In this episode, the widening gap between AI optimism and skepticism takes center stage, with employee trust in AI declining year over year even as 2026 models leap past everything before them. The real culprit behind flat productivity metrics gets named directly: work slop, unreviewed AI output that dumps the verification burden on the receiving party. Real-world proof points anchor the argument, from a $40 million business built in two years to Salesforce spending $300 million annually on Claude API tokens.
Key Takeaways:
- Work Slop Is Killing Productivity Stats Lazy, unchecked AI output forces capable receivers to redo the work, which explains the flat or negative productivity metrics companies keep reporting.
- The Learning Curve Got Steeper, Not Easier Competing LLMs, connectors, plugins, skills, and multiple product surfaces make today’s AI harder to onboard than the single chat box of late 2022.
- Earn the Right to Skip Review Only stop line-by-line checking after thorough QA proves the workflow, like SEO and GEO posts that now hit the mark 95 plus percent of the time.
- Soft ROI Beats Hard ROI Long Term Hours saved are measurable, but compounding capacity from continuous AI upskilling matters more, and most companies ignore it.
- Follow the Money, Not the Surveys Salesforce’s $300 million per year on Claude API tokens and a $40 million business built in two years with AI prove productive returns that outdated studies on deprecated models miss.
Listen to the full episode to learn how to be an AI architect instead of a work slop contributor.
(Podcast transcript below)
Welcome to Get Serious. We’re at Serious Land Capital. We have funded over six and a half million dollars worth of vacant land deals with industry leading 41% operating margin. So today, topic is within the AI space. Again, no surprise here. We are as deep into that area as you you you can imagine. You can find tons of content. we, you know, our reputation within that field continues to get recognized.
like I’m you know a senior advisor for Callan’s uncommon business team, which is tracking to do forty million dollars this year, is well here too. So like we are certainly pushing forward in in the AI skill department and you know folks continue to to recognize us. more consulting opportunities coming up here.
shortly here. So, you know, I just mentioned those bona fides up front. so you’re more likely to potentially consider my my opinion within the AI space potentially a bit more here or not you know totally up to you you can shut this off at any time or feel free to to disagree here it’s totally your choice so you know
It it still is pretty amazing to me on just the bifurcation of AI optimism versus pessimism within the you know US society, may maybe globally as well too. But you I’ve seen a lot of the stats from the the US side of things where, you know, the biggest A AI firms have not really done themselves any favors. And you know, what when you come out and say, hey, like
Decent chance a lot of white collar jobs or even more are going to be replaced over a certain amount of amount of years. Like it’s it’s not really a route to you know encourage folks to get get behind them. and you know, kind of kinda root for that you know, their their own dissolution as a employees and you know, sense of contribution and purpose and you know meaning within life here. So you know.
Totally get it from that side. but I I I think the the thing that perhaps is even a bit more surprising is just like, you know, the skepticism for AI, like even as these models are becoming more powerful, is that the trust has been decreasing even more year over year, at least over over the past, you know, couple of years here, you know, twenty twenty five, twenty twenty six, where, you know, twenty twenty six especially, like the models are just
leaps and bounds better than anything that that came in the past. And there’s just still like a lot of pessimism and I would almost say cynicism on how effective AI is from bumping productivity as well too. And especially, you know, I’ve we you might have seen some of these other charts like this you know between the C suite a lot of companies as well as the the actual
you know, individual contributors or, you know, the employees. the employees tend to be far more skeptical or far you know, more more likely to say, Hey, AI is like not helping my job or, you know, potentially even making it worse, whereas like kind of the opposite for the the C suite on, you know, perception within the company. And I I don’t know the inner work like specifically which companies were being pulled here ’cause, you know.
Just speaking from my own personal experience, I think tons of companies, even if they have AI, like a lot of them are improperly using it. So I wouldn’t be surprised if you know a lot of employees are having a lot of issues because maybe they’re not using necessarily the best models. They don’t have the right teachers. it it they might have to go through a bunch of regulations in order to utilize it with within their company or some things are blocked.
you know, certain certain services or, you know, areas of the company are blocked versus others. Whereas like, yeah, at our company we’re just fleet flowing or f free flowing, you know, smaller shop here. So like everything we can just push AI, boom, boom, boom, boom, boom. and we don’t really have any limitations on on where we can we we we can utilize it. Plus like we’ve learned from the best and we we continue to just hey AI first mentality a hundred percent of the time and we use the the best models on the planet, at least the ones that are publicly available.
