Everyday AI Podcast – An AI and ChatGPT Podcast - EP 303: What the AI experts are getting wrong (and right) about AI.

Episode Date: June 27, 2024

What does the future of AI look like? A lot of AI futurcasting starts as seedlings at AI conferences across the country. Sometimes we're lucky enough to attend and speak at some of these conferences. Sometimes, speakers and panelists are spot-on when talking about the future of AI. And at times, they can miss the mark. We're giving you all a quick recap of one such conference we attended -- Chicago AI Week. Newsletter: Sign up for our free daily newsletterMore on this Episode: Episode PageJoin the discussion: Ask Jordan questions on AIRelated Episodes: Ep 236: NVIDIA GTC Recap – 3 ways NVIDIA is going to change the AI worldEp 288: Big Tech’s Critics Have Gotten a Lot Wrong on AIUpcoming Episodes: Check out the upcoming Everyday AI Livestream lineupWebsite: YourEverydayAI.comEmail The Show: [e-mail] with Jordan on LinkedInTopics Covered in This Episode:1. Innovation in Business and AI2. Role and Importance of AI Experts3. Ethical Considerations in AI Implementation4. Trends in AI Startups and Public Companies5. Chicago AI WeekTimestamps:01:30 Study challenges traditional assessment, questions AI in academia.05:17 Recapping Chicago AI Week, common trends09:08 Unprecedented business innovation, US leading in safety.13:37 Transition to AI native is vital. Industry expertise shines.16:12 AI impacts on productivity, ethics, and jobs.20:12 AI startups need to be cautious with value.22:56 GPTs enable companies' similar capabilities.26:32 Demand for generative AI generalists will grow.27:52 Experts should understand basics of generative AI.32:32 AI content detection is not real, period.36:33 Engaging with content puts you in top 1%.Keywords:business innovation, safety, risk mitigation, generative AI, AI first company, AI native company, AI expertise, ethics in AI, reskilling, upskilling, AI job displacement, AI impact on workforce, public companies, AI investment, company mission and AI alignment, startups, value propositions, AI language models, prompt engineering, tokenization, context windows, transformers, AI generalists, international monetary fund, AI in academia, AI voice clone, AI content detectionSend Everyday AI and Jordan a text message. (We can't reply back unless you leave contact info) Start Here ▶️Not sure where to start when it comes to AI? Start with our Start Here Series. You can listen to the first drop -- Episode 691 -- or get free access to our Inner Cricle community and all episodes: StartHereSeries.com Also, here's a link to the entire series on a Spotify playlist. 

This is the Everyday AI Show, the everyday podcast where we simplify AI and bring its power to your fingertips. Listen daily for practical advice to boost your career, business, and everyday life. Meet Firefly AI Assistant, now live in Adobe Firefly, the All In One Creative AI Studio. Just describe what you want to create and the assistant handles the rest, orchestrating multi-step workflows across Photoshop, Premiere Express, and more in one conversational interface. You direct the outcome. The assistant accelerates execution. The AI experts are getting a lot of things wrong when it comes to generative AI.

I'm lucky enough to get to go to a couple conferences and I get invited to speak on panels. And I watch just a lot of other people talking about generative AI. If I'm being honest, a lot of experts are getting things wrong consistently over and over. So today I'm going to talk about that and a little bit more on Everyday AI. What's going on, y'all? My name is Jordan Wilson and I'm the host of Everyday AI. All right, we have some difficulties there. We have some technical difficulties, yes.

So normally we do a daily live stream that goes out to LinkedIn among other places. So we have some LinkedIn issues today, but no worries. You can still join us on YouTube, Twitter, all those other places. but everyday AI, it's for you. So daily live stream, when it works, podcasts, and our free daily newsletter, helping us all learn and leverage generative AI. So let's just, before we get into it, let me throw out a reminder. If you haven't already, go to your everyday AI.com, sign up for the free daily newsletter.

