Transcription
Today, we are talking about five AI transformations that are moving so fast and have such big implications that they might keep you up at night. Hello friends, what's going on? We have a bit of a quiet day today, believe it or not, in the AI space. There has been such a huge influx of news with OpenAI's fundraising, and before that, ChatGPT, image gen, and yesterday, Runway dropping their model. But today is one of those days where it's mostly about the discussion. And on days like that, given how fast things are moving, I like to be able to take a step back and kind of zoom out and try to contextualize what's happening and what's changing in ways that might help you wrap your head around it and act upon it.
A perfect setup for this comes courtesy of Greg Eisenberg. Now, many of you might follow Greg Eisenberg either on LinkedIn or on X. He's big into the space of this new generation of holding companies. His main thing is called Late Checkout, which describes itself as the holding company building community-based businesses. Now, of course, AI has a huge part to play in everything that he's doing, and he wrote this 46 bullet-long list of what's keeping him up at night. It ranges from the implications of ChatGPT 4, image gen, to the insane speed of change in AI. As Greg points out, it's been 739 days since the Will Smith spaghetti video. Imagine what would happen to GenAI in 739 more days. There is a little bit of startup advice in there, but mostly this is about AI. Now, 46 different bullet points is a lot to take in; I guess this is why Greg is staying up at night.
And so what I've done is I've organized this into five different categories of transformation or disruption that are themes not only that Greg is seeing but are things that I'm thinking about a lot and that we're seeing quite a bit at Super Intelligent as well. We kick off with what I call the revenge of the services business. And there are a few different of Greg's bullets that fit this profile. His first: ChatGPT 4, image gen is as big as the chatGBT launch, probably will birth 1,000 or more $1 to $100 million a year vertical software businesses. Greg's sixth bullet: AI is turning service businesses that don't scale into product businesses with service margins. The new unicorns will be productized services with AI doing 80% of the work. And finally, number 36: every small business will get a ghost team—automated bookkeepers, sales agents, marketers—run by one founder and five bots.
So what's going on here? Part of why I'm calling this the revenge of the services business is that service businesses have historically been in a slightly uncomfortable place, at least when it comes to Silicon Valley. Venture capital economics are not made for all types of companies. Many, many companies, even big companies that get tens or hundreds of millions of dollars of revenue, are still not really in the venture capital bucket for a long time. Then they've been a second-class citizen among entrepreneurs. The best entrepreneurs find themselves flooding to the biggest opportunities, while meanwhile, lower-margin services businesses get somehow neglected. In general, there is a big shift in this thinking right now in Silicon Valley. Part of that is because of the blurriness of where products end and services begin.
One of the big catalysts for the shift in this thinking has been Palantir. Palantir went against the grain for a very long time with this, with something that they called forward-deployed software engineers. Basically, at a time that services were completely anathema to the vast majority of Silicon Valley founders, Palantir was embedding software engineers inside the companies that they were servicing to basically do all of the heavy lifting of data integration and deployment of Palantir software. Effectively, Palantir said that for their model, it worked better to have engineers deployed inside the company who could correctly set up and use their software than it was to just trust the company to do it for themselves. Now, Palantir has subsequently gone on to be extremely successful. One of the big themes you're now seeing in new startups is Palantir for X. And so the fact that they went against the grain and were willing to bundle services and even deploy engineers inside customers made Silicon Valley start to think a little bit differently about this. Of course, the other reason that we're thinking differently about this is that AI fundamentally changes this game. AI enables much smaller teams to do much more and much better work, which means that to the extent that prices stay similar, services can be much higher margin. And many people believe that even as prices get compressed because of AI, the service businesses that are the most efficient in the future are still going to see better margins than the ones that we've been used to in the past.
What's more, as we'll see in a minute, there are simply a ton of service business opportunities that are being enabled by AI. And when I say enabled by AI, I mean two separate things. The first is that the AI transformation and the agentic transformation itself is generating demand for a huge array of services. Those fall into a couple of different categories: specific services around enabling the AI transformation is one, but AI versions of everything else that used to be a different type of service agency is another. So, content marketing agency but done agentically; PR agency but done agentically; SDR agency but done agentically. Another type of services business category that's being opened up by AI, however, is extraordinarily niche industry vertical focused services companies. Whereas previously it might not have been economical to serve laundromat owners, for example, now there are some really interesting hybridized software and service businesses that have the economic justification to go service that market because of the reduced costs implied by AI. When you're able to service the market for less, it puts more markets on the table. And all of this gets us to our second bullet and our second big category of disruption that might keep you up at night, which we'll call the comet theory.
