Eric Siu and Neil Patel map workers on judgment versus AI use, creating four categories: deadweight, slop cannon, steady hands, and turbo brains. They cover where AI wins in marketing (content, outreach, creative, analytics) and debate whether judgment or specialist skills matter more as AI scales.
Chapters
Have you seen this 2x2 chart?
Use of AI, slopping cannons. I have not seen it. Does not use AI, deadweight. Has good judgment, turbo brains, steady hands.
Yeah. So what Neil's calling out, let me just map it out for people that are listening. So this is a 2x2 chart. And so you have on the very bottom, it says people who do not have good judgment. And then to the right of that, you have people who have good judgment. Okay. And then right above you have people who do not use AI versus people that use AI. So let's just put it this way. If you are someone that you do not have good judgment, nor do you not— and you don't know how to use AI, you are what? You are dead weight, guys. You are dead weight. I'm sorry. But if you're listening to this, for sure you're not dead weight because you have good judgment because you listen to this podcast. So if you have— if you don't have good judgment and you know how to use AI, guess what? You are a slop cannon, meaning you produce lots and lots and lots of slop because you think just because you know how to use it, you're adding a lot of value. Which is, I think, a lot of what Neil kind of alludes to on this podcast is like people using it to create slop. And that is not helpful to the world.
That's the majority, by the way. That's more than 95% of what I see is dead weight, just creating more junk.
But Neil, hey, question for you. What percent of people do you think in the world have good judgment typically?
Less than 5% in the world.
Doesn't that add up? And we talked about this last week when we're like, hey, if you're using it, it amplifies your intelligence basically. Like whatever you tend to do, you've just amplified it 10x.
So I've been stealing Eric's quote recently at events. AI brings out the best in you and it brings out the worst in you. The lazy people are getting lazier and they're producing more junk, and the good people are accelerating and becoming even higher performers. That we see. But when I look at marketing right now, from an AI perspective, people are only really doing 2 main things with AI. And you can add if you know any others that are taking up big buckets. One is creating content and 2 is doing outbound to just generate more leads. Those are the 2 main things we see. If I had to add in a 3rd, I would add in more creative for AI. I mean creative for ads, like mass scale producing ads. If I had to add in a 4th and it's a much smaller percentage doing this is using it for data and analytics. But the main 2 are content and outreach. Those are the big 2 that I'm seeing. I don't know what you're seeing, but those are the main 2 that I'm seeing.
I actually want to address that. So let me finish the 2x2 here. So the 2x2 here, so let's say someone has good judgment, okay? But they do not use AI, they are steady hands, okay? So meaning that they're kind of more or less stabilizers, but if they have good judgment and they use AI, they are turbo brains. And so I think the reason I'm pulling this up is any of you watching this right now, you should be thinking about looking at this chart all the time. Whoever you're looking at hiring, you should run them through this, okay? And you're like, oh well, how do I quickly test if they have good judgment or not? You know what we do? We have this thing now, Neil, called— I think I showed this to you, but it's literally the Single Grain Beet Clod Challenge, right? And so it's basically like, hey, AI could do 80% of knowledge work now. We need people who can do the 20% that AI can't. Creative leaps, pattern recognition, judgment under uncertainty. So basically you can use AI all you want. That's baseline. If you only use AI, like, okay, then you haven't beaten Claude. So we just want to see people like fill these out and then, you know, these challenges over here, here's like a brief and the situation and all that stuff and they'll send it over to us and that's basically it. So your hiring process is going to have to become tougher. That's what I generally think. Now, going back to what Neil said, I think, Neil, you're talking about use cases, right?
Yeah, because at the end of the day, yes, people can use AI for whatever they want. As a business owner and as a marketer, we really care about revenue growth and profitability.
Yeah, so I would say in general, yes, but here's what I'll say, Neil, that's interesting. So what typically are the main kind of funnels that matter for a business, right? It's how good is your recruiting? How good is your sales? How good is your marketing? If you have a good product, you need product fulfillment and product delivery, right? You kind of need those 4 main things. Am I missing anything?
Good finance, but yes.
Good finance, yes, good finance. But you kind of need the first 4 before you get good finance, right?
Yeah, because you need money, but yes.
So we agree on that. So I do think a lot of stuff is like, okay, content creation or outreach right now, whether it's for outreach for sales or outreach for recruiting, Or like ad creative. Now, where I think this OpenClaw is going, Neil, because Bernard and I were talking today and he's like, dude, this is the first version of AGI. I was like, you know what? This is like the MS-DOS version of AGI, right? It's like the very, like, you know, the 1970s, '80s where you're just using command line interface. And so I think you're going to be able to have your own version of OpenClaw that allows you to do all these things as one. And then you just have really smart people. Neil, the way I'm looking at it is that you have for deployed marketers, like for deployed engineers, and these people are working maybe full-time embedded into a company, and they're just focused on customizing these things, and these people know how to talk business, they understand the latest in marketing, and they're creating new skills and new workflows within like a mission control dashboard. That's where I think a lot of this is going. So I don't disagree with you, like a lot of it's being used for that, but I think a lot of it's going to become consolidated. Like you don't need all these freaking SaaS products.