We're going to be recapping today's conversation and a lot more. All right. But before we get into what AI experts are getting wrong and right about AI, let's first go over what's happening in the world of AI news. So the U.S. is the most prepared country for AI, according to a new report. So the International Monetary Fund, the IMF, has released a new report ranking countries based on their readiness to adopt artificial intelligence into their economies. So the report assesses countries on four key measures, digital infrastructure, human capital and labor market policies, innovation and economic integration and regulation. So the U.S. and Netherlands lead the chart with a 0.77 rating.

followed closely by Finland, Estonia, New Zealand, Germany, Sweden, Australia, Japan, and Israel. All right, so make sure to check out our daily newsletter to find out which of the countries did not make the top list. All right, next, AI-generated content easily fools university professors, according to a new study. So researchers at the University of Reading and the UK conducted a study where AI-generated exam answers, outperformed real students, in online assessments. So the project involves submitting unedited answers created by chat GPT under fake student identities, with most AI answers receiving higher grades than human submissions. This study challenges the traditional methods of educational assessment and raises questions

about the future role of AI in academia, which we're going to be talking about tomorrow, by the way. Also, some people are suggesting that, oh, this is the Turing test all over again. I don't think so. Also, we're going to talk a little bit more about this actually in today's episode when it comes to AI content detection. But make sure to check out the newsletter for more on this story. Speaking of the newsletter, we snuck this one in yesterday that came in just after the podcast. But NBC is going to be using an AI voice clone of legendary broadcaster Al Michaels for Olympic recaps.

So according to reports, NBC plans to employ an artificial clone of renowned sports broadcaster Al Michael's voice for narrating daily streaming recaps of the upcoming Summer Olympics in Paris. So the AI generated voice based on Michael's past appearances on NBC aims to replicate his signature expertise and allocation, providing viewers with a familiar experience. So Michael's was initially skeptical, but then praised the AI version of his voice as astonishing and almost perfect, expressing amazement and a hint of apprehension. So the cloned voice is obviously powered by generative AI and voice synthesis technology will greet viewers by name. So I believe that's going to be happening online,

enhancing the personalization on the Your Daily Olympic Recap on Peacock. So according to reports, nearly seven million personalized variants of the recap are expected to be streamed during the games, drawing from NBC's extensive live coverage of the event. All right. So make sure to go check. out our newsletter at your everyday AI.com for more on those and to figure out why YouTube is

trying to make AI music deals and what Sam Multman just said about GPT5. All right, let's get into it, y'all. So yeah, normally we have our live stream audience, but yeah, sorry, we don't, our LinkedIn's not working today. Bummer. But let's just talk about what the AI experts are getting wrong and right about AI. All right. And also, while I'm here talking about this, this is kind of, you know, based on a recent event. So I was lucky enough to be invited to the Chicago AI recap. Sorry, Chicago AI Week. So doing a little recap of that as well.

And I just wanted to let people know this most of what I'm talking about here is not just from one event, right? I'm lucky enough to get to go to a lot of great conferences, online events, seminars, panels that I, you know, sometimes chair or speak at. So I'm really taking a broader look here. But I realized I had a lot of notes, right, when I go to all these conferences. And I wanted to talk about what people are right about and what people are wrong about because I'm seeing some common recurring trends that I think I just have to call out.

And hey, shout out. Shout out to all of our normal LinkedIn audience joining us on YouTube here. You know, Jason and Cecilia and Colby and Tara, thank you for joining us on the YouTube. So yeah, if you have questions, please get them in. But, and so, yeah, I was invited to chair a panel for Chicago AI Week. So thanks to Zhao Chen, who kind of ran to put this together for inviting me out there. Shout out Quentin for this little photo there on Twitter that I just saw this morning. So let me first just give you a quick recap of Chicago AI Week.

Yeah, like sometimes I don't want to push Chicago too much. I'm from Chicago. That's where we come to you live every single day. But I think it's worth shouting out, right? A great event that Sao Chen Zhang put on. I've had Zhao Chen on the show a couple of times here on everyday AI. So I think it puts together just a great program. And also his group, AI 2030, 1871 here in Chicago, a great host.