This comes from a tweet that I saw from Sam Parr, who's the host of My First Million and who runs an entrepreneur community called Hampton. Sam tweeted, "This Hampton member sold an 8-figure agency and now launches a new agency every 90 days using his Comet theory." The Comet theory is this: technology drives behavior change—that's the comet; businesses adapt much slower—that's the long trail behind it; that gap is the perfect agency opportunity. Companies scramble to take advantage of the tech, but they need experts ASAP. That two-to-three-year period before internal teams have the know-how are where agencies thrive. This is the best articulation of this opportunity that I've seen, and it finds life in a couple of Greg's other points as well. Number 15, he writes, "I don't know how to say it, the money and opportunity for the average Joe is in AI startups in vertical-specific applications that actually understand industry context. No, adding industry terms to your prompts isn't the same thing." 17: The AI middleman boom is just starting. Companies that sit between foundation models and specific industries will capture most of the value while both ends get commoditized.
So there are a couple of things going on here. The first is this pattern embodied in and described by the comet theory, where technology creates the momentum for change, but there's this long drag period of adoption. And in that long period of adoption, there is an opportunity for outside experts and consultants to come in. And the specific things that Greg is saying about that opportunity is that it's going to be highly verticalized, which gets back to the point that we were just making about all of these new very niche businesses that might not have been economically viable or at least economically interesting before becoming interesting in the context of AI. Now, what Greg is saying is that the people who are taking on those opportunities are most likely not just going to be random kids who happen to like that industry, but people who actually have deep experience in that particular niche or vertical. Now, by his second point, the AI middleman boom is just starting, and companies that sit between foundation models and specific industries will capture most of the value, reflects a growing sense in the world of incredibly fast and incredibly quickly commoditized software. One of the only actual moats left is the customer relationship. Customer relationship in many cases starts from a belief in the credibility of the provider. One of the best ways to have belief in your credibility is to actually be experienced in the particular area or vertical or niche that you're trying to service. So basically, if technology is changing incredibly quickly, companies are going to stick with the solutions providers that are run by people who actually understand them, actually have experience with them, and can help them navigate that change.
Now, there's another piece of this middleman conversation which I think is important as well, which is that the compression and the cost of the software itself actually creates more space and more margin to be captured by middlemen. You can have an overall price reduction and still see more, not less, margin that's available to be captured by people who sit in the middle because of how much the underlying software cost is decreasing because of the speed of change of AI, and not just the cost of software, but the cost of expertise that the software represents. TL;DR: As the price of intelligence craters to near zero, even in a world where people expect to pay less for intelligence, there's still going to be room for them to expect to pay for the expertise to implement that intelligence and contextualize that intelligence for their specific problems. We could spend entire weeks of episodes on this comet theory; I think it's fascinating and a great way to describe one of the most important opportunities we're seeing out of this AI transformation.
The third area of AI transformation that might keep you up at night is what I'm calling creative expansion, which is, of course, as you imagine, the way that AI is going to expand the capabilities of what people can do. Greg talks about this in a few different bullets. The first: the sketching economy is the real AI revolution; when anyone can turn rough sketches into production-ready designs, taste and ideation become the only scarce resources. The second: we're witnessing the birth of a whole new job category—AI workflow designers. People who can map human processes into AI-augmented workflows will be the highest-paid consultants of the next decade. Let's take these in that order. Now, with the first, Greg is trying to point to the shift that AI is creating in what skills are actually economically relevant and viable. He's talking about how taste becomes more important in a world where the means of production is commoditized, and I think that that's absolutely true. However, implicit in this is also the point that everyone can turn rough sketches into production-ready designs, and that is a really, really big point. We are moving into a world where, thanks to AI, everyone can be a designer with tools like Lovable and Bold; everyone can be a developer with tools like ChatGPT; everyone can be a strategist, a writer, a marketer; everyone can be everything thanks to AI augmentation.