So that’s a big distinction there too. And maybe some of these industries, like who knows, could be in areas that I yeah, I I still have never seen an industry yet that like couldn’t be, you know, you know, high high highly benefited by, you know, targeted use of AI. Like for instance, I’m having a a scope of work AI consulting meeting next week that’s in the
mulching space like landscaping and you know you you you you would think okay yeah it’s about like as physical as it gets but nevertheless like you know when we’re when the company is trying to target additional roll ups or you know handling service orders and so forth and who who’s going out to whatever space like yeah there’s a lot of room for AI. So like that’s just another example where
There’s still plenty of opportunity to, yeah, reduce overhead and just make, you know, smoother, smoother workflows more applicable. so yeah, you you do see a lot of that. And like I’m as much of an AI optimist as there is. I’m sure there’s probably more. Like, I I don’t don’t get me wrong. I I I think there’s still plenty of negative or questionable aspects of AI. Like, I mean, yeah, take a look at all like the AI avatars and the
loneliness epidemic and you know people not communicating with with each other. I think those are really, really tricky issues and dangerous issues for for society and need to be handled appropriately or like you know the gr the wealth disparity you know if a handful of companies like you know take most of the you know potential economic throughput within the the overall
economy, like that’s I don’t think the best thing really in the short, medium, or long term for society. So yeah, plenty of warning signs, nevertheless. like I think the pros outweigh the the cons by a long shot at the moment here. And a lot of that is just like through my own usage, our own company’s usage, just our productivity like meaningfully in a hard way as well as a soft way. You know, hard is like, okay, can you actually measure
you know, total hours cut in order to do this task and how it relates to ROI and soft as in like we know that our continued investment in upskilling ourselves in AI like is building our capacity for even greater growth in the future, even if that’s harder to put into numbers right off the the bat. So th those are a couple of ways to to measure it. I would argue that the softer ROI is
more important, especially in in the long run here. But but it’s an area that that a lot of companies do ignore. yeah, because you know you you got a lot of especially the publics and so forth, like you got a lot of bean counters and just very short term metrics to to target here. So totally get that. But you with with you know
All of this in in mind here too, and like looking at some of those stats where people say, Yeah, is AI really helping be that productive? And like when were some of these studies done? You know, some of these were on like much older models that are just like completely deprecated at at this point, not really relevant. in in today’s day and age, like post-cowork and, you know, now like Opus 4.8 and you know, now like Fable just came out for for anthropic here, like the models are just keep churning out on a week by week basis, like becoming
more and more capable and I’ve been testing these and they’re like, Yeah, it’s just getting ridiculous. So a lot of this data on, yeah, is it really affecting productivity is just becomes out outdated week over week. So it’s really hard to figure out, okay, what’s truly accurate here. But again, I just want to point out like I can definitely speak to our company. Our productivity is way beyond, way beyond what we were ever ever able to do in in the past here. you know, my my own self personally, like
you know, it’s always more disciplined worker and so forth. Like it it it’s just ridiculous what what what you can be capable of. So like I I j I just don’t get the counter argument from from the other side here, just like based on people who are really applying AI and like not just our company. Like it definitely improves productivity. So to me it’s just more of like the education piece and like what blockers are in place or regulatory hurdles and so forth that are preventing some of these other
companies from from bumping forward. And I think like all of this can really boil down to the education side of this and how are you using AI and like how how are you actually architecting things? And you know, even when I’m in advising you know new customers within Callan’s accelerate to automate course which like has a high threshold to enter he has a nine thousand dollar product for you know
12, 13 weeks worth of of training here. So people are very motivated to take advantage of it. though I have been surprised, just like to it it still takes so much to get folks over the hurdle of just like using AI as a thought partner to solve problems. like that, that that’s really the core skill set for being an AI architect. and I I I I just don’t think that many people have like fully
grasp that or you know how to orient the correct questions and like keep pushing and knowing how to utilize your judgment for what to accept versus not. And you know, part of this is there’s a certain learning curve to AI. and interestingly you know and I think even counterintuitively is that even as these models
are becoming more powerful and technically can you know produce output that requires a lot less initial effort in order to get out compared to older models here. I would argue but that the learning curve is actually steeper than it had been with the early models, just because there’s so many other connect connectors and like different
routes to utilize AI now that you know for somebody who’s not well versed in you know how how to approach all the different facets of the LLM models and which LLM in the first place, ChatGPT or Gemini or Claude and so forth, like it it becomes very overwhelming to start to figure out okay, what what should I even do here? Like in the past, you know, think ChatGPT three, you know, the or
2022, end of 2022, and when this first came out, it was like a chat function. Like that was pretty much it. again, the models were much more limited there. but for a learning curve, it’s like, okay, once you kind of figure out that piece, yeah, sure, you could start learning prompt engineering and there’s different routes to you know, get better better outputs. but the the the funnel was far more narrow.