I agree with you. That, that's a big chunk of the future, maybe not all of it, but I don't think it's going to be 4 people. I think it's going to be many more than that because what we see is unless you have people who are really good specialists, when they use these products, the output isn't as good. So you really need the specialist. And I know when you were saying 4, you're not saying it's going to be 4 people. You're just giving a random number.
Yeah, yeah, yeah.
I just want to clarify that for anyone listening because I already know Eric thinks the similar way to I do. We wouldn't want a paid Facebook advertiser to be using this technology for SEO. They're the wrong person. You're going to have a really mediocre output, even if they're amazing at Facebook ads. And if they're amazing at Facebook ads, but they haven't done much in Google ads, we wouldn't want them to be using this technology to help with Google ads. We want them to stick in their lane. The moment they start spreading outside of the lane, that's when you start getting mediocre outputs. Mediocre outputs don't work because that's what the majority of the competition is going to do, and you're not going to end up winning with that, right? And what we're seeing is we're running a test right now. I should end up knowing more, call it first week of March. All right, so my team set up a system where we're having AI outreach as me and set up meetings. At the same time, I am— which we've talked about— I am setting up my own meetings. And we're seeing what's going to end up driving more revenue by the end of, by sometime in March, right? At least from a pipe perspective, because we know based on how many meetings we get, based on how the first call goes, that after the first call, we usually know what's in our pipe that's qualified and what percentage will close. But we need the first call. Just the email intro or someone willing to take a call doesn't tell us enough, even if it's company size or right title. We genuinely need the first call. Sometimes we need 2 calls, but if we need 2 calls, it's because typically they need to bring in someone else where we weren't able to get all the information we need on the first call. So in that case, we still classify the second call, even though it's the second one, we still classify it as a first call. But so far, no joke, it's setting up way more appointments than I am personally, of course, because it's working more hours and it's A machine. The quality, no matter how much input we give and training we give, the quality is not there yet. And I want to use the word yet because no matter how many qualifications and criteria you give, it's not as simple as inputting 10, 15, 20 years worth of relationships from your head that's not out there in email or that's not out there on your LinkedIn or whatnot, and having the AI learn all of it. Like, it does take time, but I do believe if I kept doing it through the AI for, call it, another year, 2 years, which I know people think is a long time, but in reality it's not. Like, look how long you and I have been entrepreneurs for. It's like short-term pain for long-term gain. I do believe that output for appointments that it's setting, the quality would be much higher. But I'll quickly see what's going to happen within 30 days from at least appointment to close.
And Neil, just to clarify, you're sending 1 per day. Is that what it is?
Me personally? Me personally, I'm doing 5 to 15 a week. So I'm not consistently at 1 a day because it depends on how often I'm traveling.
Yeah. I mean, if you're traveling, you can do more, right?
No, it's really hard because like, let's say I'm in Mexico City. Morning meetings, morning event after that, then interview with Wired, then another speech at a university, then meeting with the university because we work with them. Then I had to go to Bloomberg for another interview. Then I had to do a meeting with my team to catch up with them. And then here doing this.
Yep.
So like, I haven't had much chance to work at all today. That's the hard part. When I'm on a plane, it's different. But if I'm not on a plane, That's what I mean, on a plane.
Yeah, not when you're on the ground. Um, quick break. One of our ClickFlow users just told me that they're saving 90+ hours a month on content alone. ClickFlow is our AI SEO product that actually does the work, not just here's what you should do. It does it. It creates content in your voice. It helps you with other areas such as finding the right internal links and creating FAQs. We spent years making it, perfecting it with our clients so it doesn't just sound like You can get a 14-day free trial at clickflow.com. If you're serious about AI SEO, you should check it out. Back to the show. By the way, Neil, I just, let's do this for fun. So what do you think the LLM IQ scores were on average in 2023?
100, 110, 90, I don't know, somewhere around there. Let's go 90 to 110.
Okay. Which is like, you know, that's actually, Average is 100.
Okay.
So, so get this and then let's, let's go a little more. 2024, what do you think it was?
I would go with maybe 120s and I would say today is probably closer to 130 to 140.
Okay. So pretty accurate here. Let me pull this up real quick because yesterday some of my team was like, they're just robots. They're just robots. I was like, are you kidding me? Like, so, and I'll explain why I said, are you kidding me? Because it's like, oh, like, We think just because we're the creator, we're going to be smarter than it forever. But check this out over here. No way.
It'll be smarter than humans.
Yeah. You're saying what?
There's no way it won't be smarter than humans. Like, humans will not be smarter.
We're on the same page. I'm like, it's just going to crush our intelligence, right? And so look at this, 70. Okay, then it jumped up to— 70 was initially 2023, and then GPT-4 was right in the middle of 2023, 2024, it jumped up to like 90. And then you go a little more to like early 2024, 100. And then last year, 2025, it jumped up to about 115 or so. And then now we're talking 2026, it's basically genius level, okay? 130.
Oh, close.
