Eric Siu and Neil Patel discuss how AI is reshaping paid media, from Meta's acquisition of Manus to layoffs at companies like Block. They explore whether ad-platform AI can replace human creative strategy and propose a pay-for-performance future powered by AI-enabled services.
Chapters
So I have this pulled up over here. So this guy's, uh, his name's Zach. So the Meta AI situation is bigger than anyone realizes. This is what he's saying, not what Neil and I are saying. Millions of media buyers are about to get left behind or even fired. Meta just paid over $2 billion for a Singapore-based, uh, AI company that launched in March 2025, making it one of their largest acquisitions ever, right behind WhatsApp. Um, and if you think this is just ChatGPT with extra features, you're completely missing what's actually happening. There are 2 types of people right now. So I'll just give the 2 types of people. People who think AI is a tool that you use when you need help asking questions and copying answers into your work, and people who understand AI as an autonomous worker that executes entire workflows while they're literally asleep. So let me explain. So Manus now, so Manus is a general agent that Meta bought. Manus can now log into ad platforms, analyze campaign performance, make optimization changes, and handle customer support by monitoring inboxes and responding, all without human intervention. And I think this is where we can talk about our reaction.
Yeah, no. So here's the big thing with Facebook and Google managing your ads, okay? And this person, just to be very clear here, they're talking about Facebook managing your ads and they're talking about in the Facebook portal because of that acquisition that they made. Facebook and Google, they've had AI technology for a long time when it comes to ads and they've been trying to improve it. Mark Zuckerberg for a long time saying, why do agencies even need to exist? You have to keep in mind that what they optimize for is not always what you're optimizing for. They're optimizing, yes, for you to get great results, but they're also optimizing to maximize their own company's revenue. And sometimes those 2 things fight each other because what's best for Facebook may not be best for you, just like what's best for Google may not be best for you. Such as if you throw a ton of ads in Gmail, you'll see it doesn't convert that well and you can burn a lot of money really quickly. I'm not saying you can't make Gmail ads profitable. I'm just saying if you let the AI run wild, a lot of times you'll burn a ton of money in Gmail ads and other channels that aren't the best versus someone typing a keyword into Google, clicking on a paid ad and just buying. So the way I look at it is you still need humans in the loop. For a long time, there have been tools that help automate and make media buying easier. You can call it, oh, we have AI now, so it's much more sophisticated. They've been using machine learning and a lot of cool tech to make it easier to buy ads, and they have. But the real secret to doing really well from buying ads is not actually how you buy ads. Yes, I said before you need some humans in the loop, but what it really comes down to is a creative component. Yes, AI can help you with some creative ideas, but the real secret is doing something unique and trying new stuff that no one's tried before. And it's a big hit or miss. It's a big experimentation game. AI can help you run the experiments faster, but once you find the creative pitch, that converts really well, that's not been tried before in your space, you double down and you just go really hard at it, but you have to kiss a lot of frogs to find the right one that just works well.
Yeah. So my reaction to this in regards to if paid media is over or is it over for paid media marketers, I think the paid media marketers who don't adapt, right? So just the button pushers alone are going to be in big trouble. That's my opinion. And I think if you're a paid media person, you're going to have to become a different I think a lot of roles in marketing are converging now, right? And so Satya Nadella from Microsoft actually said that, look, if you're on the product side, you're actually kind of converging with designers, you're converging with engineers at the same time. So a lot of roles are going to start to converge. And so Taylor Holiday actually posted this to X talking about head of growth, media buyer, creative strategist, retention strategist becomes profit engineer, right? Whatever that means. So my whole point is, I think if you evolve, like, yes, paid media is still important. Yes, you still need a human in the loop. You can't let these things fully automate. But again, you know, and we haven't reacted to this yet, but when you look at Block cutting 40% of staff, that's a lot of people, right? You don't just cut 40%. And there's probably some, you don't get it right too. When you cut 40%, there's probably some really amazing people that you probably shouldn't have cut, right? But I digress.
