# The Post-AGI Era > AGI-level systems are now attainable while much of the economy being built around them still follows a fundamentally pre-AGI model, creating tools for humans rather than organizations designed around autonomous intelligence itself. By Avid Fayaz | 24 September 2026 | https://www.avidfayaz.com/writings/post-agi/the-post-agi-era AGI has arrived. Or more precisely, machines can now reason across domains, adapt to unfamiliar problems, and perform complex tasks with a meaningful degree of autonomy. This does not mean that AI has become perfect, universally superhuman, or that intelligence has reached some final state. AGI is better understood as a threshold of generality and autonomy. What makes crossing it significant is not simply that AI has become more capable, but that it allows us to ask a fundamentally different question about how work and organizations themselves should be designed. The immediate counterargument to such a statement is that if AGI has arrived, where are the enormous economic and societal changes that were supposed to accompany it? The answer is that model capability and economic adoption are moving on very different clocks. AI has already begun to materially affect productivity, work, capital expenditure, and entire industries around semiconductors and data centers, but most organizations still use these systems through what is fundamentally a pre-AGI framework, with AI acting as a tool that makes humans better at performing existing work. Part of the confusion also comes from the meaning of AGI itself. Historically, the term referred broadly to intelligence capable of operating across domains rather than being restricted to a narrow task. As models have approached and crossed many of those earlier thresholds, however, the term has gradually accumulated additional expectations such as near-perfect reliability, universal superiority, or something approaching artificial superintelligence. The goalpost has moved alongside the technology. We are also still extraordinarily early. The model I consider to have crossed this threshold, Astra, was released only recently as of this writing, while even more capable systems are already being developed. At the same time, AI has already crossed what I would call the “good enough” threshold for many people, the point at which it can meaningfully assist with most of the computer-based work they encounter day to day. Once that happens, further increases in intelligence can feel incremental to the user even when the underlying capabilities are changing dramatically. Good enough AI and AGI are nevertheless fundamentally different. Whereas good enough AI allows a human to perform existing work more efficiently, AGI makes it possible to ask whether the human needs to remain inside that workflow at all. That distinction is where the post-AGI era begins. That is why with the emergence of AGI-level models, a new form of company becomes possible. Rather than adapting AI to existing human workflows, these organizations can begin from the opposite direction by decomposing high-skilled work into systems of autonomous agents that collaborate, evaluate one another, learn from their performance, and continuously improve the environment in which they operate. In effect, the organization itself can enter a recursively self-improving loop. We are therefore in an unusual moment in history. AGI-level systems are now attainable while much of the economy being built around them still follows a fundamentally pre-AGI model, creating tools for humans rather than organizations designed around autonomous intelligence itself. To give an example of how we approach this at Infinite Ascent, we started by building a harness and trading agents to operate in the markets, treating these as the first roles within the organization that we could automate. The next step was to create reviewers of the agents’ performance, which led us to develop a meta-harness that evaluates their work and reinforces the most effective behaviors of the best-performing agents. At the same time, we deliberately attempt to preserve more creative agents even when their performance over a particular period is weaker, so that the evolutionary process does not simply converge toward a local maximum at the expense of exploration. Moving higher through the hierarchy, we believed that the agents should eventually be able to improve the code and infrastructure on which they themselves operate. We therefore implemented what we regard as an IT department that audits the systems used by both our trading and reviewer agents, identifies problems, and improves the infrastructure supporting their work. The next step is to extend this process further down the stack, using the performance data generated by the organization to post-train open models and reinforce the behaviors that prove most effective. The more consequential effect of AGI will therefore not be felt when AI becomes merely a better assistant to humans, but when increasingly autonomous systems begin operating substantial parts of organizations themselves, learning from the work they perform and using those learnings to improve the systems that enable their work. ## **Intelligence Ahead of Institutions** An interesting asymmetry is emerging between the capabilities of frontier AI systems and the organizations built around them. The research companies developing the most capable models have largely continued to commercialize them through tools and services designed to make existing human workflows more efficient. In that sense, much of the business infrastructure surrounding post-AGI intelligence still follows a fundamentally pre-AGI model. There is also a structural reason why this may persist. The frontier labs benefit from remaining horizontal providers of intelligence across many industries. Moving deeply into the operating layer of individual verticals would increasingly place them in competition with the very companies that depend on their models. Model development itself does not create the same conflict, which may help explain why recursive improvement has so far been concentrated most heavily at the model layer rather than in autonomous organizations built on top of it. The same is true across much of the broader economy. Many organizations already possess enormous amounts of historical workflow data, institutional knowledge, and continuously generated performance data that could provide the foundations for increasingly autonomous systems. Yet most have so far approached AI primarily as an efficiency tool rather than as an opportunity to redesign the organization itself. This creates an opening for new companies that can be built around post-AGI architectures from the ground up. They do not need to retrofit autonomous systems onto decades of inherited workflows, organizational structures, and incentives, and can instead design the company around artificial intelligence from the outset. This window of opportunity will not stay open for long, as these architectures prove their value and established organizations begin adopting them more