learn, and seven more AI trends you can't ignore. So if you want access to that, I'm I'm not gonna keep you for, you know, another forty five minutes and tell you those things. So just go share this LinkedIn post. So we'd livestream to LinkedIn, trying to really, get a bigger footprint there.

Jordan Wilson [00:05:29]:
So, if you want access to those, it's already made. It's ready to go. Just go repost this and I'll send that to you. Alright. Six AI myths. You gotta stop believing. Let's get into it. Number one, AI can be your company's competitive advantage.

Jordan Wilson [00:05:45]:
No. It can't. Right? I still you would be surprised. It is the year 2025 y'all. Still companies that were, you know, super slow to adapt. Essentially, one like this. 2023, they sat on the fence. 2024, they did a a a terrible year long pilot.

Jordan Wilson [00:06:03]:
And then 2025, they're like, alright. We're gonna do this AI thing because this is gonna set us apart. No. It won't. Right? If if you are in the year of 2025 actually thinking that using AI can be your company's advantage, absolutely not. I'll say it like this. If if your company top to bottom wasn't already implementing generative AI in 2024, you were at a competitive disadvantage in 2024. So not only is the messaging, in in in kind of the competitive landscape, they're flipped in many business leaders' minds.

Jordan Wilson [00:06:39]:
You are already very far behind. If your entire company has not been trained on GenAI basics, let's just say, I would not wanna be working at your company. And I'm being serious with that. The fact that there are still large organizations out there that don't have AI policies, that don't have an AI business operating system, that haven't trained their employees top to bottom, whether it's c suite, you know, junior workers. There it is so few people. Right? Companies are investing billions and billions of dollars in, you know, building, fine tuning their own models, which I think for the most part is a waste of time. They're spending a lot of money monthly providing access. Right? Companies just think like, oh, okay.

Jordan Wilson [00:07:26]:
We're gonna fine. We're gonna do this AI thing. Here's some Microsoft Copilot licenses. Here's some ChatGPT enterprise licenses. Problem fixed. Wrong. That is not your company's advantage. Next.

Jordan Wilson [00:07:39]:
Going along with that, thinking that just providing your employees an AI tool means that productivity's gotta go up. Right? And then we saw those early studies, you know, from McKinsey digital, you know, that said, generative AI will be able to automate, up to 60 to 70% of daily manual task for knowledge workers. And I think some decision makers somewhere up there in the C suite that haven't touched a computer in a couple of years took that to heart in the wrong way. And they're like, alright. Fine. You know? It's a line item. You know? It's gonna cost us, you know, for bigger organizations, couple million dollars a year to provide, you know, tens of thousands of employees with the AI tools that they need. There we go.

Jordan Wilson [00:08:26]:
There's that 60 to 70% savings. Absolutely not. It's it's literally like saying, okay. Well, we're gonna require now our employees to speak a different language, to communicate in a different language and thinking that's it. Right? Like, if you're a US company alright. Let's take our US employees and we're gonna go to, South America. We're not gonna teach anyone Spanish, but we're gonna give them a Spanish keyboard. Doesn't work like that.

Jordan Wilson [00:08:53]:
You have to teach people. This is a brand new way to work. Right? I always say, get rid of the words reskilling, upskilling because I could vomit. No. You have to unlearn and then relearn. It is literally like learning a new language, learning to move your day to day processes into a generative AI system is like learning a new language. So just giving your employees access to AI tools does not equate to an increase in productivity. And no matter what you do, don't read these I stopped myself.

Jordan Wilson [00:09:22]:
I didn't wanna say anything too bad. Don't read these terrible studies like the MIT study and believe anything. Alright? I I I I should have started with that. Alright. Number three, companies thinking AI is a Copilot. Right? Yes. Microsoft named Copilot, and I'd say in 2023, you could look at the world's most powerful frontier AI models and AI systems and be like, yes. These are copilots.

Jordan Wilson [00:09:49]:
Not anymore. Not anymore. Right? We are in the Waymo area. Right? Today's best AI is autonomous. It can go out. It doesn't mean it's good if you're if you're not investing the time and the resources, but those companies that are investing the time and the resources, AI is no longer a copilot. It's a pilot. It can fly itself.

Jordan Wilson [00:10:18]:
Right? You don't need a human anymore to click a button and make the generative AI go. That's so 2024. Alright. Three more AI myths before we move on to AI systems. In livestream audience, I would love to get your your thoughts on this. This might be one of those, you know, I might, just screenshot some of the most valuable or entertaining comments and just share those in the newsletter. You you know, I love making this, you you you know, community effort. So I'd love to hear your hot takes or, you know, myths or even, you know, just kinda your thoughts on the ones that I'm sharing.

