We Predicted The Future of AI, And The Results Just Came Back..

So Neil, I did something funny with AI the other day. I basically had it look through the last 6 months of our YouTube channel. I used the YouTube API for this, and I asked it to look at the predictions you made and the predictions I made and score us over time. How do you think you did with your predictions?

I have no idea. Call it 6 out of 10, 7 out of 10.

Okay, so check this out. So it basically, okay, So let me, I basically, here's the prompt that I use for everyone here, okay? So if you use the YouTube API for the Marketing School Podcast over the last 6 months, how has Neil's predictions been versus Eric's predictions over time, right? And it looked at 320 videos over the last 6 months, okay? And it looked through all the transcripts, okay? So where Neil was right, number 1, SEO isn't dead, it's shifting to transactional and revenue keywords, okay? That's number 1. Number 2, AI is overhyped in the short term. Okay, the verdict is yes, he's correct. Number 3, you still need humans in the loop for paid media. Creative is the moat. Yes. And then number 4, economy is starting to turn around. Partially right. And then number 5, multiple AI platforms will win. Okay, so it scores you on these things. And then it looks at— it basically gives me like, it looks at your things and it looks at my things too, right? And so— Basically, here's the key takeaway here. And I'm going to give you a couple of YouTube API use cases, but I just thought this would be funny to mention to Neil because he's traveling right now. So Eric, where he's been wrong. So number 1, I said madness will take over in 2026. I overshot. Okay. Will AI beat human IQ in 2026? Directly right on benchmarks, but framing was sensational. Okay. And then I said UBI is going to happen, deflationary utopia timeline, right? Too early. So Where I'm generally wrong is I will overshoot my predictions, right? Now, for you, where Neil has been wrong, Neil underestimates the speed of AI displacement and then downplaying the coding revolution, right? So it gives all these quotes here. So overall scorecard, to your point, Neil, you said 6 out of 10. It said Neil is right, batting about 65%. Okay. Strength is market structure calls. So multi-platform, SEO shifting, not dying, creative as a moat. He reads the macro correctly. Weakness is he's consistently 6 to 12 months behind Eric on AI velocity. Eric, I'm batting about 72%, okay? My strength is technology trajectory calls are generally correct. He's a practitioner, not just a commentator. His best calls come from firsthand usage. Weakness is he tends to extrapolate personal experience into universal timelines. The excitement runs ahead of the adoption curve. Okay, here's the main thing for everyone, okay? So the pattern, AI does a good job with patterns, and this is what I'm trying to give to everyone here. You use a YouTube API, you use an X API, it'll spot the patterns for you and you can get better from it, right? And Neil, I never thought about it this way, but it says the meta pattern why this podcast works, Neil, is Neil is right about what persists. So SEO, creative, multi-platform. Eric is right about what's changing— agents, coding, K-shaped work. Together they cover the full picture better than either does alone. That's probably why the podcast works. Isn't that funny?

It is. And my favorite part out of this whole thing was How it states you are sensational. I thought that was entertaining because it was like, it doesn't matter if you're sensational or not, but it was just funny how it picked up on that. And it's like, you know, but it was spot on. And I figured roughly what my batting average is because I said 6 or 7, I believe, right? And it was right in between, 6.5. Yep.

And like, by the way, like if Neil's at 65%, I'm at 72%, neither of us is really that much better than each other. That's why this thing works, right? And Neil has a lot of experience. I have a lot of experience. That's all to say that we're going to be wrong a lot of times. Like 3 out of 4 wrong is still a lot. And so every— any podcast you listen to, you have to kind of apply your own level of experience and thinking to it. But my point of saying all this, like the way I use this for YouTube, we use it to generate new ideas. We use it to look at other competitive channels, to look at ideas they're coming up with. When you leverage a YouTube API, the X API, do these things and you combine them, it actually helps a lot from a content generation standpoint because these things are really good at pattern recognition.