whereas now again, you have all these different LLMs competing and you have like different facets of them. You have like the coding functions you with you know within Claude, you got the Claude Chat versus cowork versus Claude Code, and then you have projects within all of those spaces. you you have connectors and plugins and now you know cloud skills and so forth, or you know, there’s also like skills within ChatGPT’s format and so forth. So and you know, by the way, you can
Connect to you know, a whole bunch of other different softwares out there, like through these MCP connections, like connect to your Notion database or your Gmail or your Google Drive and and what have you. and all of a sudden, like, and the potential is way beyond what it was, you know, a handful of of years ago. and you know, your ability to create is
Again, leaps and bounds what what it was e even even a year ago. but from a learning curve perspective, I would say, yeah, that the this is actually tougher. And like out of all those functions that I’ve mentioned, even though I like I’m in AI like all day, well I shouldn’t say that, newborn at home, but like every day, certainly, which only a fraction of the population like uses AI every day. Like it’s it’s it’s min minority.
of of working adults that that do so e even like the younger generation, the Gen Zs and all that, more tech savvy, you would always think. like that that that is the case. Like not everybody’s using using this technology all the time. we are certainly up there and there’s still plenty of stuff we’re just learning every single day, figuring out, this works this way and stuff I haven’t really even explored, like Claude Code, I’ve like barely even touched. Just haven’t haven’t needed to. Like I know it’s there. and I can jump into it if if I need to, but like
So far, everything I’ve wanted to do, I can accomplish with other tools, you know, specifically Claude Cowork, sometimes within the chat, Claude Chat, that, you know, still allow us to be again like world class. I I I I think is safe to say, given the recognition we we’ve achieved within this space. with you know still having a limited
deeper understanding of of certain parts of of AI. But just again, like the key thing to know is like to be an architect is just understanding, hey, I can like the the the this is a generalist world now. And I I would still call myself generalist, even as specialized within real estate and so forth too. Like I just have so many other interests and you know worked in so many other spaces and like AI just gives you the ability to you know grasp an understanding and and
Deep dive into particular projects and get it get up to speed far faster than than in the past, which is why I like am able to advise so many other businesses on how to apply AI within their their specific situation professionally or personally here. So that that’s just a another piece that you know just just needs to be conveyed here and probably like understood a bit more. And like again.
Part of this is just like the evolving nature of the technology. yeah, like utilizing the internet and so forth too. It used to be super big, you know, you gotta dial up, get online, and then like things certainly expanded. You know, you eventually got all these different social media platforms, or what what how how does SEO work? or now you get like major e-commerce marketplaces, and you know, maybe your specific Amazon only seller.
for instance, or maybe you’re doing drop shipping by connecting through all these like world supply chains. or you know, maybe you’re running on, you know, crypto rails, for instance, too. So like everything expanded. And so you would say, yeah, like the overall learning curve was just tougher, even w a as the output increased along the way here too. So the way that I see it is, you know, people have been saying, Yeah, it’s a like
Does it really even make sense to like coach folks on on AI or like you can start to get it really, you know, kind of quickly? And then it’s like using computer. Yes, to a certain degree here, but again, like we’re still in such early stages that the learning curve is just tougher. And again, you just have to account for so many, so many so many human beings like don’t want to take the effort to like really, really immerse themselves and they’re plenty busy and got a million things going on to it. I fully admit too. Yeah, like a three and a half year old and a
And two-week old at at home here too. Like a mind is all over the place. Like I gotta have super targeted actions. Don’t have all all day to just learn. but you know, I’m very targeted with the learning that I do. So that that’s just something worth bearing in mind as well. And again, just how you how how you would actually be an architect going forward. And I I think the the point I wanted to
You start to conclude here. And this is again kind of tangenting back to why people don’t feel as productive or companies are not seeing, you know, the time savings or increased income and so forth too. Is like there’s there’s so much. I I saw one of my other mentors comment on this call, work slop. so you’ve you’ve heard all like the slop for you know, AI slop, right? You know, a lot of it, you know, like meme related or entertainment related. And it’s just like, okay, it’s kind of just garbage.
i i imagery or videos and so forth that yeah yeah it’s just slop, right? or maybe even articles or something like that, i inundating every corner of the internet. but now you have work slop, which is like, okay, people, their intent is to utilize AI to actually do something productive, but they don’t have the guardrails in place or they don’t know how to check their work, or they’re just too lazy to you know
properly discern what is actually solving somebody’s problems. And, you know, what what they’ll do is, you know, they’ll generate a report, send it over to somebody, but it’s a complete slot. Like it’s it’s useless. And so then it puts the onus on the receiving party to you know, especially if they’re capable and and they’re, you know, ass is on the line for okay, making sure, hey, this deliverable is appropriate here or the decision to to make.