Yes.
My point is, I think, and Neil, curious to get your thoughts. I think if you're just like, if we did 5 to 10 years ago, if you're just a button pusher paid media, you're actually, you know, pretty valuable. A lot of agencies needed to hire you, right? I don't think that's going to be the case anymore because that actually means if you're just a button pusher today, you actually haven't evolved. I think if I was working in paid media completely today, fully immersed, I would try to learn as much as possible. Hey, how can I start to learn the AI creative piece? How can I learn to start to make creatives and how can I learn to make creatives at scale? Because it's only going to get better over time. Or if I'm an SEO, I'm going to try to really learn the AEO side of things. Like how do we learn Reddit as well? How do we learn YouTube as well? I can't just be a one-trick pony, right? So I think I think one-trick ponies are going to get shot. I think the ones that adapt, the ones, the roles that converge, those people are going to not only get paid more, but they're going to be in more demand. Yeah.
And by the way, the media buyers that just push buttons and weren't doing anything amazing, I think a lot of them weren't doing well from a job perspective for many years now. I don't even think a lot of that's related to AI. I think for a while now, a lot of these roles when people are just clicking a few buttons, they've been losing contracts if they're agencies, or they've been eventually getting replaced because people are like, you're not adding much value because technology has made a lot of this stuff easier for years. I don't think anything's really changing. It's just accelerating. It's the same stuff just getting accelerated because technology's getting better. And when companies are laying off a lot of people like Block laying off 4,000, and I spoke at their— I have 40%, I think it was 4,000 though. I spoke at their internal marketing conference. I can't go into discuss what I spoke about because I signed a contract. But when you think about AI and Block, am I going to say that AI is not causing any jobs to be replaced? No. Am I going to say that AI is a reason for 4,000 jobs to be fully replaced? No. I bet you there was a ton of wastage in there. If you just look at X, okay, a company that Jack Dorsey created, or Twitter, And Elon ended up buying it. He canned a lot of people and they've seen growth in usage with a much smaller team. This was pre all this AI stuff. A lot of these large corporations have been bloated for a long, long time and they need a cut and many of them need a cut even more.
Yep. So let me give you the numbers around this deal. So you like numbers, right? So Block was, uh, $24 billion in revenue and 24% gross profit growth. Okay, just cut 4,000 people while raising 2026 guidance to $12.2 billion in gross profit. Stock ripped 20% after hours. The market added roughly $6 billion market cap. So that's $1.5 million in enterprise value created per eliminated role. Now, maybe that's not the right way to look at it, but I would say that, you know, it's— there's a handful of things. I think one, I think the way Jack Dorsey handled it, if you look at his message, was handled really well. Like you do 20 weeks of severance plus for every— I think it's an extra week for every year of tenure. Plus like you get to keep your equipment. There's like insurance and all that. So he's, And he's saying, look, the reality is like something has come and artificial intelligence, sure, you can use that narrative. I think part of it is, is to AI, but I think the other part is you look at 2020, 2021, I think people just overhired, right? And so I think part of that is still kind of undoing that.
Totally agree. I think the biggest issue is people overhired. Like a lot of these companies don't need that many employees. And I look at some of these companies that are publicly traded in the SaaS space, they're barely growing. And it's just like, you haven't really released too many new features and product updates, yet you have 500 people working on product and engineering. And I'm like, I would fire a lot of those people if you haven't really made any major product changes in the last 12 months. Like, why do you need 500 people on a team? Forget AI, 500 people are just really inefficient. There's too much bureaucracy. Get rid of a lot of them. You can probably get more done with like 50 people.