What I've learned over the years, and I think this has helped us actually do quite well, if you look at the batting average here, right? So call it 3 out of 10 times we're both wrong. All right. Eric and I love to be proven wrong. So when we have ideas on what we think work for marketing, we love when people internally or externally want to challenge us, not because it's a fighting battle to be like, I'm smarter than you, or I have more experience. It's actually none of that. It's more so, great, prove me wrong. Because if the data shows that we're wrong, we're really happy because we found a new channel, a new avenue for growth. And I think that's the biggest mentality shift that marketers need to have. You're going to have ideas on what you think works or doesn't, but if you disagree with someone, be open to looking at data and people who are trying the stuff that you disagree with. Because you'll be shocked at how many times you'll find some winners in there. And it's not because you're wrong, it's more so because algorithms change. So what works today won't necessarily work tomorrow, and what doesn't work today may work tomorrow.

Yeah, so your relationship with change— we've said this already— your relationship with change and adaptability has to be really strong in this era. And I'll tell you what, Neil, I just came from a call where I was arguing with this head of growth from a well-known AI startup, and he's like, I challenge you on this. I'm like, well, I challenge you on this. He's like, well, I push back on this. He's Indian too, so I knew it would be fine because the bickering that we do on this podcast, right? We bicker all the time and it was just like that. We're bickering. I was like, I just want you to know I'm not mad. And we just kept going at it. It was public. There are other people in the room. I'm like, you're wrong here. And now we're pulling up data proving each other wrong. But in the end, we respected each other more for it. That's also part of it too.

And it's really cool because as marketers, you should be open to when people share data or experiments, even if it's a really small sample size. Don't push it away. Look, because no joke, you can always end up learning something new. And I was giving a talk in Toronto. I don't know if we talked about this a while ago. I forget the exact revenue numbers now, but there was a company that grew from like a million to call it 20-something million in monthly recurring revenue within like a matter of like 3, 4 months. It was a really short time period. Crazy growth. And sometimes when people are doing these kind of things and they're leveraging affiliates, a lot of their tactics are aggressive and a lot of the way in their messaging is aggressive. So people could be looking at it like they only grew because it's shady, and that could be the case. But when I look at companies like that and Eric looks at companies like that, what we look at is how are you able to scale that fast? Oh, okay. Yes, maybe the ads were overpromising. Maybe they're pushing boundaries on areas Eric and I would not want to push on. But oh, the root behind this, or the root, one of the root causes for growth is not the sensational type of marketing because you were doing that 3, 4 months ago. It's actually used AI to create 4,500 ad variations every single month. And now you could say, okay, that's not that hard. Let's just use tools and crank out 4,500 ad variations. But if you're not doing it in a systematical approach and you don't have some reasoning behind what you're creating, You're just going to create 4,500 ads. A lot of them will be duds and you'll lose money on a lot of your ads because bad ads means inefficient advertising. But we start looking at the processes and what people are doing. We then take those tactics and implement it in legitimate businesses and legitimate strategies. And we don't push too aggressive and we don't promise the world. And we found over the years, because we met a lot of these people at these masterminds, that's caused a lot of I would say our growth over the years, like I met a ton of them at War Room, right? Like just people being really aggressive with forced continuity.

Yeah. What I would say is, by the way, you want to explain forced continuity to people? I don't think many people know that.

Sure. So to give you context, this was maybe more than 10 years ago or something like that. Eric and I would be sitting down in a roundtable. So mastermind, think of like a group of marketers all getting together and people talking about growth. There were some marketers who talked about like, yeah, I took a business from $0 to $100 million. I know that's more common in today's world with AI. 10 years ago, that was not that common. You're like, $0 to $100 million bootstrap, no venture capitalists? How? Oh, I sell anti-wrinkle cream. Okay, what makes your cream so unique? Oh, just the same stuff. But check this out, we do a free trial and then we charge every single month. So I was like, but how does that convert? Why wouldn't most people just not pay? They're like, oh, we just say it's a free trial and we keep charging people every single month until they cancel. That's forced continuity because you're charging people when they don't realize they're being charged. To me, that's not an entrepreneur who is good or clever. That's just someone being unethical, especially in marketing.