So I wanted to get those things out of the way. And I probably listened to more than 30 startups. you know, AI startups kind of pitch their companies to kind of a panel of judges. And I also listen to dozens, dozens of speakers, great panels, right? So panels anywhere from, you know, three to I think some of the bigger ones were six people. So listen to experts from companies like Nvidia, Microsoft, the IRS, Salesforce, Amazon, IBM, Big Banks, Big Consult. consulting firms, et cetera. So basically all industries were represented there. Also a fantastic women in AI panel last night. So shout up to the group for putting that on and really

elevating and highlighting some of the women doing great work in AI. So with that, quick recap of Chicago AI week. Now, I want to zoom out a little bit and talk about what experts are getting wrong about AI. All right. And again, I'm not singling out the Chicago AI week. That's not what I'm talking about. Like I said, lucky enough to get invited to speak on a lot of panels, online events, you know, host things for companies, et cetera.

But I've seen so many just recurring things that people are getting wrong. I know it's Thursday. I almost have some hot take Tuesday vibe. here. So hopefully our live audience and podcast audience is okay with that. But let's just go ahead. I'm just going to go into what people are getting consistently right first and then get into what experts are getting, I'd say consistently wrong. All right. So what most people, experts are right about when it comes to generative AI, right? And that's just what I'm talking about. I'm not talking about traditional AI. I'm not talking about, you know, deep learning, machine learning.

that's not what I'm talking about. I'm just talking about generative AI, kind of this recent, you know, boom over the last three years as companies everywhere are scrambling to change how they work, right? And I think that this is unprecedented. And I don't think anyone can make an argument otherwise, right? This level of business innovation, especially here in the U.S., right? Like we talked about that story there at the top of the show,

about the U.S. kind of leading in, you know, preparedness, safety and other things, you know, from this new IMF study. So I am talking through the lens of the U.S. here. I know we have an international audience. Thank you for tuning in from all over the world. But I think what people are getting right here, the experts are getting right, is the importance of safety, right? talks I've listened to, you know, ones I've attended over the last, you know, three, six months, safety is being rightly spotlighted. And it should. And I love what one of the panelists said yesterday. I wish I could shout them out by name.

But talking about this concept of not everyone can play offense, right? And I think it's a great concept to talk about. Because generally, when we talk about generative AI, we talk about things that we can create. We talk about ways that we can create new business value, ways that we can do things faster, better, cheaper. And that seems to be where most of the conversation around generative AI seems to be focused on because it's easy playing offense, right?

For talking about a sports, everyone wants to be the skill position. If you're football, you want to be the quarterback, the wide receiver, the running back. you know, if you're basketball, you want to be the shooting guard, you want to be the point guard, you want to be the oop to the alley, right? People aren't talking about, you know, oh, the defensive line, right? People aren't talking about, you know, if you turn on, you know, the NBA draft was last night. They're not talking about people's ability to, you know, play defense and they're, you know, plus minus when they're on the court when it comes to, you know, the NBA draft. They're talking about, you know, can you score from all three levels, right?

And I think AI is the same, right? But I do think people are getting safety right because not everyone can play offense. All right, when it comes to this thing. We need more people playing defense. We need more people talking about safety. But those talking about safety are getting it right. And it is paramount, right? Especially as we come up here in the U.S. on our election season.

I've said this before. I think AI, generative AI, is going to rehab, have it. on our election systems. So it's good. Most people talking about safety are getting it right and they are highlighting it and rightfully so. And we do need more people not just playing defense. We need more people talking about defense and talking about the importance, right?

But it's not going to fill seats. It's not going to fill seats, right? The thing that's going to fill seats is, you know, here's how to create 50 new pieces of content in 30 seconds with a large language model. That's what's filling seats. But we need more people focused on. safety. We need more people focused on risk mitigation because whether you want to admit it or not, think about it or not, there's a duplicity to generative AI. It is both something that can create

immense value, but it is also something that is inherently dangerous, especially if we don't take the time to understand it. All right. Another thing, experts are getting right on AI is shifting from this talk of integration to AI first, right? I think at a lot of these conferences, maybe nine months ago, a year ago or more, some of these earlier ones that, you know, even I attended, so much of the talk was about integration. And yes, that's still important, right? It is still an important conversation to have.

