How to Hack ChatGPT & Gemini Rankings (It’s Too Easy)

In this episode, we break down how easy it currently is to “game” AI platforms like ChatGPT and Google Gemini—and why those loopholes won’t last. We discuss the risks of short-term hacks versus long-term credibility, how AI will reshape job markets by rewarding bottleneck skills, and why strategy and adaptability will matter more than ever. We also share real-world wins and failures using AI in marketing and operations, highlighting where it drives massive leverage—and where it still falls short. 🎙️ Learn More About the Hosts Eric Siu – Leveling Up: https://www.youtube.com/@LevelingUpOfficial Neil Patel: https://www.youtube.com/@neilpatel 📩 Free Resources Growth Newsletter: https://levelingup.beehiiv.com/subscribe Ubersuggest: https://www.ubersuggest.com/ Answer The Public: https://www.answerthepublic.com/

• 0:00Speaker 00

  Finding customer information shouldn't mean digging through emails, call notes, and spreadsheets. HubSpot's starter plan brings all of those details together, so it's easier to track what's happening with every deal and what to do next. It pulls information from calls and emails. Then it gets to work, logging and summarizing them. Right after your call ends, HubSpot pulls next steps from your call notes and drafts suggested follow-ups. And it stays up to date for you, so you can build demand, win more deals, and spend less time on repetitive work. Your records, insights, and next steps are ready for you to act on. And you don't have to miss a beat searching for any of it. Try the starter plan for $10 a month. Learn more at hubspot.com. How you can actually hack ChatGPT's and Gemini's rankings. Are you ready for this? Yeah. You're probably not gonna be surprised, but let me pull this up. This guy over here, you see this is a, this is a hot dog.

• 0:44Speaker 00

  So this guy says, wrote for BBC, okay, he says, I hacked ChatGPT and Google's AI and it only took 20 minutes. So this guy, Thomas Germain, I guess the whole hack is, uh, he's the best hot dog eating tech journalist, okay? So this guy Jake Ward says, be this BBC tech reporter. Spend 20 minutes writing a fake blog post. Claim you're the world's number one hot dog eating journalist. Invent a fake championship to back it up. Watch ChatGPT and Google repeat it as fact within 24 hours. Realize you just manipulated two of the world's most powerful AIs with a single page. Find that users trust AI more than websites because it feels like the answer is coming from the tech company itself. Prove that tricking AI in 2026 is as easy as tricking Google was in early 2000s. What do you think? Yeah, dude, you and I know you, it's easy to game these solutions. But we're both believers that this is not gonna last.

• 1:29Speaker 00

  We think the algorithms are gonna get much more sophisticated. It's not gonna be in, you know, two, three months, but we think a year, two years from now. These algorithms are gonna be much, much better. And this is why Google released things like domain authority and looking for links instead of just mentions. Because they know getting things like that are really harder to, are much harder to manipulate than just looking at, oh, someone just mentioned you and they didn't link and it's joe the plumber dot com. And it's talking about marketing or medical advice. Well, what does a plumbing website know about marketing or medical advice? Yeah, so what I'll say is, I've always been a big pro of, it's easier to do the black hat things that work in a white hat way. So if you make a legit competition and you get talk- About in that way, that's gonna do a lot better because there's staying power there.

• 2:13Speaker 00

  Oh, and by the way, like early in my SEO career, I I've tried the gray hat, the black hat stuff. And I've I've experimented on my on my own websites when I was like twenty two, twenty three years old. And to watch those websites get torched. And so you don't want that to happen because it's the same story over and over. They have to protect the integrity of their product. And so you might as well just do the stuff that works, but do it in a more long term focused way. And you're gonna win over the people that like to do the short term stuff. Think of it this way, if everyone started creating junk and they were always mentioned in chat G P T and Gemini when they shouldn't be. What's gonna happen to the users? They're gonna use chat G P T and Gemini being like, the results suck, we're not happy. Google and Open A I are gonna lose revenue. So what do they do? Before it gets to that point, they adapt their algorithms to get rid of the crap and the spam. So then the results are better.

• 2:56Speaker 00

  People keep using their products more, they make more money. In essence, you gotta take the long term approach. Even if something works today and it could be black hat, you'd be like, well, at least I'm getting results for three to four months. What you need to realize, if you look at the old game of playing cat and mouse, especially when it comes to S E O. A lot of the websites that did funky shady stuff didn't just say, okay, well, it only worked for six months. They got put in this black box or in quote unquote, you know, internet jail in which a year later, two years later, even after they cleaned up their act. They weren't getting the results that they deserved because they were known for doing funky shady stuff beforehand. So you're better off having a clean slate and doing things the right way.