To me, that's a— I mean, to be fair, Neil, I would say that's a short-term thinking entrepreneur, not a long-term thinking entrepreneur. Very entrepreneurial still. We've always said this on the podcast ever since we started this podcast close to 10 years ago now, Neil. Do things, do black hat things in a white hat way. So the black hat things that work often, they're innovating quite a bit, right? And if you figure out how to do it in a white hat way, which is clean, then you're going to get the best of both worlds. And so that's why Neil and I, we're always trying to learn. We've just always found too that when you try to do things short term, it never ends up working out.

Dude, the first person that actually told me on the black hat thing, the white hat way, was Ryan Deiss. Because when I first met him, I flew to Austin and everyone was doing like a soap opera series with email marketing, which is a way to like drive people through different emotions and then eventually cause them to convert, or doing a tripwire where you get someone to buy something really cheap that's an amazing deal, and then you upsell them on higher priced items to boost your conversion rates and your LTV. Both those models work. So when I met up Ryan in Austin, I remember him saying, he's like, look dude, this whole space with info marketers, it's a shitty space, but some of these marketers do some really clever stuff. Take those tactics and apply them to legitimate businesses and legitimate offers.

Quick break. If you want to run personalized LinkedIn ads and have personalized landing pages to convert your customers at a much higher rate on LinkedIn, check out Karrot. That's K-A-R-R-O-T.ai. Karrot allows you to do things that LinkedIn ads does not allow you to do. Right now, check it out. There are publicly traded companies such as Semrush and Sitecore that are using this right now. And you can use this to get ahead because if you don't use this, it will take you weeks, if not even a little longer than that, to make these ads. Because again, you cannot do this on LinkedIn. So again, go to www.karrot.ai and we'll see you on the other side. Dude, you want to speak about— so I was speaking of arguing this week. I was, you know, I was debating someone. No, we can call this an argument. So here's the thing. We have the Single Brain website up, okay? And so this person who's a marketer, she was debating with me and she's like, this is AI slop, we have to redesign this page. And so here's what I said, okay? I just said, hey, do you know how well this page converts right now? Like good or bad? She's like, I don't know, right? She thought I was trying to like trap her or something. And I said, no, I'm just curious. You want to change this, you want to allocate resources to redesigning this. How do you think it converts right now? She says, I don't know, okay? So then here's the thing, Neil, that page, singlebrain.com, which is for sure AI, it looks AI generated, right? Doesn't matter.

Wait, pull it up and share your screen. We're on a piece.

I'm going to share my screen, Neil. I'm going to get your feedback on this. Okay. So, and you're, Neil will give his honest feedback. You know, I'm not telling him to say positive or negative things here. Take a look at this. You see this?

Yeah.

Okay. So this is SingleBrain, same headcount, 3x pipeline, 2x deal size, half the rev— the cycle time, secure AI revenue agents. So that's what it does.

Okay.

So we build and manage team of AI agents that run your revenue operations 24/7. They source leads, write outbound, create content, recruit candidates, and get smarter every week. Work with us.

I'm assuming you're pitching this to B2B companies, not B2C.

Correct. This is B2B. These are revenue agents. And you scroll down, your team spends time here. Here's the numbers over here. Here's your agent team over here. And then this is enterprise-grade by default because we run it on NVIDIA, their enterprise runtime. So all that to say, okay, Does it look AI-generated? Yes. But would you qualify this as AI slop? No.

So I would only make a few main tweaks to it. I would reduce the white space, you know, like there between the logo and the headline and between each section, there's a decent amount of white space. And I would also add more call to actions. But other than that, dude, this page isn't bad at all.

It's not bad for being AI-generated. And I actually agree with Neil.

Okay.

So it took me like an afternoon to make this, Neil, because I didn't want to wait like 6 weeks to get a page up, okay? Once we got this up, you know who came in? Um, we got, uh, Vodafone came in, okay? Enterprise company.

Great.

They went through this, you know what I mean? Like, it doesn't matter what it looks like. Like, ready? Like, this is all you need. And then we had who else? We had GymPass, now known as WellHub. They came through this too. So if it converts multi-billion dollar companies, who gives a crap, right? Just get it out right now and optimize it. Come on, man.