But at least I think now experts are rightfully so, not focused as much on integration and more about being an AI first company, which is great. I think what we need to move to is an AI native company. And instead of, you know, thinking about first, how can we, you know, do this with AI, do this with generative AI, right? It's still kind of like a second step, which is fine, right? But I think we need to move to AI native where all of our processes, where all of our operations start with generative AI or most of them, right? I mean, I'm not, you can't speak blanket

across all industries, all sectors, but I think we also need to take it a set further and go from AI first as in like, that's your first step to being AI native, like that step zero, right? Generative AI at the core, you know, so we don't have to integrate it with our first step. We're already there. But I think experts are getting that right, you know, no longer talking about when an if to implement, but we're past that, which I think is good. Another thing, I think specialties are shining. There are so many great experts, and that's something I love about attending conferences is there are people with deep expertise, right?

All of a sudden, you have people that have been doing this for 20, you know, 20 years sometimes, right? And maybe five, 10 years ago, people were looking at them strangely, right? like, say, like, why are you talking about artificial intelligence, right? I was lucky enough on my panel talking about kind of the intersection between AI and legal tech, right? I was lucky enough to be able to talk to two very smart people who had decades of experience in artificial intelligence. So now you have people with these specialties that are now shining, right? it's it's a little different right where like even something cloud or mobile or internet right when

the internet first came out there's very few people that could come out and say yeah we've we've been working on this behind the scene for decades right so artificial intelligence is not new it's been around for 60 plus years it's been being used in various sectors for many decades so now you have these people with their specialties shining and they can easily translate their very specific expertise and talk about now how generative AI can we can all take advantage of that. So I think experts, specialties are shining. And also talking about the ethics, again, I think ethics and safety experts are getting it right. I think we need to take it a step further on ethics, though.

I think the conversation that we should be having, which I talk about here on the show all the time. So we need to be talking about what happens when AI works, right? Yes, ethics are important. We need to bring in all the key stakeholders. But we need to be talking about what happens when AI works. What happens when your organization sees a 30, 40, 50, 60 percent increase in productivity? What are you going to do if you're a public company? How are you going to deal with the stakeholder pressure?

Do you have a reskilling, upskilling plan in place? Right? It's another thing when we talk about ethics that, you know, we always talk about, oh, you know, and there's all these studies. Some of them I agree with their methodology. Some of them I don't. But, you know, a lot of large studies, you know, from the IMF, from the World Economic Forum, you know, are saying, oh, you know, yes, AI is going to, you know, take away tens of

millions of jobs, but it's going to create tens of millions and it's going to be a net positive. I've said this before. I'm not a doom and gloom person, but I don't care what anyone says. AI will take away more jobs than it will create. But regardless, AI is going to create tens of millions or hundreds of millions of jobs that do not exist. But companies right now do not have great plans in place for upskilling and reskilling. So when we talk about ethics, we don't just need to, you know, have a, you know, Well, this would be nice.

No, you need to have a plan in place. If you don't already have an ethics plan in place on what happens when AI at your organization works, you are behind. What are you going to be doing with those jobs? How are you going to be upskilling, reskilling your employees? How are you educating them? How are you going to retain the good employees? How are you going to fight off stakeholder pressure, right? As we see big tech companies, the same companies that are investing billions or tens of billions

of dollars into AI yet laying off tens of thousands of employees, right? Public companies play follow the leader. So when the biggest tech companies in the world are investing their money in generative AI, but their headcounts going down and their stock prices going up, other public companies are going to feel that same pressure. So when we talk about ethics, Adobe just introduced an entirely new way to create, bringing the power and precision of its creative suite