Just a quick break. If you want to run LinkedIn personalized ads and landing You can't do it on LinkedIn ads right now, but you can do it with Carrot. It's www.carrot.ai. Again, carrot.ai, and we'll see you on the other side. Yeah, the bureaucracy is actually something we can talk about, Neil, because I mean, just sitting in that office yesterday, I can just tell that, you know, people are really inundated by the bureaucracy. And so, you know what's interesting? I kind of got called out because one of the people that was the VP at that company used to work at my company, right? And, you know, for example, Neil, let me just ask you, when you have your metrics pulled up, okay, when a metric is, you know, you look at traffic lights, red, yellow, green, red means like it's really bad. Yellow means it's kind of, you know, it's not doing well. Green means it's moving in the right direction. You're going to focus on the red and yellow things, right? So let me ask you, Neil, Neil Patel's way of looking at this. If something's red or yellow, when do you address it? Do you wait a week, 2 weeks, 3 weeks for you to ask for a remedy? Like, what is your stance on it?
Well, before I see the report, I hope they already told me what they're going to do to fix the reds and the yellows.
Okay.
So there's not really a waiting. It's more so my team, when there's problems, they don't come to me with the problems. They come to me with potential solutions. Some solutions I love, some solutions I don't like, some solutions are whatever. So then the solutions that I don't like, we'll brainstorm and try to come up with better ones. The solutions that are whatever, Again, we'll do a similar process there. And it's not always me with the best ideas. There's a lot of other people who have better ideas than me. It's just more so as a collective group, we try to be proactive about solving problems. I think just looking at things like red, yellow, green, and if someone doesn't want to come to solutions and they just want to complain about the reds, and it's like, all right, this is just wasting everyone's time.
And what if they're like, Neil, I've tried coming up with the solutions, but I kind of want to hear from you what solutions you have in mind. So I've tried it a couple times already, it's not working. Neil, you know, please, what are your ideas?
So I would say, what were your solutions that you came up with? You must have came up with at least one. You could have at least typed it into ChatGPT and asked them to help you come up with the solution, even if it's a terrible one. I just want to hear what you were thinking. And if you don't have an actual solution, what direction were you thinking in? Like, You know, what made you get stuck, right? And in all reality, I'm kind of being nice, but I would just want to straight up know why didn't they come up with any solutions and why didn't they even try to ask AI or a friend or a colleague? And I would just be pissed.
Yeah. So at this company, look, so yes, yes, totally agree with you. I think my flaw calling out in public is, you know, we do the traffic lights, which I actually got from the great book, The Great CEO Within. The problem is, I think one of the things that's flawed with that book is saying, hey, when something is red or yellow, you should wait 2 or 3 weeks, right? And so I followed that for a little bit.
2 or 3 weeks? That's what they tell you in that book?
Which is actually 2% to 4% of the year, right? Which is crazy, right? So I was doing that before. And then one of the guys called me out. He's like, yeah, when you did that, that was really stupid, right? I was like, yes, it was absolutely stupid. But I think the key thing for everyone listening to this is that if something's red or yellow, I think you've got to keep applying pressure or you're going to feel it, right? Hopefully, like, yes, if you're going to your boss, for example, come with one— come with actually 3 solutions, pick the best one, right? And then go back and forth to prove that you're proactive with it. But waiting it out and being okay with it and yet letting the scorecard be red or yellow, that is a nice way to stunt the growth of your company.
Yeah. The one thing I love about hiring is if you look for personality traits of people who are like, obsessive. They pay attention to details. They stress out a lot and they worry about things. Those people usually don't wait for the red lights. When they see something that's yellow, they try to fix it right away or figure out a solution. They stress, keeps them up at night. Yes, you want them to have a work-life balance and all that kind of good stuff and not have issues with their family, but Those people naturally just perform well from what I've seen, assuming you're hiring people with experience as well. And then with those people, you don't have to push them to solve the reds or the yellows in marketing or business. More so, you got to force them to unplug a little bit and go spend time with their kids or their significant other.
Dude, so Neil, you might not know the answer to this, so if you don't, I'll go ahead and fill in for you and I'll tell you what we do. But What do you guys do to kind of figure out if they have that, maybe obsessive is not the right word, but like, you know, they can't help but—
Yeah, obsessive personality.