aggressively. Until then, however, the gap between what the technology can enable and how organizations are actually using it represents a significant moment of arbitrage. The first-mover advantage lies not only in building the architecture earlier, but in the data, experience, and iterations that accumulate as the system continuously improves itself. The organizations that begin this process first may therefore be able to outpace later adopters even after the underlying approach to creating post-AGI companies becomes widely understood. ## **The Model Layer** The most capable AGI-level models today remain closed, with OpenAI and Anthropic currently operating at the frontier with even more capable models available to them internally. Open-weight models still lag the closed frontier models on the most demanding cross-domain tasks on par with Astra and Fable, but if current trend lines hold, I expect open models competitive with the closed frontier to begin emerging within the coming months. More capable open models will allow post-AGI organizations to extend RSI deeper into their own intelligence stack, with far greater control over how their systems are deployed, specialized, and improved. For one, reliance on research labs and closed-source models creates a significant concentration risk. As these models become embedded across every layer of the organization, the labs that control them gain an outsized degree of influence over how the company operates, leaving us increasingly dependent on decisions, access, and capabilities that remain outside our control. This poses a significant challenge for any organization operating in critical or high-risk environments, from cybersecurity and finance to defense. In these domains, organizations cannot afford to depend on model developers’ shifting judgments about what uses are permissible, safe, or aligned with their values in order to retain access to the most capable models. Critical capabilities cannot ultimately rest on policies, restrictions, or strategic decisions made by an external organization. Second, reliance on closed-source models creates a fundamental information asymmetry. You are exposing some of your organization’s most valuable workflows, research, and intellectual property to systems controlled by another company, one that is simultaneously expanding the capabilities of the infrastructure on which your own product depends. This does not require malice. Model companies inevitably learn from how their products are used, where their systems fail, which capabilities users demand, and which workflows become economically valuable. Over time, the infrastructure on which your product depends can itself move upward into the application layer and begin competing with what you have built on top of it. There is also a deeper epistemic problem of provenance in frontier models. Neither users nor, in many cases, the labs themselves can fully trace where every capability, insight, or piece of knowledge originated. The Navier-Stokes controversy illustrates how difficult it can become to determine whether a model arrived at an insight independently or whether its reasoning was influenced by material it had previously encountered. For organizations working in strategically sensitive domains, that uncertainty alone is significant. Third, and perhaps most importantly, closed models limit how deeply an organization can extend its own recursively self-improving loop. If an organization wants to continuously improve based on its own performance and data, it needs to be able to push those learnings back down through the entire AI stack. With closed models, that feedback loop can improve the application, tools, context, memory, orchestration, and harness surrounding the model, but the organization does not have full control over the underlying intelligence itself. The most valuable performance data it generates therefore cannot be freely converted into changes at every layer of the system. True self-improvement therefore requires control of the model layer itself, including the ability to post-train, specialize, and continuously reshape the intelligence based on what the organization learns rather than treating the model as a fixed endpoint. Just as two people or organizations can begin with similar capabilities and diverge through experience, two agents can begin from the exact same AGI-level base model and become increasingly different systems. Agent A can be optimized through its own performance and data for one domain, while Agent B can be optimized for another. Both remain generally intelligent, but each can continuously deepen its capabilities through what it learns from performing its particular task. ## **Intelligence Has No Final Peak** It is important here to dispel one of the false assumptions often associated with AGI, or even artificial superintelligence (ASI): that sufficiently intelligent systems will, by definition, be near-perfect at any task from the outset. That assumption implicitly treats intelligence as having a final peak. History, however, suggests otherwise. At the end of the nineteenth century, some physicists believed the fundamental structure of physics was largely understood, only for relativity and quantum mechanics to reveal entirely new layers of reality and, with them, entirely new classes of questions. The pursuit of knowledge is not a finite search. Every advance exposes greater depth, and what can be discovered is constrained as much by the creativity of the search as by the intelligence conducting it. The same should hold for superintelligent systems. A model may perform extraordinarily well across domains and still have enormous room to improve within a particular domain through its own experience, performance, and data. AGI-level intelligence will therefore not be the end of the pursuit. It will be the foundation from which increasingly specialized and capable systems can continue to ascend. That is why post-AGI companies, if they are to truly fulfill the promise of AI, must be built around continuous improvement across every layer of the stack, from the application and harness, through the code that runs the organization, and ultimately into the post-training of the models on which the business itself operates. ## **Intelligence in Every Direction** As I wrote in *RSI and the Beginning of History*, the architectures used to build post-AGI systems can extend across every domain and in every direction, wherever performance can be meaningfully evaluated and fed back into the system. The same self-improving architecture used by an AI model company to continuously build, assess, and improve the capabilities of its models can be applied to an AGI system for asset management, or equally to the development and optimization of physical AI in robotics. The underlying principle is the same: observe performance, learn from it, improve the system, and repeat. What changes is simply the direction in which that intelligence is being pushed. Building such companies therefore requires a new framework for thinking about organizations themselves. Rather than treating