into one conversational experience. Meet Firefly AI Assistant, now live in the Adobe Firefly app, the All In One Creative AI Studio. Powered by Adobe's creative agent, Firefly AI Assistant lets you start with your vision, just describe what you want, and shape the outcome as it takes form with the Assistant. The assistant orchestrates multi-step workflows, drawing on 60-plus pro-grade tools across Adobe Creative Cloud apps,

including Photoshop, Illustrator, Premiere, Lightroom Express, and more to help bring your ideas to life. You can also get started with creative skills, a growing library of pre-built workflows for common creative tasks, like batch editing photos, creating mood boards, portrait retouching, and creating social variations. Every step the assistant takes is visible, so you can refine, redirect or take over at any time. You stay in the driver's seat as the creative director. Adobe Firefly AI assistant now in public beta. See it today at firefly.adopi.com.

Does your generative AI use align with your mission, your vision, your values of your company? Those are going to become more important than ever. All right. And you have to have that in place. So those are some of the things that people are getting right. So some of the things people are getting wrong. Woof. A lot.

A lot. Even the experts, y'all are getting a lot of things wrong. Let's talk about some of those things. First and foremost, all right. I'm not picking on any of the startups that we're pitching. Like I said, I listened to, I don't know, 30 some over the last couple of days. So many startups, y'all are in for a rude awakening. I don't understand it.

I really don't. Building your USP, building your unique sales proposition around such a flimsy, such a flimsy value prop, right? There's still so many, so many AI companies that are starting in 2020. You know, quote unquote, AI startups, you know, sure a lot of them, you know, maybe our side projects, you know, just seeing what sticks, right? I think right now that's a lot of what startups are doing. You know, you throw a new startup at the wall every week. There's people that do that.

They have teams of talented developers. They put a new startup up every week, see what sticks. But, you know, a lot of these startups that I listen to, I felt bad. I felt bad. You have to understand, right? If you are an AI startup, if you're in venture capital, if you're in private equity, you need to be very careful, right? And here's why, a common thread that I saw not just at this conference, but a lot of conferences in general, AI startups are putting their true value on something that can be wiped away or is probably, if I'm being honest, is already wiped away.

And it's essentially creating, you know, a large language model for this or a fine, model for that, right? And trying to raise millions of dollars and to, you know, put the future of your company on that. Stop doing that. Stop. Right. Again, this is my hot take.

This, you know, I'm sure there's, you know, a minority of AI startups out there that can still do this and find great success. I'm not always right. But so many of these are so flimsy. And I just, I just wonder, right? Do people not understand, right? They think that they're going to get, you know, hundreds of enterprise clients. Oh, we're going to build a, you know, a large language model for, you know, I don't know,

retailers, you know, because retailers need a special large language model. They can't work with Claude. They can't work with open AI. You know, so we need to raise, you know, $3 million to build a large language model specifically for retailers, right? And you throw in all your, your buzzwords in there, you know, and, you know, rag, fine-tuning, you know, large language models. These models are bad.

No, your startup idea is bad. It's terrible. You know, if you think, if you think that you're going to be able to get dozens or hundreds of enterprise customers, I don't think you understand what's happening with large language models. The giant companies are going to squash you. Consider this free advice out there, AI startups, venture, B.C. firms, PE groups, working with quote, unquote,

quote, AI-powered startups. Can first of all, can we stop just using AI-powered blank? Please stop, right? Big companies are going to squash you. They're going to squash your company. They're going to squash your investment in those startups. Stop creating startups with such flimsy value props. They're not good.

All right. Let's talk about what's happened recently. Anthropic Claude just released projects, right? Much larger context window. You can train kind of your projects. You can put in your own knowledge base, all right? GPTs. We already know that from chat GPT allows companies to do the same thing.

Google Gemini coming out with gems. They announced this not released yet. So we've already seen from Anthropic Claude, this is available. We've seen this available out of GPTs. I think GPTs actually aren't as good as people hope. But the technology is only going to improve. So, hey, for all you thousands, literally thousands of startups out there who are trying to create, you know, specialized, fine-tuned rag models, you know, specific large language models. Do you really think you can compete with Anthropic?