If you're looking at the YouTube version, you'll notice when Eric was sharing, there's things like the logos aren't centered and stuff. And you'll notice that I did not say I would fix the logos here, that I'm not saying you shouldn't fine-tune those things, but most of those things do not have an impact on conversion. Your number one goal is speed and optimize for ROI. Then someone else can go in there and fine-tune the copy and adjust and tweak and A/B test. But it's like, to me, Yeah. You, you said you did this in a weekend. How many hours?

It took, dude, I did this in 2 hours on my OpenClaw.

And getting it up and launch form fields.

It puts it on Vercel. It, it connected to my domain. So I didn't know, you know, when you have to do like the whole like CNAME and like trans— it did all of that stuff. So it's got it set up on a domain and it got it live.

That's a great website for 2 hours. If someone has an issue with it, tell them to create, go create a better one in 2 hours.

Exactly. And here's the point. Neil said this, guys, we're not saying don't optimize and don't— like, design matters at the end of the day. But in the beginning right now, because the clock speed is so fast and you can roll these things back quickly, you might as well just get it out now, start collecting data, and then optimize it. Don't say I'm going to take 4 weeks or 8 weeks and do this because the human clock speed has changed. It doesn't work anymore. Human versus AI clock. Dude, I've been talking— my Claude coworker has been sourcing all these people. I've been talking to all these people from WPP. Did I tell you about this yesterday? I think I did.

So no, not WPP. You didn't tell me about that. Oh yeah, you told me a little bit about that.

I've been talking to these holding company people, and by the way, these are like, we're talking COO, GM level, like, you know, executive VP. So they're high up at these companies, right? I'm like, so why are you guys looking to leave, right? Um, and they're like, yeah, you know, we think the holding company model is done. You know, I, I like, a lot of these people have said, oh, there's, there's only so much I can take where I need to keep outsourcing to, you know, um, um, other countries for $20 an hour. That's what the model is. I'm like, okay, cool. And they're like, yeah, we're not just, we're not AI-forward enough. I was like, okay, Cool. So you must be very AI forward. What are you doing right now? Like, oh, I love ChatGPT. I pay $40 a month for ChatGPT. I make these agents. I was like, there's no $40 a month plan for ChatGPT. What are you talking about? Right? And then you know what they do? They hide behind the whole like, oh, WPP, they only allow us to use these closed models or like, we can't use all these other things, whatever. I'm like, just show me something that you built. You talk about these agents, just show me something. And then by the way, Neil, here's the funny part. When I get the compensation, I'm more than happy to pay this compensation. I'll just put it to you this way. If you pay someone $500 grand and they make you $5 million, $50 million, whatever, yeah, worth it. Good ROI, right? Or if they want $1 million—

Assuming you have more than a 10% margin at $500 grand and $5 million, but yes.

Exactly, right? That's— but the whole point is we want a return on our investment. But what I'm seeing right now, Neil, is that people want, like, let's say they want $1 million in compensation, but they don't— they haven't adapted and Like their work is being deflated away. It's becoming deflationary. Like maybe they're doing like $100 grand worth of work. And so all that to say is that clock speed has changed and like they can see the clock speed's not working at their organization anymore, but they're not changing their clock speed.

Dude, totally. And it's so funny how people are like, yeah, I do this and I do that and they don't really actually do it. Or when they'll make the excuse that, oh, my company does not. Oh cool, so what do you do on your spare time at home? Because your company's not restricting your home internet and it's very little. But overall, the single brand website, it's good. I used to tell my employees before AI really pushed hard is quick and dirty, just give me something and get it up. Now I tell them quick and decent because AI, you don't need something dirty. It's good enough where you can get something decent in the first round really fast.

I like that. When did you come up with that? Quick and decent. That's good.

Like 4 months ago when I was telling them quick and decent.

I like that. I like that. Quick and decent.