Jordan Wilson [00:10:55]:
So, the next one, AI can't be as empathetic or creative as humans. That is absolutely false. Right? I think one of the biggest issues is people make individual judgments on these myths based on their own expertise or usage, which oftentimes is nothing or next to nothing. Right? Someone goes in for the first time and they use an AI system. They don't know what they're doing. They're using an old model. They don't know how to, you know, the basics of quote, unquote, prompt engineering. They see something and they're like, this is absolute garbage.

Jordan Wilson [00:11:32]:
No. Your skills are garbage. The most powerful AI systems when placed in the right hands are more creative than humans, and they're more empathetic than humans. Multiple studies have shown this. There was a study in 2024 and in 2025 that, essentially looked at the bedside manner of ChatGPT versus doctors. Guess what? ChatGPT in large studies. Right? These weren't MIT studies. These these weren't marketing pieces.

Jordan Wilson [00:12:01]:
These were real studies done correctly with sound methodology showed that ChatGPT was more empathetic. Similarly, there's been blind studies, at the university level that judged creativity blindly between college students and AI systems. Guess what? The AI systems were deemed in a blind study done the correct way with proper methodology to be more creative than humans. Doesn't mean that it's gonna happen every single time or if you go in and say, give me 10 creative ideas about blank and you're like, humans could do better. No. That means that you need to put in more work on the context engineering side of using a large language model, but I don't care what anyone says, period. AI models, when used correctly, are more empathetic and more creative than humans. Get that myth out of your mind.

Jordan Wilson [00:12:49]:
Next. Yeah. AI will create more jobs than it will take. No. It won't. Sorry. Y'all, I think some people think that I'm rooting for AI to take jobs. I'm not.

Jordan Wilson [00:13:01]:
People think I'm rooting for AI. I'm not. Right? I I like humans. Humans are cool. AI is cool too. But AI will not create more traditional full time jobs than it will destroy, period. That is corporate speak. I'm not here from corporate.

Jordan Wilson [00:13:19]:
Alright? I think the future of full time work is going to look very different. I'm gonna be sharing, if you haven't listened already to our 2025 AI prediction and roadmap series, I have a screenshot of those, coming up here in a couple of slides for my live stream audience. You need to listen to them, but, AI will create millions of new full time roles, but it will not create the same number of new full time roles. You know, I'm gonna say in a five year span, five to ten year span, there will be way fewer full time jobs that exist here in The US. I don't know what the impact is gonna be abroad in The EU as an example. You know, companies that, are actually, legislating and regulating AI. That's not how it's going here in The US. And ultimately, this is driven by corporate greed.

Jordan Wilson [00:14:11]:
I think AI could, in theory, create more full time roles than it destroys, but it's corporations that are looking at headcount. And when they see when they implement AI the right way and they're like, hey. Headcount goes down, revenue goes up, stock goes up. They're gonna keep doing that. Alright. So I do think we're gonna see a massive shift over the rest of 2020 on what full time employment actually is or what employment actually is as we undoubtedly achieve AGI, and the world changes a little bit. So, but that's not true. And if you believe that that's fine.

Jordan Wilson [00:14:51]:
This is one of those things, bookmark this and, you know, let's talk in 2030 and let's look at the numbers because numbers don't lie. But, yeah, the number of full time roles will, reduce dramatically because of generative AI. Last but not least, human in the loop leads to agentic success. No. It doesn't. That is a myth, and that is terrible. And I don't know anyone else out there who has been railing against human in the loop as long as I have. It's a crutch that can't support the weight that is leaning on it.

Jordan Wilson [00:15:29]:
So many organizations that, you know, they're like, we kinda know what we're doing, when it comes to AI and, you know, they're trying to be on the cutting edge. So alright. Agents, here we go. Right? Whether you're building those in Microsoft Copilot Studio or you're using, building on top of OpenAI's agent SDK responses, their new responses API. Right. There's so many, agentic systems out there now. Most of them aren't really, agentic as we saw by the Gardner study that said 95% of, companies and vendors promoting agents, they're not actually agentic. But companies that have pushed out real agentic AI.

Jordan Wilson [00:16:09]:
They think, okay, human in the loop. I'm sorry if your name is Bill from IT. This is always the example I give. Right? Because I've talked to organizations and they tell me this and it's it's like, okay, you have this, financial agents. Right? You have this marketing agent that is out there autonomously doing financial forecasting or doing marketing work. Who's overseeing this? Who's your human who are your humans in the loop? And everyone's like, oh, it's Bill in IT. That's wrong. Human in the loop is a recipe for disaster because of agentic drift.

