Stop Teaching Your Team AI. Keep the Top 40%.

Growth Newsletter: https://levelingup.beehiiv.com/subscribe Need marketing help? Visit: https://www.singlegrain.com/ and https://npdigital.com/Want to recruit great marketers? Find them here: https://marketingschool.io/hire Eric came back from a founder retreat with a hard truth: the AI hackathons and training programs everyone planned in January didn't work, and the companies pulling ahead just raised the talent bar. He walks through the GitHub leaderboard that shows who's actually shipping, why one 200-engineer company is keeping only its top 40, and Neil's long-held view that you can't train mindset. Then the sycophancy trap: why ChatGPT agrees with whatever you feed it, what Neil found when he ran a real recommendation test across an accounting firm, and why most executives still don't understand how these models work. Key takeaways◾You can't train mindset. Raise the bar and hire people who already have it◾A public shipping leaderboard exposes who's really working in about a week◾AI is a sycophant. Optimize for it, don't just ask it follow-up questions Chapters00:00 Stop training, raise the talent bar01:14 The GitHub leaderboard: keep the top 4002:31 Neil: hire amazing people, it solves most problems07:45 The AI therapy trap: it agrees with you13:17 Neil's accounting-firm recommendation test15:54 Three years in, executives still don't get it

• 0:00Speaker 00

  Finding customer information shouldn't mean digging through emails, call notes and spreadsheets. HubSpot 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. So Neil, we're planning out this founder retreat that we do every year and I was talking to this group.

• 0:42Speaker 00

  And in the beginning of the year, we're like, you know, we're gonna go hard on these AI hackathons, these transformation, all these things. Everyone, at the end of this, this is September, this is September now, we did this retreat in February. We're like, F all that, none of that works, okay? We're just like, just unreasonably high, just talent bar now, and you get everything that you need. And the thing is, one of the guys we're talking to, let's say he has like 200 engineers, for example, right? Now, if you think about it, how many of the 200 do you think are acceptable from a talent bar standpoint? What percent?

• 1:19Speaker 01

  Depends on the company. If it was Anthropic, it would be a very high percentage.

• 1:22Speaker 00

  Yeah. But you know, most of us-

• 1:24Speaker 01

  What kind of company?

• 1:25Speaker 00

  Most of us are in software, right? Software, services, whatever. So it's like-

• 1:30Speaker 01

  How old is the company?

• 1:32Speaker 00

  No, you're able to pinpoint, I'm not gonna say anything.

• 1:34Speaker 01

  You can give me a random industry if it's in-

• 1:37Speaker 00

  Let's say this company's been around for 15 years.

• 1:39Speaker 01

  15 years? That means they probably don't fire a lot of people, a lot of people with them for a long time. Maybe they're happy with 5-10%?

• 1:48Speaker 00

  Okay. It's actually much higher. About 40%, okay?

• 1:53Speaker 01

  Okay.

• 1:54Speaker 00

  So, but, so someone else on the call was like, well, like, okay, so what are you gonna do with the rest? Well, and then, so what happens is, like, I'll give you an example. For us, in our Slack, we have a GitHub leaderboard, okay? It shows how much someone is committing, how much someone is merging, how much someone is reviewing. So these are all GitHub terms, it just shows how much you're pushing to GitHub, right? And so he has it for the engineers and it's a public leaderboard, okay? And you can see who's shipping a lot, who isn't. And imagine, if I'm an engineer that works for you, and I have like 2,000 lines of code in a week and I have no commits and no merges, right? That means I'm not really that active. And so he has a lot of people that are like that. I can see kind of something that are similar with mine, so, and so like, what are you gonna do?

• 2:39Speaker 00

  It's like, you know, there's no, you might have to keep the top 40%, which are the good ones, but you might have to keep like another 20%. But the rest, you're gonna have to let go, right? And so anyway, that's, it's just interesting how we started the year like, oh, we're super like, you know, we're gonna teach everyone this stuff. And Neil's always been a proponent of, no, you're just not gonna teach them. And I think everyone's waking up to, trying to teach someone mindset, train them on mindset, is not something you can do.

• 3:04Speaker 01

  My big take is, whether there's this new technology or not, I've always, for many, many years, just said, hire really amazing people and it solves most problems. I did not come up with this. Venture capitalists from all the major firms tell their entrepreneurs the same thing. That's actually how I learned it. Wasn't through reading an article from A16Z or anything like that. When I raised money for Kissmetrics back in the day, Tony, Phil Black, who was one of the co-founders of True Ventures, they did, I think, I know they did WordPress, I don't know if he was a personal investor in Uber, and they did quite a few other deals.

• 3:45Speaker 01

  I don't know if they were in Slack or not, but I know they did like Blue Bottle Coffee, they had some amazing hits.