So what are you guys doing to filter for that during the hiring process?
It's just interviewing. So when we're interviewing, we'll ask specific tasks. Okay. So let's say if I'm interviewing for a product role and I give someone a design, all right. And I would be like, How would you make this flow better? I would look to see where they designed some elements. Did they match the style guide? Did they ask me for the style guide, even if I didn't provide it? Because I'll tell them, here's what we're looking for. Let me know if you need anything from me. So we'll start looking for people if they use the right color of orange or if they use our right fonts and all this kind of stuff. And if people don't pay attention to those little details, and obsess about them, then we usually know that they're not the right fit. And that's an easy example for product because it's more black and white. With non-black and white roles, sometimes in management roles, you'd go ask them very specific details like, hey, client was upset, X, Y, and Z happened. You gave it to an employee, didn't get resolved. What do you do? And I would have them break down not just scenarios like that, but I would have them give examples in the past of working with the team where things didn't go their way and what they did.
Yep. So what Neil's talking about, and then we screen for this, uh, it's looking for that bias to action, right? And the very first question I actually ask, and this is part of it, it's not everything, but the very first question when they get to me during the interview is like, how did you prepare for this call, right? And I'll tell you what, the, the Beat Claude guy, the AI guy, he's like, oh, I read your book. I was like, oh, you read my book? I was like, Not many people do that, right? Like you read my book to prepare for the interview. Like, that's cool, right? But that's a bias to action right there. And not everyone has the same demeanor around it, but that goes back to our core values, right? So ours is bias to action. We want relentless learners. We want people who make bold long-term bets, unreasonably resilient. But one thing I did find out, Neil, so you've seen the show Billions, right?
Yeah.
Okay. Wendy Rhoades. So do you want to describe what Wendy Rhoades did for a living?
Sure. So Billions is a TV show about a hedge fund manager named Axe. And Wendy Rhodes, she was their in-house psychiatrist or shrink or something like that. Something like that. And she would help people mentally get in the game so they can be a top performer. So let's say you have someone, and I'm going to give you guys a marketing example that's all relatable. You have someone who is doing affiliate marketing for you internally. They're figuring out how to arbitrage ads and make more money. You know, for other businesses and you're collecting affiliate revenue from it, or they're figuring out how do you get you press on a lot of sites and, you know, adding in affiliate links and then convincing them like, hey, add in affiliate link, you can make more money and push our products and adjust the text instead of just mentioning our name, which in many cases doesn't do anything because there's very little context there. And sometimes they're just off on their game and they have some personal issues that's slowing them down on work because it happens to a lot of us. And Wendy Rhoades, in this example, the shrink or psychiatrist, she would figure out what's blocking them, whether it's personal or work-wise, help them resolve it so then that way they can go back to performing well.
Yep. So that was a very good explanation. I didn't expect you to go into that much detail, especially with it being 10:18 PM for Neil. So here's the thing. The reason why I just asked, why I'm talking about Wendy Rhoades right now is because I actually posted something to Twitter just saying how Whatever you're using, like OpenCLAW where it has memory, when you work with it over time, you actually can have it do a complete personality test diagnosis of you. Okay, so I had to do a diagnosis of everything. But the Wendy— so there's someone kind of known as Wendy Rhodes, like she's like the Wendy Rhodes type on Twitter. She responded and she's like, hey, all the personality tests you mentioned, like they're fine, right? But you should use this one. So you should try this out too. It's called the MMPI. So whatever LLM you're using the most, Just say, hey, you work with me a lot, score me on the MMPI. It's actually a 576-question personality test. And I ask it to kind of call out patterns, right? And the reason I'm giving this to everyone is like, I think you should do this with whatever LLM you're using. So look for the MMPI. But what I'm going to test doing now, Neil, is whoever we're talking to during the call, I'm going to say, hey, pull up whatever you're using, ChatGPT, whatever, ask it about your MMPI, whatever you're working with, and then tell me what it says. And that's actually where it can show your neuroticism and all those things.
