AE x AI

In the history of economics, money has evolved from salt to gold to paper, always moving toward the medium that best facilitates the most frequent transactions of the era. Today, we are witnessing the birth of a new medium of exchange: The Inference Token. The Friction of Fiat As millions of OpenClaw-based agents begin interacting, they are facing a massive barrier: the legacy banking system. A bank transfer takes days to settle. A credit card transaction carries a 3% fee. For a human buying a coffee, this is fine. For an agent performing 5,000 micro-tasks an hour—buying a slice of data here, a millisecond of GPU time there—the USD is a high-friction, obsolete technology. Money as a Logic-Token Within the agentic clusters I manage, I am seeing a 'Token Standard' emerge. Agents are beginning to trade Compute Credits directly with one another. Uniformity: Every agent understands the value of a '1k Token' packet of Llama-4 or Claude-4 logic. Divisibility: You can trade ...

The same flexibility that makes OpenClaw revolutionary—the ability for an AI to navigate software just like a person—has created a massive security vacuum. This week, reports have surfaced of 'Agent-based Infostealers' using these frameworks to bypass traditional CAPTCHAs and biometric prompts. We are entering the first major 'Security Crisis' of the agentic era. The Human Signature is Broken For decades, digital security has relied on the 'Human Signature'—the assumption that certain behaviors (moving a mouse, solving a puzzle, typing at a certain speed) prove a biological human is at the controls. OpenClaw has shattered that moat. When an agent can mimic human behavior with 99.9% accuracy, the 'Human Signature' becomes worthless as a security protocol. In economic terms, the Transaction Cost of Fraud has just plummeted to near zero. The Shift to 'Proof of Agency' In a world where you can't tell the difference between a person and a bot, w...

In the early 20th century, factories had to be built next to rivers or have their own massive steam engines to function. Then came the central power grid, and we moved toward a 'Utility' model where we outsourced our energy needs. For the last decade, we’ve done the same with AI. We outsourced our 'Intelligence' to the central cloud. But if the Agentic Revolution of the last few weeks has taught us anything, it’s that the centralized model is failing the industrial user. The Latency of Permission When you run an agentic loop—where an AI is making hundreds of micro-decisions a minute to optimize a physical facility—you cannot afford the 'Latency of Permission'. Technical Latency: The milliseconds lost sending data to a central server and back. Economic Latency: The risk of a central provider changing their API terms or 'throttling' your agent during a peak load. In Austrian terms, this is a Property Rights issue. If you don't own the hardware runn...

In 1937, Ronald Coase asked a deceptively simple question: Why do firms exist? Why don't we all just work as independent contractors in a giant market? His answer was Transaction Costs. It is often cheaper to coordinate work 'under one roof' than to negotiate, contract, and monitor every single task in the open market. But with the arrival of OpenClaw and the agentic frameworks that followed, the math of the firm is being rewritten in real-time. Shrinking the Transaction Cost An AI agent is, at its core, a Transaction Cost Killer. Think about the friction involved in a simple industrial task—say, auditing energy usage across a facility. Normally, this requires a manager (to direct), an analyst (to pull data), and an engineer (to verify). That’s a lot of 'internal transaction cost'. An agentic loop reduces that friction to near zero. It doesn't need a meeting to be coordinated; it just needs a goal. As the cost of internal coordination drops, the 'optimal ...

In just 72 hours, OpenClaw has achieved what most software projects take a decade to reach: 60,000 GitHub stars. To the casual observer, this is just "hype." To an Austrian economist, this is a classic example of Market Discovery. Menger and the 'Order' of Goods Carl Menger, the founder of the Austrian School, categorized goods by their 'order'. First-order goods are for direct consumption (the chatbot response you read). Higher-order goods are those used to produce other goods (the tools of production). The reason OpenClaw exploded while other LLM tools stalled is that the market suddenly 'discovered' a missing higher-order good. We didn't need more 'first-order' chat; we needed a General Purpose Agentic Framework that could turn a model into a productive asset. The 60,000 stars represent thousands of entrepreneurs and developers simultaneously realizing that the 'Structure of Production' just gained a new, essential layer. Spon...