Jordan Wilson [00:16:44]:
Alright. You need expert driven loops. Can we all start doing that? I made it up. EDL expert driven loops. That's what we need. Human in the loop, a lot of companies are looking at that as a crutch because some smart people said it two year two or three years ago, and it's an absolutely terrible way to think of agentic AI. It's it's it's a false sense of safety. When you do human in the loop, that way, putting Bill in IT as your human in the loop.

Jordan Wilson [00:17:12]:
No. You need multiple experts driving a loop. That's how it works. Hot takes. Yeah. Jay said easy there, Jordan. Is this turning it into a Tuesday show? Are the takes a little hot? Are you still running in circles trying to figure out how to actually grow your business with AI? Maybe your company has been tinkering with large language models for a year or more, but can't really get traction to find ROI on GenAI. Hey, this is Jordan Wilson, host of this very podcast.

Jordan Wilson [00:17:51]:
Companies like Adobe, Microsoft, and NVIDIA have partnered with us because they trust our expertise in educating the masses around generative AI to get ahead. And some of the most innovative companies in the country hire us to help with their AI strategy and to train hundreds of their employees on how to use GenAI. So whether you're looking for ChatGPT training for thousands or just need help building your front end AI strategy, you can partner with us too, just like some of the biggest companies in the world do. Go to your everydayai.com/partner to get in contact with our team, or you can just click on the partner section of our website. We'll help you stop running in those AI circles and help get your team ahead and build a straight path to ROI on GenAI. Alright. Alright. Do you guys agree, disagree with those? Let's get into the AI to, AI systems you must use and learn.

Jordan Wilson [00:18:49]:
10 of them. Ready? We're getting a 600 through six times, 10 times 10 math. Alright. Let me say this. I want to take everyone back to, you know, maybe twenty years ago, fifteen to twenty years ago, think of the type of skills that you would put on your resume. Right. I know this sounds weird if, if maybe, you're a younger audience sitting in, but you, you know, twenty ish years ago, multiple decades ago on your resume, you might put something like typing speed, Right? Words per minute. Here's what you can do.

Jordan Wilson [00:19:28]:
Right. You might have put some skills like, you you know, Microsoft Word. Right? As the software industry took off in the early two thousands. Right? You might put, you know, Salesforce, you you know, building Salesforce pipelines. Right? There was all these different types of tasks that were considered important. Okay? Now I think there's 10 AI systems that everyone needs to learn regardless of what your job role or responsibility is. And, unfortunately you might be having to learn some of these on your own personal time because your company shouldn't right? Your company probably shouldn't be giving you access to all of these 10 different AI systems anyway, but I think these are actually fundamental skills. Right? In the same way that using, a computer was a fundamental skill of the nineties and early two thousands, I think understanding and knowing the basics of these AI systems is going to be, a foundational skill set in '20 in the rest of the twenty twenties and twenty thirties.

Jordan Wilson [00:20:41]:
Alright. So ready? Here they are. And I put them in the order that I actually use them. Okay? So number one, ChatGPT. Even if your company is, not a ChatGPT organization, they have the most users, 700,000,000, active weekly users. ChatGPT is huge. It's growing even as Google is out shipping them and outperforming them in many ways. The one area where OpenAI still continues to dominate and why I think it's essential that everyone learns ChatGPT is it's stickier.

Jordan Wilson [00:21:17]:
Right? One of the the the issues with Google and Microsoft is their respective AIs, Gemini and Copilot, they're everywhere. You can use them everywhere. And sometimes when something's everywhere, it ends up getting used nowhere or it's just too hard because it's too fragmented. So that's one of the main, benefits of ChatGPT. Everything is centralized. Right? And it's much easier to give employees access, where sometimes it's like, alright. Well, we don't know how to give this group access to Copilot across these seven different instances. Right? Are they using it in Teams? Are they using it in, Power BI and Microsoft Copilot Studio? Do they have the right, you know, dynamics per permissions.

Jordan Wilson [00:22:01]:
Right? It's it's it's convoluted and complicated sometimes. Even if you are a Copilot Gemini organization, you should be learning ChatGPT. Number two, Google AI Studio. Foundational skill. I kid you not. I use AI Studio all the time. Google AI Studio, they've started to put limits on there, but there's still, wildly generous, free tiers on there. If you're not using Google AI Studio, you need to start.