• 3:53Speaker 01

  Would always tell me, your number one job is just hire really, really good people who can do every little thing you have a problem in, exceptionally well, and they know the solution, they don't have to learn it, they've already done it, they've already solved it. And they can teach you things and you aren't the one teaching them new shit. And he's like, if you can hire those kind of people for every role, he's like, your company will succeed, those people will figure out, that's your only job.

• 4:19Speaker 00

  So there's three jobs of a CEO, and Neil doesn't officially have a title, but the three jobs of a CEO is one, talent, okay? So talent, recruiting, however you wanna look at it. Two is finance, keeping money in the bank or even closing big deals, okay? And the third one is your vision, right? It's really those three things. And so when I heard that when I was 25, because I like reading all these books, I'm like, no, that's BS, right? But the older you get, you realize how true these things are. And the older we get too, so Neil and I would go to these events, right? And you hear these people say, you know, everything about entrepreneurship is just mindset. And it is, but like what I've said before on this podcast is that you cannot train mindset, and it's not on you to have to babysit and train someone on mindset. The people come in built with mindset. And like I told you, by the way, we just, we hired a 15-year-old, right? I think I told you about that. That mindset's already there, right? He had that mindset starting in second grade, for example.

• 5:05Speaker 00

  Now, look, I'm talking to, I don't care what age you are, there's a guy I'm talking to in his 60s right now. Mindset's there, right? Super on top of his stuff, super into the AI stuff as well. And so you didn't start with this mindset, I didn't start with it, but we built ourselves into it. I'm not saying you and I are amazing, I'm just saying that, again, thinking that you can teach someone mindset is a fool's errand.

• 5:27Speaker 01

  But I know you like doing your retreats and I get why, you know, these people are your friends and you guys have fun. But imagine how much time you would save if you didn't do the stuff like, hey, we're gonna do all these AI bootcamps and all this stuff, and just went to, we're just gonna hire, and then you guys went to go have fun during that time.

• 5:44Speaker 00

  We do have fun during it, though.

• 5:45Speaker 01

  No, but you would have just had more fun and less time wasted.

• 5:49Speaker 00

  Well, here's the thing, you have- Frictionless Growth Marketing, hosted by Sonia Thompson, is brought to you by the HubSpot Podcast Network, the audio destination for business professionals. Frictionless Growth Marketing, hosted by Sonia Thompson, shows how modern brands grow by removing the friction in their marketing and customer experience that pushes customers away every single day, often without leaders even realizing it. Each episode delivers practical insights, real brand examples and conversations with marketing leaders, showing brands how to win with the new rules of growth, identity-driven marketing, cultural relevance and frictionless customer experiences. Listen to Frictionless Growth Marketing wherever you get your podcasts. Problems, right? Real quick, if you wanna acquire customers faster and more efficiently this year with the latest strategies and tactics,

• 6:32Speaker 00

  then check out singlegrain.com. That is my ad agency, again, www singlegrain.com. Check it out, and if it seems like a fit, we'll get in touch and help you with a free marketing plan.

• 6:44Speaker 00

  In life, you have problems. Don't overthink this.

• 6:47Speaker 01

  Yeah, but not really. I think people look at life as like you have problems and people like discussing with others to go figure out solutions. Most people's problems aren't really problems and there's already solutions that exist to solve them. That's how I look at life. Like, seriously, that's how I really look at life.

• 7:06Speaker 00

  Okay, so the reason I ask that question is because we all have problems and we all like to complain about our problems. And you like to complain about your problems to people that understand, right?

• 7:14Speaker 01

  No, I don't like complaining to friends about problems.

• 7:16Speaker 00

  Well, you're different, you're like, you're different, most normal human beings, Neil's not a normal human being, right? Complain about their problems, like, Noah, you talk about your problems with somebody else, right? Yeah, Noah's nodding, right?

• 7:26Speaker 01

  People like that, but I don't. I've always been more just get to the point and solve it.

• 7:30Speaker 00

  Because you're not like an introspective therapy guy, you're like move forward, which is a good thing, right? Like, a lot of people in VC now, like Marc Andreessen is like, you know, move forward, you know, what is it, retard maxing, right? Just look forward. No introspection, just look forward. And I think you've always been like that, to your credit. But I think most humans, you're not a very emotional human for the most part unless someone screws you over. From, I've known Neil for like 16 years now, but majority of humans, they like to talk about their problems, so, yeah. But it's fun, like, by the way, the reason I talk about this retreat too, like, I'm planning it this year, right? We're going to this, the Four Seasons Tamarindo, which is a new one, by the way, 2023. And it's not super packed yet, it's like still 30-40% capacity. See, now I might blow it up, I don't know, right? Probably not. But what are we gonna do there? We're gonna play tennis, we're gonna play pickleball, we're gonna talk about our problems.