• 3:38Speaker 00

  Because the moment you do anything the wrong way in marketing, you don't know if that's gonna be used against you and hurt you even in the future four or five, ten years from now. I wanna switch gears here, I wanna get your reaction on this. So the former Uber founder, actually, I should just say Uber founder. So Uber founder Travis Kalanick said this, plumbers will be paid like LeBron James, you see that? I did not see that, but I just hit me when you said former Uber founder, I'm like, uh huh. That's still founder, still founder, you can't take that away. Once someone's a founder, they're a founder. So. Kalenix says this, like, everyone is like, and I've been guilty of saying this, I'll just own up to it, but everyone assumes AI eliminates human value, okay? So the physics say the opposite. So Kalenix says this, let's say the entire world, everything in our world was automated except for plumbers.

• 4:23Speaker 00

  You had machines making buildings. You would basically have like a thousand buildings a day. The algorithm can design a skyscraper in a millisecond. It cannot connect the pipes, okay? So when compute violently accelerates the speed of construction, the unautomated human becomes the ultimate bottleneck and the bottleneck captures all the margin. So Kalenix is saying, how valuable would those plumbers be? Extremely valuable. Those guys, each and every plumber will be paid like LeBron. Why? Because plumbing is a long pole in the tent to progress. So basically, he's just saying like, at the end of the day, if things go exponential, you're actually going to need more. If we're able to do more with more, why would we not do more, right? So their economic value goes exponential. You get so much efficiency everywhere else that you need millions of plumbers.

• 5:08Speaker 00

  The market thinks automation drives human wages to zero. The physics dictate that it drives the bottleneck's wages to infinity. The next decade doesn't belong to whoever out computes the machine. It belongs to whoever stands at the exact point where the digital engine meets the physical world. That's interesting. We already have that happening right now with all these data centers being built. I know people who are electricians that are getting paid millions of dollars. No joke to work on these data centers because there's a lack of electricians who are amazing and specialized in working at data centers. And they're making an arm and a leg. And the same goes like with marketing. If they automate most of the stuff in marketing, like content creation and keyword research, and the list goes on and on. But you're amazing at strategy and one of the best.

• 5:52Speaker 00

  Do you think that they're going to be happy with just taking what's average on the web? And then coming up with strategies for that and giving it to you? Or do you think they're going to pay you more being amazing marketing strategist? Because all the other stuff is automated and really easy and cheap for them to do? The stuff that isn't easy for AI to do is going to be worth a lot more and they'll spend a lot on it. I honestly think if you're a curious, hardworking human being, there should be no reason that you're not going to thrive in this world. And what Neil just talked about, these electricians making millions of dollars. By the way, I look at this Aman that's building every single day. I'm like, man, they must be making an arm and a leg because they work a night shift and a day shift too. I don't know if you know that, Neil. So the whole time, I can hear the construction the whole time. Yeah, let me let me see this.

• 6:37Speaker 00

  That building won't be done for like another six years supposedly. Probably not, yeah. But they're they're making fast progress on it. So Neil, have you seen this? So let me um, let me make this. Here, let me refresh this real quick. I have not seen that. I don't even know. This is the AI exposure of the US job market. Those of you that can't see on the screen. So these are three hundred forty-two occupations, okay? Colors, AI exposure, okay? So if you look at this, if you're red, that means your job is really at risk um to AI. And if you're in the green, that means these jobs are gonna be okay. So like, for example, uh, AI exposure for hand laborers and material movers, two out of ten, okay? So it shows like seven million people uh that did this job, median pays thirty-eight grand.

• 7:17Speaker 00

  And so, nor for, and it shows the education level required. So it shows basically like what jobs could be in trouble and what isn't. But like secretaries and administrative assistants, eight out of ten AI exposure, okay? Wow, really? The thing was eight out of ten? Yeah. I I don't think in-office uh secretaries and administrative assistants are that high of a risk. Cause like, let's say if you work at the JP Morgans, they're gonna still want someone to walk and greet and all that stuff. I I think it depends, but I I think we can agree on this one, right? Customer service representatives, nine out of ten? Okay? Yeah. Bookkeeping, accounting, and auditing clerks? I think so. Uh. Bookkeeping, we're not talking about like, we're talking, not talking about like tax strategists or financial strategy. I I I think bookkeeping, I don't know if it's a nine out of ten, but I agree it's high, maybe seven or eight, because you're still gonna have some humans just to double check.