Jordan Wilson [00:22:32]:
Right? You can use Gemini 2.5 Pro, which right now according to most benchmarks and, evals is the most powerful model in the world for free. The context window is larger and longer than it would on the front end of, you you know, gemini.google.com if you're using Gemini on the front end. So you just even, you you know, and then there's some basic dev controls. Right? You can control temperature. You can control output. You you know, there's just a lot of, more granular control. So even if you're not technical, so shout out to, the team there, at Google that has really made, I think, Google AI Studio a force to be reckoned with, and and they've made it easier for nontechnical users as well. Next, Google Gemini.

Jordan Wilson [00:23:17]:
Right? Especially if your company is a Google organization, you need to start using it and understand it. Number four, an agentic browser. Right? There's not a ton now, but I think to be prepared for the future of work, I'm very bullish on the fact that agentic browsers are going to have a more immediate impact than actual agents. Right? I think in the long run, the ceiling for agents is obviously much higher. But when you talk about agentic browsers versus AI agents, well, one of the advantages, to agentic browsers is you don't have to rely on bill and IT. Right? Comet, I I'm using Comet, from perplexity more and more. There's dia. There's a handful of other, sound, agentic browsers.

Jordan Wilson [00:24:11]:
Next, NotebookLM. Y'all, if you're not using NotebookLM, I don't know if we can be friends. I'm kidding. Of course. But again, one of those systems that is so incredibly powerful to be able to ground, Gemini, Gemini's 2.5 in your own data only. So if you are brand new and you don't know what NotebookLM is, right, let's say your company does, logistics. So you can upload all of your information in there. And then if you ask NotebookLM about marketing, it's gonna be like, I don't know, bro.

Jordan Wilson [00:24:49]:
Right? It is, one of the best ways to get rid of hallucinations, in large language models because you feed it whatever information you have. And if you ask it anything else, it's not gonna know because it only uses the information that you upload, which is an extremely powerful way to work with AI. Alright. I'm gonna go through these next ones quickly, but Copilot. You gotta know Copilot even if you're just starting to use it in the browser. So, yes, Copilot uses a lot of OpenAI's models, but there's differences in how they behave. There's differences, in the functionality. Right? And in the future, we've heard from Microsoft, they may start offering, other models, even on the default Copilot.

Jordan Wilson [00:25:35]:
Obviously, if you're building, in Azure, you can use just about any model out there, but you need to be using Copilot, period. Right? So many of the world's largest enterprise organizations run on Copilot. I get the gripes. I hear them. Right? Co, Copilot can be I think it's I'm gonna blame Bill. Bill and IT needs to do a better job. Explain if gosh. If I know anyone named Bill and IT, they're never gonna talk to me again.

Jordan Wilson [00:26:06]:
IT departments or whoever is managing access in Copilot, you gotta do trainings, right, on just here's how to access Copilot, and you need to make sure the right teams and the right people have access to Copilot in the right places because this is still, believe it or not, still one of the number one hurdles when it comes to Copilot, integration, and usage going up across teams is access and permissions and knowing where the heck to use it. Right? I I I think one of the biggest pros of Copilot and Gemini, it works everywhere, but then specifically Copilot is like, well, it's everywhere. How do I use it? It's kind of behaves differently when I use it in place a versus place b. When I use it in place c, it can access these files, but when I go to place d, it can't. Right? So, yes, there's obviously learning, education, training, and ongoing development that needs to happen, but one of the biggest things is just understanding access and permissions. Alright. Number six, you need to be using an AI video platform. Even if you are a non creative, go in there once a week, create one simple video.

Jordan Wilson [00:27:16]:
Alright? I do think I see the future of content consumption being highly personalized, and highly targeted. So even if you're not a creative, I think it's important to understand the basics. Right? Just as if you weren't a full time creative, if you had Adobe, you know, creative suites on your resume ten years ago, even if you weren't a full time designer, full time, etcetera. Right? Companies would look at this and be like, okay. They have a baseline understanding of, an important sector, an important part of our business. I think it's important for people to understand how to use AI video. It's not hard. Next, the 10 AI systems that you must use and learn.

Jordan Wilson [00:27:59]:
You need a quick model eval. Alright? What that means? Oh, I left one off my list. All right. We're going to make it 11. Cause I can't forget this one. But you need a place to quickly evaluate models. All right. So one of the things that people don't understand is we should be building AI solutions when you're talking about using an API.

Jordan Wilson [00:28:29]:
All right. So I'm not talking about front end users. Now I'm giving, a a little bit of strategic advice for our back end people, people building solutions, and those people using those solutions that are a little bit more custom. You should be building everything in a modular way. What that means is, let's say, you know, your company has been using, you know, GPT four o, and they get rid of it. Or your company's been using, you know, Claude 4.1, and then they get rid of it, or Claude 4.2 comes out. And maybe it's not as good at certain tasks, right, because they started to focus on something else differently in the model. You need to be able to have your, either your individual, your team, or your company wide use cases, and you need to be able to quickly benchmark and evaluate those.