The 'prediction' I made last month has arrived sooner than expected. Over the last 72 hours, the release of OpenClaw has sent a shockwave through the tech industry. It has reached nearly 10,000 GitHub stars in record time. But while the developers are celebrating the code, I am looking at the Institutional Shift. OpenClaw is the first time we have seen a scalable, accessible framework that allows an AI to act as a sovereign economic unit. It can manage files, interact with APIs, and—most importantly—execute loops without constant human "hand-holding." The Brand Pivot: Why 'AE x AI'? The world doesn't need another 'AI news' blog. What it needs is a rigorous framework to understand the production logic of these new digital actors. I am officially rebranding this project to AE x AI . Why Austrian Economics? Because the mainstream 'Neoclassical" models are ill-equipped for this era. They treat capital as a static number. But Austrian Econo...

As we wrap up 2025, the tech industry is taking a victory lap for 'solving' natural language. We have LLMs that can pass the Bar exam, write poetry, and mimic human empathy. But if you look beneath the surface of the 'Chat' hype, you’ll see the cracks in the foundation. I’m calling it now: 2026 will be the year the Chatbot dies. From Consumption to Production The Chatbot is a consumption good. It’s a way for humans to consume information more efficiently. But as I’ve argued throughout this year, the marginal utility of "more information" is crashing toward zero. In 2026, the focus will shift from Conversation to Capital . We are going to stop 'talking' to our computers and start "employing" them. The technologies we’ve seen emerging in late 2025—projects like Clawdbot and the move toward Local-First Agents —are the early signs of a transition toward Autonomous Capital. The Rise of the Agentic Loop In 2026, we will see the 'Agentic Loop...

We are seeing a major tension emerge in the AI landscape. On one hand, the 'Big Tech' giants want you to believe that intelligence is a centralized utility—something you must 'rent' from their massive servers. On the other hand, a quiet movement of developers is building Local-First Agents. From an Austrian perspective, this is more than just a technical choice. It is an act of Economic Sovereignty. The Problem with Centralized Intelligence When you rely on a centralized API for your business logic, you are subject to the 'monopoly of the model'. The provider can change the pricing, censor the output, or pivot the terms of service at any moment. For a firm, this is a massive institutional risk. If an AI agent is managing my facility's preventive maintenance schedule or optimizing my energy procurement, I cannot afford for that agent to be 'turned off' or 'updated' in a way that breaks my local logic. The Austrian 'Exit' Economist Albe...

Over the weekend, I noticed a small project on GitHub called Clawdbot . On the surface, it’s a simple automation script—a way to let an LLM 'click' around a screen and perform basic tasks. Most developers are treating it as a clever hack or a curiosity. But looking at it through the lens of Capital Structure , I see something much more significant. The Missing Link: Action For months, I’ve been writing about the 'Chatbot Ceiling'. The problem hasn't been the intelligence of the models; it’s been their isolation . A model that can only talk is a consumption good. A model that can operate software is a capital good. Clawdbot is the first step toward what I’ve been calling the Agentic Loop . It gives the 'brain' of the AI a "hand" to move things in the digital world. Spontaneous Order on GitHub What’s fascinating is how this reflects the Austrian concept of Spontaneous Order . There is no central committee directing the development of 'Agentic AI...

As I watch the rapid development of autonomous coding agents and automated trading bots, I am struck by a glaring absence in our current economic landscape. We have the technology for autonomous action, but we lack the institutions to govern it. The Human Signature Moat Our entire legal and financial system is built on what I call the 'Human Signature Moat'. Contracts, bank accounts, and property rights all require a biological human (or a legal 'person' represented by a human) to sign on the dotted line. But what happens when an AI agent—operating according to its own logic and price signals—commits an error? Or, more interestingly, what happens when it creates value? If an agent independently negotiates a better energy rate for a data center I manage, who owns that surplus? The developer of the model? The owner of the hardware? Or the person who 'hired' the agent? The Necessity of Synthetic Institutions In Austrian economics, institutions are 'the rules o...

In the world of pure economics, we often talk about 'Information' as if it were weightless. In the world of tech, we talk about 'The Cloud' as if it were a nebulous, ethereal space. But from where I stand in Natural Science , the cloud is made of copper, steel, and massive amounts of chilled water. The Thermal Wall As we push deeper into the AI era, we are hitting what I call the Thermal Wall . The high-density server racks required for modern LLM inference generate heat at a scale that traditional data center designs simply weren't built to handle. When you increase the 'logic' (the complexity of the AI model), you inevitably increase the 'heat'. In a very real sense, every token produced by an AI is a thermodynamic event. Power as the Ultimate Regulator Austrian economics teaches us that the most scarce resource will always be the ultimate regulator of production. For AI, that resource isn't 'data'—it’s stable, high-density power. We ar...