Like they have to go through the work slop and they’re either gonna have to go back to AI and figure out, hey, what are all the issues here? Or like, you know, manually have to run down, okay, all what what what’s even correct here versus not. And that that’s what I think is driving a lot of those, you know, negative or flat productivity metrics that we’re seeing reported across these industries. is people just being lazy with their AI output or not, you know, again.
utilizing their own judgment, being an AI driver, as I’ve talked about before, they’re AI passengers, like, yeah, throw it in there, get get an output, boom, just spit it out and move on to the next thing. Like that is not how to to utilize AI, especially for like more important professional things that that you don’t have nailed nailed in from you know very solid project or prompt and so forth. Like, yeah, if if you’ve
Done your quality analysis here, your QA, and tested things thoroughly on you know, like for instance, like our SEO and GEO posts, like we’ve really, really nailed that down. So now we don’t need to read line for line because we’re confident that you know, we’re gonna nail it, you know, 95 plus percent of the time with with the output here. But if I’m doing something new, like you gotta look over that. and and make sure you actually understand what you’re asking the AI to to send over. Like I
I’m I’m seeing it more and more like even within the the land space and people sending over, here’s a parcel that’s yeah, gonna be worth like ten million dollars because it’s just in this you know in industrial development pathway. and you know and again, like I I can see all the formatting that’s like, okay, this is just straight from ChatGPT. but it’s so generalized and you know, so overly optimistic.
Obviously, because it doesn’t actually have the understanding of what what do I need to do to discern the underlying market mechanics? and what’s actually feasible from a dispo side, especially in in this market here. and again, like especially for like, you know, major multi-million dollar potential in industrial play. And you know, the whole project is boiled down to like, you know, four.
bullet pointed sections. like that it’s it’s just pure work slot. Like you you if you send stuff to that, send stuff like that to me, like it’s not going to get a response. Because it’s not my job to clean up your work slot anyway, just like, you know, vice versa here. So you know, again, there’s like plenty of time. And I we we use tons of a you know AI generated output that we will review and make sure and before we you know send it out. So it’s like
I I’m not saying hey everything needs to be done manual or like you know any any AI output or you know stuff that’s formatted through an AI side like isn’t appropriate. No, we we definitely utilize that, but we are checking and making sure, hey, does this actually make sense? Is this accurate? usually using our human judgment to to ensure that we’re not sending work slop. but yeah, every time this stuff is coming coming over, it’s like I I kind of chuckle a a bit too. Like I wish it didn’t come to us, but it’s like
Man, just people like it it’s so bad. and it, you know, it’s completely wasting our time. Like, yeah, okay, get out of here. We’re we’re we’re not responding to this. and if you think you’re gonna pull one over on us, and and get a five million dollar check out of it, like no no way in hell. so again, that that’s an example here. I’m sure you have all encountered some of this work slop here, but like, don’t contribute to the mess.
and if you don’t contribute to the mess, like you will stand out even more. and your company will thrive even more. As again, some of these other companies clearly are struggling with productivity metrics and getting their employees to buy in and so forth too. So if you can rise above and like again, take our example, take you know, so many of these other examples from like, you know, Callan’s business to, you know, so many others that have learned underneath them and not just Callens, but like the butt you know,
Mention them because you know we we learned from them as well too. And like Gallen built a $40 million business in two years. a lot of that with with the help of AI, like you know, she she walks the walk, right? So you know, I it you just have have to mention that like there’s there there’s real examples now, right? or I mentioned, you know, last week of you know, Salesforce, you know, spending $300 million worth on a year for for cloud API tokens. So like
big companies, and and the ones who are like really, you know, continue to change the game. Yes, they see the potential that this is technology. They would not put those investments and and dollars to work if they clearly were not getting productive returns from them. So like just use your brain, you you use your own experience to discern where a effort should be put into. Don’t contribute to work slop.
Be an AI architect, just know how to ask the right questions back to the AI models like that. Those are the key takeaways here. And just understand, like, yeah, that this can help you know your business get to heights that you’ve never thought possible with less time needed from your end than than you ever thought. So like that’s why we keep pushing every single day and have such enthusiasm for it. So
Hopefully you can soak that up from me. I don’t know how to make that any more clear. if you disagree, again, feel free to to let me know or I like I’m again open sponge for this type of thing, because like clearly this is a major schism within society right now too, which I I find so fascinating, especially being on like one particular side fairly firmly.
so with all that in mind, subscribe and share, everybody. If you’ve got a land deal, send it over, seriousland.capital, you know where to find us. looking forward till till next time and take care. Bye.