Do you think you can compete with Open AI? Do you think you can compete with Microsoft? They're offering similar functionality. Do you think you can compete with Google? some of you may, most of you can't stop doing this. Come up with uniquely valuable SaaS companies. Come up with uniquely valuable products. Don't have it rely on just one thing.

Enough on that. All right, another thing, prompt engineering 101 is so sorely needed. So many, you know, quote unquote experts, you know, aside from listening, to, you know, 30 or so different panelists talk. All these startups pitch, I, you know, even in the last two days had dozens of conversations with leaders. And this goes back, right? I talk to people on the show almost every day, right?

We have guests about three days a week. I've talked to hundreds of people. I go to these conferences all the time. Prompt Engineering 101 is sorely needed. Even, you know, quote, unquote, AI experts are getting the basics wrong. Right. Why do I know this? Well, the examples they're talking about, the screenshots they're sharing, right? Oh, look at, look at this response from this model, right? And, hey, I asked chat GPT to do this and I got this response. Well, yeah, if you put in a 10 word response into chat GPT and if you think it's going to, you know, spit out your, your company's KPI is for the next three quarters, you have it wrong. And so many quote, unquote, unquote, AI experts still do not understand how large language models work. They don't understand tokenization.

They don't understand context windows. They don't understand transformers. They don't understand so much. So I also say this as a word of caution, right? Because when people stay in their specialties, AI experts, it's great. But now, you know, AI experts, quote unquote, are being asked to talk about more and more things that may not be inside of their specialty, y'all. If there's ever something I don't understand on the show, that's why I bring on experts, right, in their specialties.

I've spent thousands of hours talking about large language models, prompting, right? And I'm still not going to say I am the expert. I'm not, right? But people, companies, you need to be investing into prompt engineering 101. Yes, models are going to get better, right? This concept of prompt engineering is going to get easier and more intuitive as models get more capable. However, so much of the future of work is interfacing with large language models. And I think we're going to see a lot of small language models from big companies for specific

purposes. You need to understand how models work, prompt engineering 101. All right. Also, another thing that I think, well, I don't know if this is something experts are getting wrong versus just an observation is there's very few generalists out there. I would consider myself a generative AI generalist, right? I know a little bit about a lot. There's not a lot of people out there, right? Maybe there's not a market for it yet, but I think it is going to be a in-demand market in the future, the ability to be able to speak generative AI, right? If you're a listener of the show, if you read our podcast, you're in good company, or read our podcast, listen to our podcast, read our newsletter, you're in good company, right? Because I've noticed because of great

guests that have come on this show, I can speak generative AI fairly well, right? I can go chair a panel on legal tech, do a pretty good job because I've spent hours talking to experts, reading, listening, right? But I've noticed there's very few generalists, right? So when an expert that works in a specific niche gets asked a question that is slightly outside of their niche, sometimes their answer is not very good, which is surprising to me. Yeah, and I've been seeing this for you know, years, right? I follow the space very closely. But that's just something important to keep in mind for the average everyday person out

there, right? Just because if you hear someone from a meta or, you know, a big company, you know, a Google, give a response on something, you need to be careful. You need to say, is this their expertise or are they just being asked questions about things that are maybe outside of their scope, right? You need to be careful, all right? and also experts out there, you need to start understanding the basics of generative AI 101, right? You don't need to, you know, know the ins and outs of diffusion models, but you need to understand

the difference between a mid-jury and a dolly, the basics, right? You don't have to, you know, be an expert at, you know, chain of thought prompting, but you need to understand the difference between a zero shot and few shot prompting, okay? You need to understand the basics. If you are an AI expert, if you fancy yourself that, if that's the hat you put on in the morning, you need to understand things outside of your specific scope as well. All right. Some other things that I think experts are getting wrong is there's too much focus on the creation

and not enough on the curation or what goes into large language models. I'm talking about large language model. So much of the discussion, even experts, is focused on what you can create with a large language model, which is great. Right, don't get me wrong. Right. Again, that's what fills the seats. But I think the conversation needs to be more focused on the curation or what is going in or what is not going into models.