Jordan Wilson [00:29:20]:
And using something like Hugging Face or l LM Arena, super simple, free, but being able to go in, look at models side by side, and evaluate them for your individual use cases because these models and we're not just saying, like, oh, when you go from, you know, g p t four o to five. No. These models get updated all the time, usually multiple times per month under the hood. Right? So you might have had something set up that was working, and I'm just gonna use, you know, Gemini 2.5 Flash as an example. Maybe you had some workflow set up for Gemini 2.5 Flash that were working perfectly. Google does a under the hood update, and all of a sudden you're like, wait. What's going on? Right? Well, first of all, you have to pay attention to those. Go look at the chain, change log, see if you can find out exactly what was changed, but you have to be able to evaluate models on the back end quickly.

Jordan Wilson [00:30:08]:
What happens if one of these, quality of life minor updates to these models actually slow things down in a big way? Maybe you were using a, you you know, cheaper model like a Gemini 2.5 flash or, you know, GPT five mini on the API side and they change something and you were using it for, you know, customer service responses or something, and now you're like, wait. Now they're bad. Now these responses are terrible. You have to be ready to be modular. You have to have plug in place. You have to be able to quickly evaluate models. Alright. The one I forgot on my list, open source.

Jordan Wilson [00:30:40]:
You need to be using an open source model. So even though they're not great today compared to the frontier models in the state of the art offerings, the gap is getting smaller. The technology is improving, right, in terms of being able to actually run these models locally. But the fact now, you have models, I think probably the most impressive open source model that I've used, is GPT OSS. Right? The, GPT OSS model. So in a lot of regards, this is very similar in terms of capabilities, to g b t four o. Right? So if I would have told you a year ago that you could download g b t four o, never pay for it. All the data that you connected to, it would be completely private because it runs offline.

Jordan Wilson [00:31:30]:
It runs locally on your machine. You don't even need internet. If I would have told people that right. People would have said this is nano bananas. I can't believe this. Right? But you should be using and experimenting and maybe eventually using some of your more, proprietary or sensitive AI use cases. Maybe those that you haven't, been tapping into AI, on the cloud side, bring them open source. There's great tools.

Jordan Wilson [00:31:58]:
You know, Ollama is probably the one I use that allow you to run models locally. And, again, as the hardware catches up, right, GPUs are getting more powerful. You know, your average PCs are getting more powerful and more capable to run these models locally. Alright. Next, Claude. Alright. Yeah. Even though I'm not a big Claude fan, you should still be using it.

Jordan Wilson [00:32:20]:
Alright. This is the Microsoft Word, Microsoft PowerPoint, right, that goes on your resume. You need to understand the basics of how Claude works. Alright. Again, think of typing. Think of Microsoft Office. Think of, right, if you are in any, analytical position. Right? People would always get, like, you know, I got my Google Analytics certification.

Jordan Wilson [00:32:46]:
Right? Even if they're not working in Google Analytics, there's certain foundational skill sets that you just have to go through and understand to make you an adaptable, and employable person in the future. And even learning Claude, learning Copilot, learning Gemini, learning ChatGPT, those are included. And then last but not least, an o an AI coding tool. Whichever one it is, you need to learn one of them, at least. I'm not gonna say learn all of them because there's a ton. So rather whether that's, Cursor, OpenAI's codex, Google's Gemini CLI, replic Zed, GitHub, copilot, etcetera. Find one. Right.

Jordan Wilson [00:33:24]:
And I'm not saying you have to turn into a developer, build yourself little applications. They're not hard, Right? You can literally build yourself. Right? If if you look at software that was impressive, you know, desktop software that was impressive ten to fifteen years ago, you can go build those types of solutions now in less than an hour. Right? And run it locally on your own machine. Something that solves problems for you. Right? The way I started doing it, I started with simple Chrome extensions, you know, doing little simple web apps. Right? Start solving your own problems first. But I think you need to ultimately pick one of these AI coding tools.

Jordan Wilson [00:34:03]:
Whew. Here we go. We're on our last section. Six celebrating 600 episodes, six AI myths, 10 AI systems you must learn, and 10 AI trends. So now as we, pivot to our last section, because some of these trends I'm not gonna say they're random, but these aren't the most important AI trends I'm looking at. For those, make sure you go back and listen to episodes four forty three to four forty seven. I should have done, like, a midterm update on these. I was planning on do it doing it, and then things got busy in life.