Now, that does not mean that a generalist, non-specialist, non-creative, non-designer is going to outperform a designer using AI. It doesn't mean that a vibe coder is going to outperform a regular coder thanks to AI, but it means that the average ability of everyone to create anything has increased dramatically, and we have so little idea of just how transformational that's going to be. In short form, what it means is that everyone is going to create a lot more, and this theme of more is something that I'm going to keep coming back to. Now, on this other idea, we're witnessing the birth of a whole new job category—AI workflow designers. I think, on the one hand, this is totally self-evident, and what Greg is probably talking about: we used to do content and write tweets without AI; now we do it with ChatGPT. These AI workflow designers are the people who can take that sort of change and move it across every different part of the company. Except that if you've listened to me at all, you know that I don't really think that this is going to happen that way—as a one-to-one transition from old AI way of doing things to new AI way of doing things. I think that's how it'll start and how most companies are imagining it: instead of hiring a human copywriter, I'll hire an AI copywriter; instead of hiring a human customer service agent, I'll hire an AI customer service agent. But I think that this radically underestimates the implications of the cost of intelligence and the cost of production coming to basically zero. Why hire one tweet-writing agent when you could hire 300, each of whom has a different voice, a different style, and can come at things in a different way? And while you're at it, why don't you also hire some synthetic audience designers who go out and create panels of synthetic people to react to those different tweets and give feedback in ranking? And then maybe there's an integrator AI or agent, a judge whose job it is is to view all of those tweets and all of those reactions from all of those synthetic panels and put together the report with a set of recommendations that gives you the top three. And your job becomes managing that whole process and using your human taste to select the best one. Or alternatively, we live in a world where what you're actually selecting is the top 10, and those get delivered in different places to different audiences based on who they are, which is probably mediated in some way by their own personal agents.
The point of all of this is that I think that Greg is absolutely right that AI workflow designers are going to be important, but I think that their job is going to be much more significant than just mapping human process onto AI-augmented workflows. I think these are going to be managerial entrepreneurs; they are going to teach an entire new generation of people; they're going to teach us all how to become leaders of teams or even armies of agents doing work in ways that simply weren't possible to imagine before. And by the way, as an aside, this is why I think that literally every upskilling initiative and company right now is more or less completely useless to the world that we're headed into. The transformation is not going to be about can you use ChatGPT and prompt it well; it's going to be about can you conceptualize what's possible if you can hire infinite intelligences to do your job in a different way. It's much harder and much cooler. And as if to put a fine point on this, you might have heard in the ad earlier that we were talking about Lindy's big update, which includes agent swarms. The whole idea of these swarms is that you can get literally hundreds of agents to work for you at the same time. We've been using this internally for research, for content production, and for sales, and even though we are babies bumbling around in the dark with it, it's hard to describe how different it is when what the prompt produces isn't an essay but a Google folder of 50 essays that happens in just a few seconds.
Now, on this theme of mass expansion of creativity, there was an interesting conversation between Anthony Pompliano and Mark Cuban that I thought got some of these implications as well. Pomp teed off by saying, "Being worried about inflation right now is a surefire sign that someone has no clue how the economy works. Inflation has been crashing for the last 3 months, and artificial intelligence is one of the strongest deflationary forces we will see in our lifetime." Now, what Pomp is talking about is the idea that because AI makes the cost of intelligence and production crater, it means that the cost of goods that are produced with intelligence simultaneously crater, leading to deflation. Mark Cuban reposted it, though, and said, "Disagree. Generative AI will be the greatest growth in productivity engine ever. We will see amazing new things that replace the old things but will cost more, and we will bust our butts to get it. Some will be purely digital, but it will be the good digital stuff that everyone wants and will pay up to get. Others will be new products that we didn't know we wanted that we will pay up to get. Entrepreneurs will create new companies, some of which are based on ideas that AI generates that will have to compete with other startups that competitive AI dreams up, which will compete with companies who don't know how to compete in this new world. Their old things will get cheaper because they're worse, and yes, if you are watching along, you can see that I am live censoring Cuban's tweet for the sake of keeping this family-friendly." He continues, "There will be millions of models that compete and trillions of agents that compete, but intellectual property will be king as people try to determine what we value in a new world. That will lead to robots and self-driving everything; anything that needs to be physically transported, humans or otherwise, will happen in a self-operating device. Some will be drones the size of a pill; others will be enormous that will move really big things that AI invents, and that won't be cheap. Life in companies will move to be native AI rather than AI bolt-on like we're seeing today. There will be two types of companies: native AI companies that are great at it and those who are on their way out of business. Deflation happens when we get more of what we already got at a cheaper price. When everything is new, entrepreneurs will charge what the market will bear. We haven't reached the crazy stage yet because it's all changing so fast and going in an amazing direction. If you aren't excited by what's happening in AI and playing with every tool and learning all you can about it, you need to go back to your IBM PC."
Now, when it comes to the question of inflation, I don't actually think that these points are mutually exclusive, but that's not really the point here. What Cuban is describing is something that I've talked about in the past as the theory of more and is my base case for why I'm not worried about job displacement in the AI economy. Cuban's point is that basically, for any amount of money that we save on the lowered cost of today's existing goods and services, we will easily spend even more than that on all of the new things that get created thanks to the new capabilities of AI and AI-enabled people. Remember, we talked about all those people who are now designers and developers building stuff? Well, they're going to build things; some of those things are going to be awesome, and we're going to pay money to get access to them. We're going to see more content, more games, more personalization, more of absolutely everything. The question, though, is how to win in that new environment. This is our fourth theme for this episode. In a world where things are changing so fast, in a business environment where things are changing so fast, how do you actually win? What's the playbook? Greg, once again, has a couple of bullets about this. He writes, "Really smart strategy is to rebuild traditional products with AI as your unfair advantage, hiding the complexity behind familiar interfaces. Basically, just look at proven apps that have no AI; make them AI-first. If it adds a ton of value to the end consumer, use AI features; don't sell AI in creator-led marketing. This is the playbook." He also writes, "Distribution is the only moat left. Your product, tech, and team can all be replicated; your direct connection to customers cannot."