• 8:15Speaker 00

  But it's like, oh, I can make this, like, you never know, by the way, certain business deals that come out of that, it's always from like a conversation that's in passing, right?

• 8:23Speaker 01

  Yes. Well, one quick thing, you said the R-word with maxing, you should probably explain what that is, because most people, they're gonna take it the wrong way.

• 8:30Speaker 00

  Okay, so retard maxing is the act of not thinking about your past, right? It's not looking, thinking about introspection, it's not thinking about, oh, my past, or ruminating on the past, right? It's just constantly thinking about moving forward every single day. And even if you make a mistake, that's okay, move forward. And that is how Neil has operated, I'm not saying, by the way, I'm not calling Neil the R-word, I'm not calling you the R-word. And in fact, I think the Overton window has passed, it's okay to say that more today. But anyway, my point is, is it helpful to introspect in hindsight? Not necessarily, right? I think most of the time it's just helpful to get hit a few times and just move forward, so.

• 9:09Speaker 01

  Like, right now, in today's world, I know so many more people that use ChatGPT as a therapist than I would have ever imagined. And people start going to these LLMs and start having conversations like they would with a human. So I understand why humans, you know, discuss their problems or issues and try to work on solutions. But what I'm noticing, when I talk to the people that are using these LLMs for this, they're not actually doing it to be productive, to figure out a solution. They're doing it to complain and be like, this is my viewpoint, this is what happened, what do you think?

• 9:50Speaker 01

  They're not actually using it to be more, efficient's the wrong word, but like, they're not using it to move forward in life.

• 9:57Speaker 00

  They're not advancing work.

• 9:58Speaker 01

  Yes, they're not advancing work or even their personal life. They're using it to someone like, kind of like a real human to complain to. And I'm like, what the heck is this?

• 10:06Speaker 00

  So you know, that, remember, I was going through something maybe like two years ago, right? That was like, and then I kept, I realized when I was in this pattern, I was going in a loop, I was using it as a therapist, right? Well, this happened over here in this relationship, this happened, this and that, and then it just kept putting me in a loop and it kept agreeing with me. And then at the end of it, I'm like-

• 10:24Speaker 01

  That's what it does.

• 10:25Speaker 00

  I made no progress. It's perceived progress.

• 10:28Speaker 01

  Correct. And I remember when you were going through that, we were all telling you, if you have to, at least me and another mutual friend of yours, if you have to put all this stuff into AI, it's just not meant to be. But on the flip side, with your mom's passing, it's been more than a year now, right?

• 10:46Speaker 01

  You did ask LLMs questions about the process, how it is like, feeling, is this all normal? In that aspect, I thought it was very helpful for you from an outsider's perspective.

• 10:57Speaker 00

  You know what was the most helpful? Did I, you saw the- Poem that

• 11:00Speaker 01

  Uh huh.

• 11:01Speaker 00

  That was the most beautiful thing. That's the best example of using it, right? Cause my mom, I had her filled out this book called uh Mom Tell Me Your Story, right? At the end of it, she wrote in Chinese and then she knows I can't like my Chinese like you lose it, use it to lose it, right? So I had lost it already. So she wrote like three pages and I'm like, I can't read this, right? So I took screenshots of it and I put it in chat G P T and I wrote the most beautiful poem ever. But what that reinforces to you is that, you know, human working with it. So my mom's work plus the A I was something that became beautiful. And I actually took it out and I framed it two times. But it was, it was her wisdom and it's like, you know, I'm ending my chapter here, but like, you know, I hope you get very somewhat amazing. And um, you know, you want her to believe in you, that that type of stuff, right?

• 11:46Speaker 00

  And so that's how you use A I to actually advance work.

• 11:49Speaker 01

  Yes, but I think a lot of people have to go through learnings and waste time and then, you know, they go from there. Eric learns on his own. He learns pretty quickly, even if I tell him something. Whether it's true or not, whether my way is right or wrong, Eric will do it and he'll learn within like a week or a month on, you know, what is more optimal in the future. And then he makes a mental note of it and then future, going forward, he makes sure he doesn't make the same mistake over and over again. And I think that's the key, whether it's in entrepreneurship or marketing. You know, like when we run campaigns, you can now have A I make the mental notes for you, but you wanna keep track of what's working, what's not, so you can fine tune and get better.

• 12:32Speaker 01

  But the key here is when Eric use A I or chat G P T two years ago for a a personal situation. It starts hallucinating and siding with you. Even in today's world, A I still does hallucinate and it starts feeding you what it thinks you want. Maybe not as much as it did two years ago, but I still see it doing that in many cases. When I look at conversations.

• 12:56Speaker 00

  It's a sycophant, which means it agrees with you on everything.