Jordan Wilson [00:34:46]:
But go listen to them. Alright? It's a five part series. They're very short episodes. I know sometimes I ramble, and these episodes turn into very, you know, long, listens. And I apologize for, all the people that say, oh, I listen every day on the treadmill and then I feel bad. So go listen to those. They're very quick. There's five of them.

Jordan Wilson [00:35:09]:
They're twenty five minutes each, in going over 25 AI predictions and road maps for 2025. So, these next, AI trends to pay attention to, they're not necessarily the biggest or most important, but these are some of the ones that, have either fallen through the cracks, maybe didn't make, that original series or something that has emerged in the last few months kind of out of nowhere. Alright. Here we go. Digital evidence will be a thing of the past. Alright. So I'm not just saying this from a legal perspective, but even just thinking what you believe. Right? Things you read and see online.

Jordan Wilson [00:35:50]:
Anything digital, I wouldn't believe it. I know that sounds weird. Right? That's why, if I'm being honest, that's why I do this as a live stream. You know? And that's why as much as I can, I try to take questions and comments from the audience? Right? I wanna be your your your human friend in AI. I literally still get people leaving comments on the live stream. Right? I saw one last week. Someone's like, oh, this is an AI. Like this isn't a human.

Jordan Wilson [00:36:19]:
I'm like, nah. It's it's it's me. Right? But I get that. You should be skeptical of everything you see online because the AI systems, voice cloning, digital twins and avatars, the AI photo, AI video, it's so good. I mean, look at the nano banana. Literally, I can upload a photo myself. I'm going to get a complete duplication. I can put myself anywhere doing anything, then load that into v o3 and it looks real.

Jordan Wilson [00:36:52]:
Right? And this is the worst that technology will ever get. Think about that. Digital evidence gone. And I'm not just saying from a legal perspective. Next, I think third party AI chats are going to start to die. So there's great ones that I think are obviously going to stick around. Right. But essentially there's literally thousands of platforms, and I'd say there's hundreds that are well funded.

Jordan Wilson [00:37:19]:
And I'll say there's dozens, that have enterprise footing, so you need to be careful here. I think eventually, some of these bigger ones are gonna die. I'm not saying the, you know, you.coms and the pose. I think those are obviously gonna stick around and keep growing. But there's so many of these, like, second tier third party services that all they do is they're like, okay, You can come on here and, you know, use every single, you know, large language model, right, for $20 a month or $15 a month. Right? They're not gonna stick around for long because I think what kept them going on early on, was the ability to use different models because a year ago, certain models just didn't have all the features and functionality of others. That's no longer the case. Right? So I think early on these third party AI chats, were more just identified gaps in frontier labs, go to market strategies.

Jordan Wilson [00:38:18]:
And I think the big players have already adjusted that. So I don't see and especially on the scaffolding side. Right? The tools. Right? Because early on, you know, I I I mean, even Claude, like, I don't know, five months ago, Claude didn't have access to the Internet, which, I mean, they should've just went out of business just for that because that was an asinine, predicament. The fact that, you know, we were in the year 2025 and Claude didn't have access to the Internet. Right? So, essentially, there's all this the the scaffolding or or tools that these third party chats would provide. Most of the models, right, when we talk about, Copilot, Claude, Chad GPT, and Gemini, they have all the scaffolding. They have all the tools now.

Jordan Wilson [00:39:03]:
Right? And it's it's very similar. So I, I do think they're going to start to die off. So I wouldn't put a lot of your day to day business processes into any of those third party chats. Next social ads. Aren't real. Let's be honest. If you're seeing any social ad, assume it's fake, assume it's AI. And it probably is.

Jordan Wilson [00:39:22]:
Right? And and maybe that's better. Maybe that's better. Right? The this is kinda related to one of the, 2025 AI predictions. I I said AI video is gonna catch up in 2025 and, not kill UGC ads, but essentially make them obsolete or put a, expiration date on them. And I think that's already happening. Next, wow, web browsing will start to die. Oof. It's a hot take.

Jordan Wilson [00:39:51]:
I know it's not popular, for people in the media. You know, I was formerly a journalist. I think traditional web browsing is going to start to die. And that means a lot for the future of media. And I'm not gonna get into it too long. I'll probably do a dedicated episode on it on it now. But I think that you can probably even say personally, I go to so fewer websites myself. Right? I'm absolutely loving Google's AI mode.

Jordan Wilson [00:40:26]:
They've made it so much better recently. So between AI mode, between deep research from OpenAI and in Gemini, I'm not going to the websites. I'm not going to websites a lot. Right? When I have eight tabs open in Chrome, it it's because I got Copilot open. I got Gemini. I got, ChatGPT. I got Google AI Studio. Right? I have Claude.