Yeah, you're very close, right? But check this out. So it's like, oh, it's doing this like linear projection over here. So it goes to like 200 plus. And I'm like, no, no, no, no, no, no, no. You can't do a linear projection on something that's exponential. So I'm like, give me an exponential scaling chart. It goes all the way to like 1,000.
So here's the hard part. For these models to get much, much better and smarter, it's so much more work than it was at the early days. I don't think it's exponential for them growing to that kind of IQ. I actually think you start seeing diminishing returns after a while.
I don't know enough. Neither Neil or I are AI scientists. I just heard on a podcast that one of the scientists was like, oh yeah, it's going like— I don't know if it's going to go to 1,000, but he's like, in a couple of years, it should be at like 300 or 350 or something like that. That I can believe. I don't even know how to fathom 1,000 because, yeah, so.
And I want everyone to keep this in mind. Having a really high IQ, whether it's a human or AI, doesn't mean that they're going to produce amazing results for you. And Eric and I can attest to this, or at least I can attest. I don't know what your hiring has been like over the years, but I've tended to have the worst luck with people who graduated from Stanford or Harvard or Ivy League schools. When you look at their IQ tests, they tend to be really high, just generally speaking. Would you agree with this, Eric? He's nodding his head yes.
Yeah, I'm sure. Yeah. Yeah. Yeah.
So, or shaking his head yes, whatever it is. Um, what I found is my hires from like Arizona State University tend to outperform the people— Eric's giving me the thumbs up. They tend to outperform the people in most cases who graduated from Stanford. Keep in mind, I'm not in a field like Elon Musk where I need all these computer scientists, so it could vary field by field. But when you look at general IQ, IQ tests test you on a wide variety of things. My kids have taken IQ tests. An IQ test doesn't necessarily mean you're really amazing at one specific industry or field. And yes, AI can learn a lot, but here's a problem, and I'm a big believer of Ben Affleck with this. So Ben Affleck and Matt Damon were on a podcast. I don't know if you saw it.
Yeah, yeah, with Rogan, right?
Uh, I think it was Rogan, and they were talking about the averages. I only saw clips of it, and they were just like, we don't think AI is going to be amazing at writing a script. We think AI optimizes for the average. You could say the average could be 136, 140 IQ, but that's just IQ. What is the output that they're producing for specific industry? Someone could be really book smart, but they could be a terrible scriptwriter. Someone could be a really amazing marketer, but they could be terrible at math. And what I'm getting at is they're scraping the internet, learning from everyone. A lot of the crap that is out there on marketing is junk, and they're pulling from a lot of that. And that's why when Eric and I were talking about how future organizations were looking like, and Eric mentioned 4 people, but he didn't really mean 4 people. What he meant to say, and what I'm trying to say is, It's going to be specialists. You're going to take really good people who are good at, let's say, Facebook ads, and they're going to be using AI to do a lot more damage. And instead of you needing a team of 5 or 6, you may be okay with a team of 2 amazing Facebook ad specialists combined with AI instead of needing 5 or 6 people.
All right. So what I'll say is, When it comes to Ivy League people I've either worked with or hired, they kind of— I'm generalizing here, but in many cases I've seen, they kind of struggle to get out of their own way sometimes, and they tend to overthink a lot of situations, which sometimes it's like, damn, you're really too smart for your own good.
Yeah.
And so I'm generalizing here, but it does happen more often than I see. And I think, I don't know, maybe it's just a pattern that we've seen in our hiring.
Dude, my favorite one is we've had a lot of Ivy League people in M&A, and with M&A, a lot of these people are amazing at Excel. Like, just, they're rocket scientists when it comes to just using Excel, and it is a skill, believe it or not. And what we'll find is like, they'll show us deals and they'll be like, Neil, this company right here, oh my God, if you bought this, the business is going to change. And they show us a spreadsheet of the synergies with the business, you'll be able to cross-sell this to these customers. They'll be able to cross-sell their stuff to your customers. You can just cut this fat right here, look at the additional profit. And they tend to struggle to take into account things like culture. Cross-selling doesn't always work that simply. Who is the person that is in charge of that other relationship? Are they even the decision maker for the service line item that we're discussing or that we have? Are they even part of that organization or part of that department, or is it something else? Is it in a different country? Like, there's so many nuances to M&A, and I'm like, you can't run M&A off a spreadsheet. And we bought companies that people told us like, ah, it's not growing that much, it's declining. I don't think it's a good company to buy, and they want too much of a multiple. And I'm like, I'm going to go buy the company. And people are like, you're telling me you want to buy this company that's not growing as much, they have quality problems, and you want to buy them? I'm like, they're one of the biggest in that region. They have a name. I'm going to go in and fix it. And they're like, why would you go through all that headache? I'm like, they have the brand recognition where they're getting hit up by most of the companies in the space. I can fix other stuff. You're telling me to go buy another company that's growing fast, that doesn't have the brand recognition, that doesn't have the right customer profile, but from financial, you think it's better. And I'm like, I'm telling you, to go get the right companies to hit you up because of your name brand is much harder to do, in my opinion, than it is to fix operations for someone like me. And that's the skillset of my team and I. So I go with the route that I think I have the edge on versus what a number cruncher is going to tell me.
All right. See you guys. Goodbye.
