-
@ Pete Winn 🔆
2025-05-16 08:18:52Value Creation at the Edge
The conversation around artificial intelligence has largely centered on the technology itself, the capabilities of large language models, the race for more parameters, and the competition between AI companies.
He with the most data / biggest model / biggest platform wins all.
As we're been exploring in recent "Good Stuff" podcasts, the true business model of AI may be much more straightforward. AI is after all a productivity tool with little technical moat, in fact the existence of AI coding and learning tools quickly chop away at this moat even quicker!.\ \ We believe that the it's about transforming traditional human heavy businesses by dramatically reducing operational costs while maintaining or increasing output.
AI is poised to create value not primarily for AI companies themselves, but for businesses that effectively implement AI to transform their operations, particularly small, local businesses that can become extraordinarily efficient through AI adoption.
The Value Shift: From AI Companies to AI-Enabled Traditional Businesses
A central insight from episode 1 of the podcast series, is that the value of AI isn't likely to accrue primarily to companies like OpenAI or other AI technology providers. Instead, the real winners will be traditional service businesses that can leverage AI to transform their operations and cost structures.
"I think we're gonna see this shift to traditional service businesses... that traditionally have pretty fixed low margins because of a dependency on language-heavy workflows that require a lot of humans as the medium of intelligence in the business."
The opportunity here is to use AI to manage the language dependency and shift the moments of intelligence, that currently exist in the heads of our staff, into software that can run 24x7 for fractions of a cost.\ \ The real limiting factor here is less a magic AGI, but instead detailed thinking and process redesign to move humans to the edge of the process. As it turns out if we think through what each person is doing in detail we see the specific decisions, outputs, moments of intelligence are actually quite constrained and can be replicated in LLM's if we break them down to a low enough level of fidelity and take each decisions one step at a time.\ \ The result? Businesses that have traditionally operated with fixed, low margins can potentially achieve "software-style margins" by dramatically reducing their operational expenses.
Transforming Traditional Service Businesses
We have developed three key heuristics for identifying businesses that could benefit most from AI transformation:
-
Language Intensity: Businesses where much of the work involves processing language (reading, writing, communicating). Language in, language out. If you are sat in a chair and typing all day, this could be you.
-
Labor Component: Where we see this language intensity so we find many people performing similar, standardized roles. For examples, if we have four people in the same role this is a big clue we have good process, checklists, role descriptions etc for how the work can be done in order to replicate work across multiple people.
-
Load in the Business: Taking these processes into account, what amount of the operational expense of the business do they represent? Where these language and labor-intensive operations represent a significant portion of the business cost, we can see there will be significant return.
Traditional service businesses that match these criteria—legal firms, accounting practices, consulting agencies, contract engineering, design agencies and others—could see dramatic transformations through AI implementation.
By automating these language-heavy processes, businesses can potentially reduce operational costs by 50-80% while maintaining similar levels of output.
The Power of Small
We believe that small businesses may have an inherent advantage in this transformation. While large enterprises face significant barriers to reducing their workforce (political pressure, media scrutiny, organizational complexity), smaller businesses can adapt more quickly and focus on growth rather than just cost-cutting.
If I'm in a 20,000 person business and I need to remove 10,000 people... that's hard. You can't do this without sending political shock waves in your local community.
If I'm a 10 person business and I need to double my revenue, nobody gives a shit. I can just do it.
For small businesses, AI removes growth constraints. When adding the "21st person" no longer represents a significant capital investment, small businesses can scale much more efficiently:
If the next nominal client that you onboard doesn't actually cause you any more additional pain, if you don't need to hire more people to service that client... you just take off the brakes off from a growth perspective.
This gives small business a unique advantage in capitalizing on AI.
From "Bionic Humans" to "Humans at the Edge"
We currently see this integration to business happening in one of two models:
-
The Bionic Human: Equipping workers with AI tools to make them more productive.
-
Human at the Edge: Redesigning processes to be AI-native, with humans entering the process only when needed (and often facilitated by bitcoin payments).
While many businesses are focused on the first approach and it can certainly see returns, it is still a process constrained by the human input. The real value unlock comes from fundamentally redesigning business processes with AI at the core.
Now we can purchase intelligence in buckets of $0.02 API calls, how would we operate different?
This represents a profound shift in how we think about work and processes. Rather than humans being central to processes with tools supporting them, AI becomes the backbone of operations with humans providing input only at critical junctures.
This is "a complete mental shift" that challenges our fundamental assumptions about how businesses operate. The human becomes "the interface with the real world" for AI systems rather than the primary processor of information and decision-maker.
The Value Trap: Understanding the Competitive Dynamic
So what happens next? Here we have developed the concept of the Value Trap to explain how the competitive landscape will evolve as AI adoption increases..\ \
Initially, early adopters of AI and "Human at the Edge" business processes, will see dramatic benefits.
If your costs have dropped from 90 to 20 this creates an immediate competitive advantage where the early adopter is "now making 80 units of profit versus your 10 units of profit.
They gain massive pricing power in the industry and can compete for growth with an unfair advantage.
Over time, and here we believe this is likely a 5-10 year period although we believe the quicker side, competitive pressures will erode these advantages.\ \ As competitors adopt similar AI strategies, price competition will intensify, and revenues will decline. The business that initially saw its costs drop from 90 to 20 units might see its revenue decline from 100 to 30 units, resulting in similar margins but much lower overall revenue, often destroying the enterprise value of the company at these new revenue / profit levels!
This evolution creates an imperative for businesses to adopt AI early, not just to maintain perpetual advantage, but simply to survive the transition. Worse they're hit with a second challenge of the value trap, how do I keep hold of the value I generate along the way.\ \ If you're reading this on Nostr you may already suspect a way out of this value trap.\ \ If not I would invite you to consider storing the immediate short term returns you pull forwards in something that would be inflation resistant, hard to seize and ideally portable.\ \ We refer to this as a 'The big orange arbitrage".
Implications for Business Owners and Capital Allocators
For business owners, especially those running small to medium-sized enterprises, the message is clear: understand how AI could transform your industry and begin planning your transition now.\ \ This might involve creating an "AI-native twin" of your current business—similar to how Netflix developed streaming alongside its DVD business—to eventually replace your current operations. If you want help please ask, I heavily favor more small businesses in the world and would love to help make this a reality.
For capital allocation, the emerging opportunity we see if in "transformation led private equity". The acquisition of traditional service businesses and applying AI to dramatically reduce operational costs and increase enterprise value.\ \ This approach treats AI not as a product but as a transformation strategy for existing businesses with proven product-market fit.
Transformation led PE is venture style returns without the risk of product market fit.
So the lesson?
The business model of AI isn't all about selling AI technology, adding a RAG chatbot to a new DB or collecting everyone's data.\ \ Consider the humble cash flow business, use AI to transform the operational processes and save into everyone's favorite orange coin.
-