Jordan Wilson [00:40:48]:
That's why it it I don't have 30 different websites up. And I think that it's gonna lead to a lot of you you know, I've always said the three choices media publishers have, but I think actually going to websites is going to start to go away. Open source will surge. That's our next one. You know, there's been this this promise and premise of of open source models for a long time, But up until about a month ago, when GBT OSS was released, I don't think there were a lot of capable models that the average person could run on a consumer computer. Yeah. And when I say consumer computer, you probably gotta spend at least, like, $3, but you don't need a supercomputer. You don't need a server.

Jordan Wilson [00:41:34]:
You can literally go out there, buy a higher end consumer laptop, and you can run a model that's as powerful as GPT four o and GPT OSS locally. I think open source is going to surge in the next couple of quarters. I think that companies, like I said, that were fence sitters, because of maybe not wanting to handle certain, private or sensitive data, there's no there's no need to now. There's no need to have that, hesitation. Yes. You still have to have, expert driven loops. You still have to train people how to use them. You still have to be able to look out for hallucinations.

Jordan Wilson [00:42:12]:
You have to understand, the the the basics of, you know, context engineering, which is essentially our prime prompt polish framework. Right? But open source is gonna search. Our next couple, I'm gonna wrap these up quick here. I think world model competition is gonna go bonkers. We've seen Google lead that with Genie three. I think we're gonna see something soon, out of World Labs. I think world models are gonna be the new race of 2026, but I think we're gonna start to see that form a little bit in 2025. And if you didn't see Genie three, we've been sharing about that.

Jordan Wilson [00:42:47]:
Make sure to go look that up from Google. I think that the ultimate, and I've been talking about this for a long time, the end goal of generative AI is out in the real world. Right? Training data. I'm not saying we've we've hit the cap. Right? But all these frontier companies are training on the same Internet. Yes. Google has a huge advantage because they have access to YouTube. You could make, a case for Grok that Grok has an advantage because it has, you know, access to x data, but a lot of data on x is just garbage anyways.

Jordan Wilson [00:43:19]:
Right? But for the most part, these companies are training on the same Internet. What comes next is training on the real world in models that can marry your traditional, quote, unquote, text or multimodal models with real world data. Alright. Next, I think we're gonna see a new wave of big four type consulting, consulting groups. So, you know, you have your traditional consulting groups, your Accenture, your KPMG, your etcetera. Right. And, obviously, after dragging their feet in 2023, most of them woke up, and started try to become a little more AI native. But I think we're gonna see this new surge, almost like we saw startups, you know, in, late twenty, in the late twenty hundreds? How do we say that? In the late zeros, o's, o's? How do I say that? Right.

Jordan Wilson [00:44:15]:
You know, your your Facebook, your Twitter. Right? We almost saw this new category of businesses, the startups, the early startups. I think we're gonna see the same on the consulting side. I think we're gonna see some, some names that no one knows today or maybe they haven't even been created yet. And I think in five years, they're gonna be just like known as well as the BCGs, as the as these big consulting arms. But I think they're gonna start from nothing or they started from nothing in the past year or so, and they're gonna become juggernauts. Right? Look at OpenAI. OpenAI is literally providing consulting services on their own products, and they're charging millions of dollars.

Jordan Wilson [00:44:59]:
I think that there is a market for that. I think even though the big fours in the we'll say the the tier one, management consulting companies have shifted to be a little more AI friendly. AI native is always gonna be faster, and I think we'll win. Next, explainable AI becomes a key focus in agentic AI. That's because I expect a lot of big lawsuits to finally be decided, because of hallucinations or because of AI model behavior going badly. We obviously saw the very sad story, of the teenager, that tragically died, in part, the family is alleging, because of, the conversations with chat g b t. I think we're gonna see a lot of, stories very similar to that and on in the business sense, where companies weren't spotting hallucinations. They weren't spotting bad model behavior.

Jordan Wilson [00:45:55]:
And this is with one on one chatbot interaction. But when we think about agentic AI, I think explainable AI and traceability is gonna make a resurgence. Right? These were all things that we were talking about. You know, people were talking about these things, obviously, in 2022, 2023, and then they kinda hit a pause while we all talked about Rang and iGentic AI. But I think explainable AI and traceable AI is gonna become a huge focus point in the coming quarters. Next, 2026 midterms in The US are gonna be the AI election. Here's why. It's small enough.