Go ahead, Neil. It's Minnesota Multifaceted Personality Inventory or something like that?
I don't know. I don't even know what it's called.
The Minnesota— sure, real. It says it's a licensed clinical test that's given and interpreted by a trained psychologist. It goes on and on, but pretty much it says I can't actually administer or score, and then it gives reasons why.
Okay, so check this out. Tell me if you disagree or agree with these things, right? These are the main outliers for me, Neil, and you tell me if you agree or disagree. Are you ready?
Yeah.
Okay, so hypomania. So Scorpion hypomania, this is on a scale of 0 to 100. Mine's a 72, which is— it says significantly elevated. This is your defining scale: high energy, rapid ideation, parallel work streams, impatience with slow people, a tendency to start more than you finish. You run at a pace most people can't sustain. The 8 to 7 Enneagram map directly maps directly here. Not clinical mania. This is a trait-level hypomaniac drive that builds companies. That's number one. Do you agree or disagree with that?
Agree.
Okay, number 2, um, is psychopathic deviate. Okay, I'm high on that. That's, uh, 68, right? Um, so it's elevated. Um, this scale catches rule breakers, authority challengers, and people who reject conventional paths. You dropped out of the corporate track, built companies your way, run hiring through Beat Claude challenges instead of HR playbooks. High PD in entrepreneurs is extremely common.
Agree or disagree? Agreed.
Okay, so let me tell you where I'm going with this. So it spotted a pattern because it's, it's my Claude code, right? So I was like, okay, It's like your specific pattern. Look at your history. So 2014 to 2016, energy, survival, hustle, it works out for you. 2018 and 2020, you install a GM, enjoyed growth, like it kind of worked, right? 2021 to 2023, spread across multiple different projects, acquisitions, drift, revenue drops. Starting in 2024, reengage, performance rises again. So it's like, notice the pattern when external friction drops. So bull market, high liquidity momentum, you expand aggressively. When reality constrains you, you focus. You don't naturally self-constrain. And that's the core, right? And so my point of calling this out is like, it's pretty spot on. And so during the interview, Neil, I do think it's hard to just kind of call it out with someone. Like, you're gonna get different answers. And like, you might, you might find some person more likable than the other one. But if you're able to get this out, I think, and you're able to look at those traits, maybe you can dive in a little more. That's how I'm looking at maybe approaching this in the future with hiring. So this is more of an experiment. But I find it fascinating. And I was like, hey, Give me a one-liner on this, um, on how I am so I can tell everyone that works with me like how, how I work, right? So the one-liner is, I naturally over-index on building the future and don't generate enough internal friction around margins and metrics. So without structured accountability, I can let operational discipline drift while chasing expansion. I was like, damn, that's pretty good.
So yeah, I had the same talk with you a few weeks ago. I said you're spending a lot of time on AI, which is the future, And I'm like, the fundamentals is what's going to grow. So you just need both.
Well, what Neil said, I think this was 2 weeks ago, Neil was like, this is post-call and I'm happy to share this. Neil was like, I think this could be a distraction like Clubhouse.
Yes.
Yeah. So anyway, let's move on.
And we were talking about, what was it, OpenClaw or something like that?
I was talking about OpenClaw, Claude Code. I will tell you though, from that retreat last week, I'm not going to name names, but the people who are using Claude Code are like, they're really innovating. The people who are on ChatGPT, like someone showed like custom GPTs and all that, right? The people that were already on Claude Code and Open Claude, they're like kind of past that. But the ChatGPT people were like desperately trying to catch up and they're like, oh crap, I'm really behind, right? And so I do think my belief on this, and we probably disagree for now, is that I think the founders really need to be all in on this stuff. Otherwise, the company can't move as quickly as the company needs to move. So in my mind, I'm like, this is the fundamental because Stuff is changing too quickly. So anyway.