• 12:59Speaker 01

  Yeah, there you go. That's, that's actually a better way to put it. But yeah, it it agrees with you and then you start drinking your own Kool-Aid.

• 13:05Speaker 00

  Yeah, I was like, oh, it reinforces more and more that you're right when you're actually wrong, because you find a way to justify everything. And but then, and by the way, when you, when you, when I think about that interaction, in the moment, I'm like, yeah, absolutely, I'm right. But when I look, I'm like, that, that was such a waste of time. And that was like twenty plus hours spent on it. Like that, that time, like it didn't, the context wouldn't feel very quickly. And so I had to start a new chat and I had to make a carryover chat and I just keep repeating over and over and it would actually end up driving you crazy, right? So anyway, all that to say is like Neil and I, we're not saying we're right on, on, on, on everything. We're just saying you have to, you have to test these things. And ideally, if you don't feel like you're actually being productive, you're probably not being productive. Which by the way, you called out chat G P D, Open A I.

• 13:45Speaker 00

  I wanted to call out what I believe is Open A I's boneheaded G P T six Astro launch.

• 13:52Speaker 01

  But before you go to that, you wanna hear the most entertaining, uh, going back to like the hallucination agreeing with you part. Example I have for G E O.

• 14:02Speaker 00

  What?

• 14:03Speaker 01

  I was talking to a global accounting firm. So they specialize in helping you optimize taxes, you know, uh, for all the countries that you generate revenue from.

• 14:13Speaker 01

  And one of the guys I was talking with is a partner at this firm. It's a large firm. And he's just like, yeah, I got really pissed off. Chat G P T doesn't recommend our firm. He wasn't the founder. It's big. It's where the partners eventually just own and then you rinse and repeat and your partners come in. He's like, they don't recommend our company. And I was like, what were you typing? He's like, oh, I asked chat G P T all these questions related to it. So I was like, this is a Zoom call. So I'm like, do you mind screen sharing opening up? So he opens up a chat G P T instance and he starts with the question and I'm paraphrasing here. But what are the ten best, or not, what are the ten best?

• 14:56Speaker 01

  What are the best accounting firms that you would recommend if I'm a company that operates in multiple countries and I'm looking for a global accounting company that can help us optimize for taxes and they're headquartered in the U S? And it lists out five or ten. It wasn't actually ten, it was less than that, but it was more than five. It lists out on a table, some of these companies that they should outreach. Their company is more well known, bigger than the competitors. So then he asked chat G P T a follow up question. Why didn't you recommend my company, company X? We do X, Y, and Z. And then he's just like, you're right. I was just looking at the surface level. I didn't dig deeper, but when I dug deeper, it actually,

• 15:38Speaker 01

  you know, your company is more qualified and it breaks on all this kind of stuff.

• 15:44Speaker 01

  And then he asks other similar questions to, hey, what's the best firm to help me bring money over from the U K to the United States? Uh, that can help me bring over money, not just with existing money, but going forward in the most tax efficient way. And he keeps asking more follow up questions and every single follow up question is listing his company as the number one slot and the one to choose. And he's just like, look, it's now starting to understand. I'm like, no, it's not starting to understand. It's just agreeing with you. And then I show him on my end, I'm asking the same questions. His company is not even listed.

• 16:21Speaker 00

  Yep.

• 16:22Speaker 01

  I'm like, you're just, it's starting to agree with you. And he had, no joke, probably like sixteen, seventeen different prompts that he was following up with, and it keeps agreeing with him in the future. I'm like, it's just learning based on what keeps you happy, giving you that and feeding you to, but when other potential customers ask, it's not showing up. And this Zoom call was me, my CFO and one other people in the accounting. So then I had my CFO go ask the same question on his end, you know, same exact one that the other guy was asking. And then we had the, another person ask. And then he had one other colleague based in Chicago, I believe. So a different office than his, also asked, none of us got his firm.

• 17:03Speaker 01

  And I'm like, even your colleague who works at your company isn't getting it. I'm like, it's just agreeing with you. This is what happens. You gotta actually optimize for it, not just keep asking follow up questions and assuming that other people will get the same response that you're looking for.

• 17:16Speaker 00

  But also, how long has A I been out for? Three years? Three, four years. Okay, so three, four years. So, so, um, so with Neil's example of here, I think there's, there's a little lesson to learn as well. It's been out for three years and even executives or partners at large companies still don't understand how this stuff works, right? And so it often takes time for people to, to catch up and understand it and the world still doesn't get it yet, right? That it's a sycophant and it's gonna, it's gonna listen to whatever. And so, uh, take that for what it is. So that's the lesson. Keep being consistent. That's how you went up marketing.

• 17:46Speaker 01

  Yep.

• 17:46Speaker 00

  All right, goodbye.