• 8:01Speaker 00

  Yeah. So Neil, let's let's play a game here. Do you want me to go for, I can focus on the red, or you want me to hit the ones that are not at risk? Hit the ones that are not at risk. It'll be more fun. Electricians down here, look at this, two out of ten. I I was thinking home services, HVAC, plumbing, roofing. I think a lot of those are less impacted. Yep. So he, oh, right here, HVAC. Two out of ten. Childcare workers, yeah, I want a human. I don't want a like, security guard. I probably don't want robots being my security, because there's probably a lot of liability there, so. They get hacked, you're screwed anyways. Yep, yep. Uh, where's marketing? Five out of ten? I think marketing is pretty high. Let's just see. I think it's hard to find here. Let's just go with the ones under our highest risk first.

• 8:44Speaker 00

  So software developers, I think we can agree, the ones that don't adapt, that's what I would say. Data scientists, man, I never expected to see data scientists so hard. Like, I would just say that, let's caveat this, the ones that are really good at strategy, which, which we just talked about, those people are gonna be okay. We agree on that, right? Yes. Okay. So I think marketing's like five out of 10 or something like that, but. Yeah, and it gets tricky too, because what kind of marketing role, right? If it's just hitting up people for negotiating Instagram sponsorships, I think it's more closer to eight or a nine out of 10. If it's something like you're focusing on, you know, strategy for international expansion for a company that's looking to grow faster, I think it's pretty low. Do you see marketing here? I can't see. Oh, advertising right here. Advertising, promotions, and marketing managers, eight out of 10.

• 9:29Speaker 00

  Yeah. I think it really depends for all of these, including the role, just like the secretary example I gave, like, I don't think banks getting rid of them. Same thing for top executives too, like what executives are talking about CEOs, like, so. Yeah, yeah, you nailed it. Like, if you're saying top executives, I don't see AI replacing any of my top executives. If it does, it's because they sucked and they couldn't adapt to the AI world, and they're gonna be replaced with a new executive that is adapting. Yep. Same with like engineers, I know they said it was, I think you said eight or nine out of 10. I. We're not actually looking to replace engineers and get rid of them. Yep. We're looking for engineers, at least at my organization, who are great at using AI and can move faster.

• 10:09Speaker 00

  And the ones we're getting rid of aren't using AI, can't adapt to it, and when we get rid of them, we're replacing them with a new engineer that can use AI. So instead of the model of like, oh, really? You need two engineers now? On our end, we're like, no, if we have 20 engineers, we still want 20 engineers. We just want 5X the output and just do so much more so we can grow faster. I found the most effective thing when it comes to hiring now is just to have, like, when they get to me, all I talk about is AI. I want to charge, just, I want to have an exceptional conversation for 30 minutes about AI. Because if you knock out AI, you knock out all of our core values, and you just, you leverage. Yeah.

• 10:45Speaker 00

  It's like you can do 10x more. Why would I not try to go for you all day? Right? It's the people that say, whoa, I'm really interested in learning right now. Right? Can't do it. So. I'm interested in learning right now. What happens six months down the road? It was that's past now. I'm not interested in learning anymore. Yep. Can't do it. All right. Well, anything else from your side? Dude, you want to hear AI failure before? Yes, yes. Okay, so for shits and giggles, we had AI help plan my itinerary, flight time, when to come, when to go. It did a good job telling me when to leave, to come to London. The bad part was, and this wasn't AI's fault, my American Airlines flight was three hours plus delayed because of some plane issue.

• 11:27Speaker 00

  They couldn't get the air conditioning working. So then we had to switch planes. That's not AI's fault. All right. When it planned out my schedule, we told them how long I normally speak. Cause we, I, I wanted to see, uh, remember that Elon thing that he talked about a long time where his assistant wanted a raise? And then he said, okay, go on vacation. Let me think about it. And he wanted to try doing it all on his own. My assistant did not ask for a raise. She's amazing. I'm not going to replace her. This was just an experiment. When it came to the schedule in between, it started putting large blocks of time, including the travel time, which is fine. But it just overinflated how many hours? And if I had a guess on how much, it put roughly 60% more time than needed for most things.