Again, your creation and what you are able to produce with generative AI is going to exponentially increase when you understand what goes into it. Again, just another thing that we're getting wrong is instead of us business leaders, enterprise companies taking time to understand and explain the black box of generative AI. Instead, they're just giving more and more people access to it, right? Because there's pressure. I get it. Oh, we're paying $60 a month per seat for this enterprise model. So we need to just give it to everyone.

Okay, sure. Companies should be hiring us to train them, FYI. But beside that, you need to understand. You need to have people on your team. really taking the time to understand models and what is actually going into them, right? Small example, small example. You know, I sat in on a panel on the future of AI and kind of media and journalism. It's my background.

I was a journalist. I was expecting to hear some talk on as an example. Oh, Open AI's partnerships with Axel Springer, with the Associated Press, with the Financial Times, right? Because this is impacting what goes into the moment. model impacts your ability to actually use it, right? And I think future models relied on a lot of sources that they maybe are no longer going to have access to, right? And it seems like that wasn't really talked about, which is fine, right?

Because we're always talking about what we can create and all the cool things we can create. It kind of goes back to this concept of offense versus defense, right? We need to also talk more about and understand the inputs of these large language models, what is going into them that wasn't there before, what are our, our next models of the future not going to be trained on that they previously were, you have to understand the models, right? It's like no one knows, you know, the Internet, right? We all know the Internet.

You connect to, you know, you have a service provider. There's, you know, or I don't know, maybe we don't know the Internet very well, you know, but we've had 30 years to understand it. You don't have 30 years to understand generative AI. You have to commit time and resources for your company to understand it now. Another thing I think people got wrong. I kind of touched on this. I think AI experts aren't doing a very good job of staying well-informed outside of their

specialty, right? Even when it comes to AI news, when it comes to, you know, big partnerships, when it comes to things that are impacting, you know, energy, climate, safety, etc. It doesn't seem like AI experts are staying well informed, you know, outside of their one very specific specialty, right? Again, just a general observation from many, many months of attending conferences, listening to, quote, unquote, AI experts speak. But I get it. It's hard, right, because there is so much happening in this space. And then last, I think last but not least, what people are getting wrong.

and I'll try not to go on a long rant on this. AI content detection is not real. Let me repeat that. And we kind of talked about this with this University of Reading study that we read at the top of the AI news. AI content detection is not real, period. So many companies, I mean, experts are talking about it. You know, there's startups that are trying to integrate something like this

into their products. If you are a decision maker at your business, if you are in charge of Gen AI implementation, large language model training, whatever, just no, that's not real. There is literally no such thing. That is not how large language models work. Period.

All right. You know, they say, oh, we can detect burstiness. And no, no, you can't. So, you know what? I'll put a challenge out there. Any AI content detection platform, you won't want to do this. But we're starting to sell sponsorships here at everyday AI. I will give you a $10,000 sponsorship here on the show.

You can get your product out to literally hundreds of thousands of people. We have a pretty wide audience across our podcast, live stream, newsletter, website, etc. I will give you a $10,000 sponsorship if you will come on this show. and allow me to break your model. Guess what? Number one, you won't do it because I will expose it, right? There's a reason why open AI, right? There's marketing reasons.

They initially put out an AI content detection platform. I think it was to give people that eads, right? Oh, it's okay, right? And then they obviously shut it down because studies found it was less, I believe, 26% accurate. So it was worse than flipping a coin. It was worse than if you, Asked an animal at the zoo choose if this is, if this content is created by AI or not. No such thing.