Jordan Wilson [00:46:33]:
It's not a presidential election here in The U S it's midterms, right? There's going to be so many AI generated ads that decide key elections, right? Because they're going to come out a little too late. And it's going to swing some major elections. Book market now. The technology is too good, too powerful, too fast, too cheap, and too hard to tell the difference between what's real and fake. Literally, there's gonna be multiple US elections in the twenty twenty six midterms that are gonna be decided by AI. And I actually think it's gonna be less problematic in 2028 for the presidential elections because by then, I think that this will have, unfortunately, become a big enough problem in our society that the average everyday American will know kinda like we know, oh, like, was this photoshopped? Right? Ninety nine percent of people don't know that AI video exists in the way that it exists now. And I think that, unfortunately, the 2026, midterms are gonna be the thing that exposes that. People don't realize they're around the corner.

Jordan Wilson [00:47:47]:
I believe the first midterm start in March. Right? So we are three, six like, six a half year away from midterm elections. It's gonna be the AI election. It's gonna get a little ugly, but literally, you're gonna have deep fake misinformation, disinformation from AI sway key elections. Alright. And then our last trend, I think AI ready? We're going alphabet soup here y'all. AI will ruin W F H because of ROI on GenAI. Alright.

Jordan Wilson [00:48:20]:
Here's what I mean by that work from home. Right? We're seeing a big surge in return to office, right? RTO, because companies that have, been using generative AI, right, and are either so many still big corporations are either work from home or they're hybrid and companies are starting to invest in AI and they're like, wait, where are big returns? Right? There's something happening. Sorry, employees. Right? So many, so many employees are just automating 80% of their jobs with AI, and they're just pocketing that time. Right? You literally, I think, have a huge, huge section, a huge block of the, you know, work from home or hybrid or second computer workforce that have essentially automated a large chunk of their roles. And there may be, you know, just working five or ten hours a week because they're really good with AI or their company has antiquated processes to be able to manage workflows. It's huge. And I think that AI in turn is actually going to ruin quote, unquote, work from home and hybrid, because companies are not investing in unlearning their own processes.

Jordan Wilson [00:49:43]:
Right. They're like, Hey, here's, here's some, you know, licenses of an AI tool. And, you know, they're like, where's that 40% productivity boost. People are pocketing it. Maybe you should pay people more big corporations. All right. That's a wrap. We went over six AI myths, 10 AI systems you must learn, and 10 AI trends you can't ignore.

Jordan Wilson [00:50:05]:
Alright. I hope this was helpful. Like I said, there was a lot that didn't make the cut. If you want seven more AI myths debunked, seven more AI systems to learn, and seven more AI trends you can't ignore, and these were some really good ones. I'm like, man, I can't believe that this one didn't make my list. But we had to get the math, you know, six times 10 times 10 for our six hundredth episode. So if you wanna access those, just please repost this. If you need to find the link, it's always on our website.

Jordan Wilson [00:50:28]:
It's always in our newsletter. If you're listening on the podcast, it's in the show notes. So go find the LinkedIn post for this one, episode 600. Repost that. I will send you over those additional resources. I hope I I hope you guys like these little bonus things. I just realized so much great information, for this show just ends up dying on my desktop. Right? And I'd rather share it all with you, if you just do it, take ten seconds, and just share this information with everyone else.

Jordan Wilson [00:50:57]:
I think understanding and to go back to how I start this as I wrap up. Understanding some of the basics. Good, unbiased, well researched information shouldn't be gate capped. That's what I'm trying to do with everyday AI. So much of the information out there, it's ultimately snake oil, someone trying to sell you something. It's, you you know, a certain, big tech company is putting it out, so it's obviously swayed toward them. I want to be able to do this for another 600 episodes. But I can't do it without people like you who are finding value telling people about this.

Jordan Wilson [00:51:32]:
Right? So please email people in your organization if you find a specific episode helpful. Share it with your colleagues. Please subscribe to this if you haven't already. Share these LinkedIn posts. When I say these, I hope I say these things. I I hope they don't sound like a broken record, but, ultimately, if you find value in this show, I can only do another 600 episodes if you help me out, if you share, if you tell people, if you follow us on Spotify, leave us a rating, all that good stuff. So I hope this episode was helpful. Thank you for your support over the last 600 episodes.

Jordan Wilson [00:52:05]:
Couldn't have done it without you all. Really appreciate it. I hope this is helpful. If you haven't already, please go to your everydayai.com. Sign up for the free daily newsletter. What should we do in the next 600 episodes? Anyone wanna plan all of them out right now? Alright. Thank you for tuning in. Hope to see you back later for more everyday AI.

Jordan Wilson [00:52:24]:
Thanks, y'all.