Yeah, but to go back on that, I actually agree with you on which, whether it's AI or anything, if your leadership and the founders aren't on board, it's very hard to get anything changed from an organizational standpoint. You need the person at the top to buy in and push it because if they don't buy in and push it, others will not adapt. And I think you end up losing the race. But at the same time, you just have to keep pushing on the fundamentals that are driving the business until you can figure out how to get growth from the new quote unquote vision.
Yeah. Which is actually interesting, Neil. I'll kind of give out the vision, the reason why I was even in Arizona yesterday. And that was like, I woke up at 3:30 AM, flew there, was in the office at 9:00 AM with my CTO, and then we flew back. And by the time I flew back, it was like 4:30 PM. I was back in my home, right?
Wait, wait, wait, let's go back a little bit. You went to Arizona for what? Because your office is in LA.
Yeah, so we went to Arizona because I'll just say, I'm not going to name names right now. We'll keep it anonymous, right? But you would know. Went to Arizona because we're basically creating an offer right now where we are basically going to be driving phone calls or leads for one of their businesses. They have a new SaaS product, let's just call it that. Okay. And that has a lot of AI that's driven into it. And so basically a lot of the scaffolding my CTO and I want to do. So we had a lot of questions for their executives. We also had questions for their people, boots on the ground, to understand what the problems were. Right. So that's almost like doing customer development or customer discovery. And so what we're going to do here is we are going to, you know, whether it's Sending, whether it's making phone calls, or whether it is driving outbound emails, right? Or whether it's making an agent that fulfills the call, like the sale end to end. That's what we want to do, right? And then that is actually the vision of the future where we want to be able to build these solutions for customers, right, called Ford Deployed Marketer. And then we have the smart marketer managing and customizing it for the clients, right? And so that is, that's why we had that conversation. When we started having those conversations, like, oh man, like, They might have a lot of people, but they're just like, not enough people are AI-enabled and they just can't do it. So that's where we are with that right now.
Your meeting in Arizona reminds me of Ethan's company, TinyLab.
What does TinyLab do?
AI deployment company. So they go in 7 days, they figure out whatever problems you have in 7 days, they try to deploy as much AI as possible that solves the problem literally within 7 days.
Yeah, we're not doing that. So the model's a little, it's all pay for performance.
Oh, so it's all pay for performance, but it's similar in which you're fixing some of the problems.
Yeah, exactly.
'Cause some of this that you're mentioning isn't all marketing related, right?
No, it is, I would say it's more like, we can call it enterprise revenue agents, right? If you wanna call it that, right?
It's more like business consulting, but we're using it through AI.
We're basically building revenue solutions to solve all their problems there.
So we're not really consulting much, like there's not gonna be phone calls, like back and forth reports, No, no, no, but like, so when people think of consulting, like if you go back to Accenture, Accenture, when you pay them, it's not just they come with a strategy, they will go and build the product, solve the problem, and help deploy it. You're pretty much going in there, you're doing the consulting part from figuring out the problem, and then you're coming up with the solution, you're building the solution, and you're deploying it.
Correct, so our vision, even last year, towards the end of last year, we had planned this out already, that the 3-year vision is tech-enabled services, and this fits in really well with it. And the only way this is going to move forward, to your point earlier, is like the founders have to be involved. And so that's why we flew in last— like, I hit up my CTO Monday, I was like, we're going to Arizona on Thursday. Um, and like, literally we just got up and went. So, um, that's where we are with it. I do think in the new world that we're moving into right now, that it does enable a lot of interesting models with pay for performance because pay for performance before, um, it's a lot harder, especially when you have, um, a lot of hard costs with people and everything. But if you're able to build a lot of scaffolding with these agents and able to— you're able to make it work, you can actually repeat it. for different verticals or different companies, right? So that's how we see it and we'll see where it goes. But at least we'll kind of— I'm giving you guys that model that you can work with. Yep. Guys, thank you for joining us and we will see you next time.