Literally no such thing. Models, yes, there's certain trends and patterns, but there's no such thing. I can both write as a human to sound like an AI model, right? And I'll say, in an ever-changing world of technology, let's join together as we delve into these exciting topics. I can write that. The model is going to say, oh, AI, AI generated, right? 100%. But then if you know what you're doing with large language models, which I do, right, I give talks literally all across the country. Companies pay us a lot of money to teach them how to get chat chbt to write more like a human. If you Google that, you're probably going to find

everyday AI or myself near the top of the search results. It's my background. I was a journalist, award-running journalists. Yeah, you can do that, right? You have to understand the tokenization process. You have to understand next token prediction. You have to understand temperature. You have to understand top P. If you understand those things, you can get a quote unquote large language model.

This sound extremely human. And to get a 0% content detection score, right? We did this. We busted every single one about a year ago. So, you know, with false positive, false negatives. Not a thing. Doesn't exist, right? But companies are trying to do this.

There are things that originality scores, plagiarism scores. Yeah, those have been around for decades. AI content detection, not a thing. So if you are working at a university, if you're a decision maker at an enterprise company and you say you need this, right? Oh, we need this to make sure it's our content's not AI generated. No, no such thing. All right.

And we're going to end today's show with a little bit of a hot take. All right. So if you listen to our podcast, if you engage, right, with the guests, do you know you can come in and when we have guests, you can ask them questions and they'll answer them? If you listen to this podcast daily, if you engage with the guests, if you read our daily newsletter, I'm going to go ahead and say, you are putting yourself in the top 1% of people that attend these AI conferences, right? Period. I go to a lot of these. I listen to the speakers. I talk to the speakers. I talk to, random attendees. I always go around, network, introduce myself, trying to get a heartbeat of where the general public, where the business public is when it comes to generative AI. If you listen to our show every day, if you engage, come on LinkedIn, ask experts questions, right? They'll answer them.

We might answer them live on the show. Read our daily newsletter, right? Where we look at the latest studies. We break them down. I'm a human. I write this. I write our daily newsletter. Former, like I said, former award-winning journalist.

If you do those things every day, you are putting yourself in the top 1%. All right? So we always talk about how can we learn AI? How can we leverage it? How can we use it to get ahead in our career? I know I go on rant sometimes, but y'all, we bring you some of the best minds in artificial intelligence on the show, covering topics all across the spectrum, sales, education, marketing, nonprofit, society, ethics, responsible AI, people who are

large language model experts. We cover healthcare, automotive, robotics. We cover dozens of sectors. the leading experts in the world, we bring you them every single day. So if you can't go to all these conferences and yeah, you can hear me, you know, riff on them. But if you tune in, if you engage, if you read our newsletter every day, you are putting yourself, I'm saying this with confidence, I've talked to, I don't want to say thousands, but hundreds of people at all of these conferences, you are putting yourself in the top 1%. So if you want to know how you can not get left behind, but get a head. I'm letting you know right now.

You are in the right place. All right, that's it for today, y'all going over some things AI experts are getting wrong, but also getting right when it comes to AI. So if this was helpful, please share this with someone. I know our LinkedIn isn't working today. But please share this with someone. Also, if you're listening on the podcast, please give us a rating and subscribe to the show and go to your everyday AI.com.

Sign it for the free daily newsletter. I'm signing off now, but I'm going to go write that newsletter. I'm a human. I'm going to go write it. So, thanks for tuning in. Hope to see you back tomorrow and every day for more everyday AI. Thanks, y'all. Meet Firefly AI Assistant.

Now live in Adobe Firefly, the Allman One Creative AI Studio. Just describe what you want to create in your own words and the assistant handles the rest, orchestrating multi-step workflows across Adobe Creative Cloud apps, including Photoshop, Premiere Express, and more in one conversational interface. You direct the outcome. while the assistant accelerates execution. Stay in control with the ability to step in and refine at any time. See it today at firefly.adobie.com.

And that's a wrap for today's edition of Everyday AI. Thanks for joining us. If you enjoyed this episode, please subscribe and leave us a rating. It helps keep us going. For a little more AI magic, visit Your EverydayAI.com and sign up to our daily newsletter so you don't get left behind. Go break some barriers and we'll see you next time. time.