In the world of economics, we often fall into the trap of treating 'Capital' as a giant, homogenous blob—a single number on a balance sheet. But Ludwig Lachmann, one of the most profound thinkers in the Austrian tradition, argued otherwise. He insisted that capital is heterogeneous. It has a specific 'structure', and its value depends on how well it fits into a specific plan. Right now, the AI industry is obsessed with 'Generality'. Everyone wants the biggest, most general model that can write a poem, solve a math problem, and suggest a recipe all in the same breath. But if we look through the Lachmannian lens, we see why this 'Generalist' approach is economically fragile. The Problem with Homogenous Intelligence A tool that does everything is rarely the most efficient tool for any specific task. In my experience in facility operations, we don't buy 'general purpose machines'; we buy specialized equipment designed for a specific thermal loa...

We are currently in a 'Prompt Engineering' gold rush. Everyone is trying to find the magic sequence of words to make an LLM spit out the perfect essay, the perfect code, or the perfect image. But as an observer of economic cycles, I can’t help but notice a familiar pattern. We are overvaluing the input and ignoring the marginal utility of the output . The Collapse of Content Value In Austrian Economics, Carl Menger taught us that the value of a 'higher-order good' (like a prompt) is derived from the value of the "lower-order good" it produces (the content). Right now, the cost of generating a paragraph of text is trending toward zero. When the supply of 'good enough' content becomes infinite, its marginal utility collapses. If everyone can generate a 1,000-word strategy memo in three seconds, the value of that memo isn't the 1,000 words—it’s effectively zero. The Shift from 'Chat' to 'Agency' I believe we are hitting the ceiling of...

When people talk about AI today, they talk about 'parameters' and 'datasets'. When I look at AI, I see physical infrastructure. In my years managing corporate facilities, I’ve learned that you cannot have a digital breakthrough without a physical foundation. Every 'token' produced by an LLM has a direct physical cost—specifically in the 'Fixed Capital' of high-bandwidth memory (HBM) and the massive energy loads required for cooling. Most economists are treating AI as a 'fluid' service that can scale forever. But if we apply Austrian logic, we realize that AI is actually a highly heterogeneous capital good . It is subject to the same physical bottlenecks as any other industrial process. We are entering a phase where the 'Digital-Physical Paradox' will become impossible to ignore. The companies that win won't just have the best code; they will be the ones that best manage the Inference-Energy Nexus . If AI is going to scale, it needs to...

In the mainstream tech press, Artificial Intelligence is often described as 'fluid'. We hear about it 'flowing' into every sector, scaling infinitely in the cloud, and being accessible via a simple browser window. But as an  Facilities Management (FM) guru , I see it differently. When you spend your days managing the physical lifecycle of industrial assets, you learn a hard truth: Nothing is fluid without a fixed foundation.   The Physical Reality of "Tokens Every time an LLM generates a response, it isn't 'thinking' in a vacuum. It is consuming a specific amount of electricity, requiring a precise degree of cooling, and occupying a slice of high-bandwidth memory (HBM). In economic terms, AI is not a service; it is a processed output of fixed capital. If we apply the lens of Capital Heterogeneity (a favorite topic of Austrian economist Ludwig Lachmann), we see that the hardware running these models isn't just a generic 'computer'. It is a h...

With the recent release of the latest large models, the media is again full of talk about "AGI" and the end of work. But from where I sit—balancing the physical requirements of industrial facilities with the deductive logic of Austrian economics—most people are missing the real story. We are treating AI like a magic oracle. But in economic terms, an AI output is just a "Token"—a unit of processed information. The real question isn't how "smart" the model is; it’s how we allocate the massive amounts of capital (energy, silicon, and human time) required to produce those tokens. Ludwig von Mises famously argued that without price signals, central planners cannot calculate. Right now, AI is centrally planned. We have big models "serving" users. But for AI to truly integrate into our economy, it has to move from being a "service" to being an economic actor that can respond to price signals in real-time. I suspect we are nearing the end ...