And much more than the specifics of how to win, I think what matters here is the fact that what it takes to win, by and large, is changing. Things that we used to think were absolute paragons of defensibility—unique technology, for example—are, in a world of AI, effortless to replicate. One of the big conversations happening on Twitter/X right now is the absolute cataclysm coming for the SaaS category. SaaS is based on expensive, perceived business models, and as roll-your-own code tools become more and more proficient, there's going to be more and more incentive for people to just recode the specific features that are most relevant to them and avoid those expensive and locked-in fees. Now, to the extent that there is good news, the one thing that I see everyone agree on and certainly is something that I feel is that relationships are still going to matter. In fact, they might be the only thing left that matters. Greg confirms this when he says, "Your direct connection to customers cannot be replicated." We're seeing this play out in the supposed ChatGPT rapper space, where companies that were once dismissed as simply rappers for the foundation models underneath are actually having a lot of success because they have deep knowledge of the industry or niche that they're operating within. They can design interfaces that use those models better because of that knowledge, and they have relationships with the customers that are trusted. That is incalculably valuable in a world where things are moving as fast as they are. Think about it this way: if technology is totally commoditized, if interfaces are totally commoditized, if products are totally commoditized, how are you going to make a buying decision as a business? You're going to go with a person that you trust; you're going to go with their version of the thing rather than some other version of the thing, not because it has a better feature set, but because you trust the person. This is one of the most important through-lines from the way that business has always been to the way that business will be in the future is the value of these relationships. The specific way in which that value manifests might be different, but it remains the case that relationships—this direct connection to the customer—is going to be at the very heart of success in the future as well.
Lastly, our last theme in the transformations that might keep you up at night is the idea that the moment for all of this is right now. In bullet 19, Greg wrote, "AI is creating win-or-take-most markets overnight. The window to establish yourself as the go-to solution in a specific vertical is maybe 6 to 12 months before it closes for a decade." This isn't helping my sleep. And then his last bullet: I don't know how long this window stays open, but we're in a moment where all the rules of building businesses are being rewritten. For the people playing with these new tools, creating audiences and communities, you've got an unfair advantage. Look, obviously, there will always be new opportunities. The future that is being built right now will beget new behaviors that create new opportunities that we can't even imagine yet. Even the people who are the most at the cutting edge. In fact, I think our ability to predict the future is, in some ways, lower than it's ever been because of how seismic these changes are. And yet, at the same time, something that I have felt strongly for a very long period of time, since before I started this podcast, why I was so intent on starting this podcast, why I was so intent on starting Super Intelligent, was that I do believe that we're in a call it 5-year or so period where the entire new foundation for the next generation of business is going to be built. All of the big foundational companies—the Googles, the Facebooks, the whatevers—are either going to be built anew right now, or the old companies, the old leaders, the legacy players are going to sufficiently evolve and transform themselves that they get to be the leaders for a new generation. But whatever the case, their success or their disruption will be determined in this period. I have never seen as much blue oceans opportunity as there is right now, and I think it's immensely exciting to be playing inside that.
Now, for entrepreneurs, this is unmitigated good news, especially because the rules of entrepreneurship are changing as well as capital constraints and talent constraints come down, once again, thanks to AI. For those of you inside big companies, it's a little bit more complicated. The natural inertia inside corporations fights against the speed of change. Now, I will say that without a doubt, this is the fastest that we've ever seen enterprises adapt to a new opportunity. The speed with which they have adopted and adapted to, or tried to adapt to, generative AI is fundamentally unlike cloud; it's fundamentally unlike mobile. There is clearly a sense that this is a different type of moment that requires rising to the challenge, but still, the lag remains. And I don't think that Cuban is at all being dramatic when he says that there are going to be two types of companies in the future: companies that natively understand and are good at AI and those that are on their way out of business. I think that's true, and I am here to help you be the former, not the latter. Anyways, guys, appreciate you listening or watching, as always. This is a fun one. Follow Greg Eisenberg on X or LinkedIn for more content like this; he's constantly pumping it out, honestly. And until next time, peace.