That's good. But one other thing, going back to hiring marketers, I was just talking to one of my leaders about this the other day. We had a person that we hired in Southeast Asia and the salaries there are quite affordable. Let's call it less than 6 figures or somewhere around there, maybe with taxes and stuff like that. I'm talking about USD. So if you do the currency exchange, they're like, oh, it's quite a bit. That's great wages there. And I remember telling the person when I first met them, this person won't work out. And they're like, why? I'm like, they're not a go-getter. They're not going to close business. They're trying to coddle people. I'm like, this doesn't work out culturally for us. I'm like, this person's not going to end up working out. You can figure out what you want to do when. I'm like, you're running this region, you figure it out. But I'm like, this is just my 2 cents and I could be wrong and I'm happy to be proven wrong. They eventually figured out the person's not going to work out. So then they come to me, I was like, oh, who'd you end up replacing with? And they literally told me like, well, we want to look at budgets, maybe sometime next year we'll figure it out. We're going to do this and this internally so we can cover it and there'll be no issues. And I said, you know what you should look for? And they're like, what? I'm like, there's this other country that's close to the one that we're operating. Technically we're in both. I'm like, you can get a leader that works on both at the same time because these 2 countries are next to each other. And they're connected by land. And I was just like, you should— and it's a quick flight and people in those 2 regions work in both countries and they share the same language. And I was like, what you should do is look for someone who costs 2 to 3 times what you were paying, ideally 3. And like, that's a lot. And I'm like, yeah, good people produce better results. You go the cheap route, you're going to make less. It's like you can't start looking at it as like, I paid this marketer $300,000 a year. Well, if they made you $60 million in revenue growth, give them a bonus. Like, who cares? More power to them. And this is where I think people get wrong. We interviewed a marketer who worked for one of our competitors and they're like, we can help you get at NPD so much more LLM traffic. Look at this playbook that I'm running. And I didn't hire them. They wanted a good amount of money. It wasn't that. Their playbook was, we just create You know, 500 articles on our own website, listicles, talking about how our agency is the best for X, Y, and Z. I'm like, I get that works now. The reason we don't do it is because we don't believe that's a long-term strategy and it's useless.

Also, welcome to like 5 years ago.

Yeah. Well, that's this for LLMs, but yeah, it's 5 years ago for SEO or 10 years ago for SEO, but they're like, this will work for LLMs too. I'm like, it does work right now, but it won't work that long.

Yeah. Yeah. It's so anyway, it goes back to how we started this too. You want to go for black hat things that work in a white hat way, even though programmatic's technically not black. I mean, Maybe it's a little gray if you spam it too hard. So—

Wait, going back to the people you were interviewing from WPP, you put out, I'm assuming, a job post, correct?

No. So this is Claude Cowork sourcing based on our careers page and our Beat Claude challenge.

So they're emailing out people or messaging out people?

It's, uh, it's, it's, so it's actually on my, my other computer right now. It's, um, and our team does this too. So it logs into the LinkedIn recruiter and it just sends messages from that. Yeah, and it personalizes it for every single one, and the response rates are great. Like, I'm like, this is better than JuiceBox, which is what we pay $500 a month for or something like that. But I've been having really good conversations. Like, if anything, it's like I could pick some of these people up later, but there's, there's been some really— I will say, guys, um, the vast majority, vast, vast, vast majority of people are still not AI forward. So—

No, and then, uh, for that OpenCLaw setup, how much did you pay for the computer? I'm assuming it's a Mac Mini.

Yeah, it's— this one's a Mac Mini, but I have the, the DGX Sparks next to me too. You see this? I just, I set that up over the weekend. Ah, cool.

How much did you spend on the Mac Mini?

So the Mac Mini, that's $700. Yeah.

Okay. And the, the DGX Spark?

Dude, these are like 5 grand apiece.

Yeah. So the reason I'm asking about the cost, some people also like make their Mac Minis like all out with RAM and stuff. So it adds up way more than—

Well, that's my Mac Studio. So we're talking on my Mac Studio, right? That's all out with RAM right now.

Okay. And the reason I mentioned that to Eric is if you notice, he's spending the money. People look at AI costs like, oh, $5,000, that's expensive. Well, if you're spending $1,000 a month in tokens and it helps you save that by a one-time cost of $5,000, you'll make your money back in 5 months. Like, you can't be penny-wise, pound-foolish.

So anyway, that's a good place to end it. We will see you all tomorrow. Don't forget to rate, review, subscribe.