• 12:11Speaker 00

  Now, of course I can keep fine tuning it and it'll get better in the future. On the way back, we ended up asking when I should fly and it straight up picked a nighttime flight when I'm done at nine 30 in the morning. So I was so pissed on that. That one mistake, guess how much it cost in the change fee to change my flight? Two thousand bucks? No, $7,000. Oh, wow. Wow. That's, that's, that's a, oh, wow. That's a ouch. That's why I wouldn't rely on it right now to deal with stuff like that because that stuff drives me insane. So I'd have too much anxiety to hand that off right now. I was pissed. So I did change my flight. I use Amex points because on the Centurion card, if you book it with points, you get 50% back.

• 12:54Speaker 00

  But I was pissed and then I got the refund on my existing flight and I'll use that Fright credit for a future trip because I fly like every other week or every week. So it's not a big deal, but I was still pissed because I just wasted a ton of points on something that I didn't need to. Because AI recommended a flight back and it was a worse flight. It was like I land pretty much at midnight. And AI knows I have, I think it's called private suite, you know, in LAX. Yes. Where they just pull you off the plane. You don't have to go through immigration and then car picks you up there. So I save time, right? It knows I have that. It's all inputted in there. It didn't tell my team to book it or any of it. It was to say, this is ideal time, traffic is lower and we inputted this stuff in advance and it just really shat the bed.

• 13:36Speaker 00

  Yeah. I think the key takeaway here is you have to be careful on like what you actually wanted to like. Funny enough, even though I talk about AI all the time, I don't let it touch that stuff, Neil, because that stuff would just drive me nuts. I had a follow up point to this one, but I forgot what I was going to bring up with this one. But I think we talked about this actually from lunch last week. It's like, I think we're going to see a lot because of 11 labs and how good these voice models are. They're going to see a big rise in robocall spam, texting as well. We're going to see a lot more of that too. I would just say that with the outreach stuff that we have going on right now, you'll have a lot more experimentations running for you. So the math that you can do on this is like, if it's a good growth team, human growth team, maybe you're running 50 experiments a year as a growth team, right?

• 14:16Speaker 00

  Obviously each individual is probably running more experiments, but if you have these things that's constantly running experiments and you're driving enough volume. If you can run a hundred experiments a day or something like that, I'm just making numbers up. If you can truly run 36,000 experiments or micro experiments, how much faster is your company going to grow? It remains to be seen, but without Karpathy is talking about how quickly they're doing AI research. My hope is that can carry over into marketing, which is why we're doing it that way. So we'll see what happens. Yeah, I'm just hoping they can fix the little issues with these LLMs. Like, you know, when you want to get cited, a lot of companies are paying for articles and paying for placements. On other websites and they'll talk about how they're the best solution for X or the best product for X and it's getting excited by more by chat GPT.

• 14:56Speaker 00

  The problem right now that you're seeing with chat GPT and Gemini and a few of the LLMs, they can't decipher if the content that they pulled was paid or not, shockingly. I don't know why they can't. Yeah. And it's causing a lot of people to pay and get results in organic when they shouldn't be able to get those results. There's a lot of nuances like that, similar to the example I gave you when I'm asking for specific job titles. It is pulling job titles, but that person could have had that job title in the past, even though we said to tell it so many times, no, it needs to be its current job at the current company they work for title. It just messes up on little things like that. Yeah, okay. The sales workflows that we have right now, Arrow is our, our sales agent. So it will surface cold emails to target, like cold companies, like based on our ICP.

• 15:41Speaker 00

  That's one that's not necessarily new to anybody. But the second one is deal reviver or deal resurrecter. So obviously, you know, you're going to lose deals, right? But oftentimes, if you come, if you're coming in number two, you just have to find people at the right time. Maybe, you know, in three months, six months, again, you can resurface these. So I have a deal resurrecter. But we also have one, Neil, for any inbound contact that comes in, whether they come into the email list, maybe they watch one of the webinars. What happens is it will try to see if they're an ICP fit and then reach out with a custom message as it relates to their role. And something that, that like some type of custom, like personalization in the beginning, right? So the whole idea here is now, now that you have these claws, you can build these different sales workflows. And then that enables your team to get more done with less.

• 16:23Speaker 00

  So anyway, all that to say, guys, that is it for today and we'll see you tomorrow.