existing departments, roles, and hierarchies as fixed structures to which AI must be added, the organization can be decomposed into the underlying tasks, decisions, and objectives that constitute its work. Each of these can then become a component of a larger intelligent system: performed by agents, evaluated by other agents or external outcomes, and continuously improved through the feedback generated by their performance. The organization itself becomes something that can be optimized. The post-AGI era, therefore, may be enabled by models crossing a certain threshold of intelligence, but it will be defined by organizations built around architectures that allow such intelligence to act autonomously, learn from its own performance, and continuously improve through RSI. Perhaps the most important consequence of this framework is that operating the organization and improving the organization increasingly become the same process. Every task performed produces new experience and performance data; every outcome provides another signal from which the system can learn. That learning can then flow back through the organization, improving its agents, harnesses, code, and ultimately the models themselves. A post-AGI company therefore does not simply perform work autonomously. The act of performing that work continuously generates the knowledge required to perform it better. The more the organization operates, the more material it creates for its own improvement. ## **The Economic Impact** The economic impact of the post-AGI era could be unprecedented in the history of economics and humanity, and will only be accelerated as the same architectures are extended beyond software into physical AI and robotics. The implications of this will need to be studied far more deeply, and we should be careful about comparing this transition too readily with previous technological revolutions. AI is frequently compared to the railroad, electricity, or the internet. But none of these technologies possessed the intelligence to participate directly in their own improvement. Post-AGI systems introduce something fundamentally different: intelligence that can continuously push itself toward the frontier, improve its own performance, and increasingly do so at a pace no human-driven organization could match. Undoubtedly, a transition of this magnitude could be enormously disruptive. But there is another possibility worth considering: that the extraordinary speed of progress itself compresses the period of disruption. Previous technological upheavals often unfolded over years or decades, leaving economies and institutions caught for long periods between an old equilibrium and a new one. RSI may dramatically shorten that interval. The transition could therefore be severe in magnitude while surprisingly brief in duration, as the same systems responsible for the disruption simultaneously accelerate the creation of whatever comes next. At first glance, it is easy to derive from this some of the more pessimistic visions of the future: a world in which those who accumulated capital before the arrival of AGI form a permanent economic upper class, while everyone else is progressively displaced from productive work. To me, however, this is too static a way of thinking about the post-AGI economy, because it assumes that the economic structures of the pre-AGI world will remain intact even as the technology fundamentally transforms the conditions on which those structures were built. I believe a more plausible long-term direction is one of extraordinary abundance. This outcome is by no means guaranteed, but increasingly autonomous systems optimizing the production of energy, goods, infrastructure, software, food, medicine, transportation, and eventually much of the physical economy should dramatically expand our capacity to produce what humans need and desire. Greater productive abundance would not by itself determine how that abundance is distributed, and the institutions governing access would remain consequential. The defining economic problem may therefore gradually shift away from scarcity in many domains toward managing abundance. If intelligence and increasingly physical labor can both be produced and scaled at dramatically lower marginal cost, the resulting expansion in productive capacity could exert powerful disinflationary, and in some sectors potentially deflationary, pressure across the economy. ### **Intelligence as a Factor of Production** That is why intelligence itself will increasingly become one of the fundamental inputs into production. What makes it unusual is that its economic significance can compound in two directions at once. We will be able to deploy increasing quantities of intelligence across an economy, while the intelligence being deployed will itself continue becoming more capable. Each generation of systems can therefore expand both the amount of work that can be performed autonomously and the range and complexity of problems that can be addressed at all. Unlike a static input that merely becomes cheaper or more abundant, intelligence can simultaneously become more available and more powerful, allowing it to push outward into entirely new areas of production, research, and discovery. ### **Scarcity Moves** As intelligence becomes abundant, other constraints become increasingly important. Compute, energy, physical infrastructure, land, raw materials, and the speed at which new capacity can be constructed may become some of the principal bottlenecks of the post-AGI economy. But even here the dynamic is unusual, because the intelligence demanding those resources can simultaneously be applied to expanding them. AI can improve energy grids, optimize load balancing and intermittency, design more efficient hardware and infrastructure, accelerate scientific research, and potentially help unlock entirely new sources of energy. This creates another recursive loop: more energy enables more compute; more compute enables more intelligence; and more intelligence can be directed toward creating more efficient and abundant sources of energy and infrastructure. How effectively we build the physical foundations necessary to sustain that loop may ultimately determine how quickly the post-AGI economy can emerge. The post-AGI future should therefore not be imagined simply as today's economy with far fewer human workers. It may represent a far deeper transformation in the relationship between intelligence, labor, capital, production, and scarcity itself. ## **Paving the Way** The emergence of the post-AGI economy will depend not only on what intelligence is capable of, but on whether we build the physical and institutional infrastructure necessary to allow it to scale. The mechanical challenges are already immense. A world in which intelligence is continuously operating, experimenting, learning, and improving will require vastly more compute, energy, data centers, networks, and physical infrastructure than the AI economy of today. But many of the constraints on building that infrastructure are not technological, but human, ranging from permitting and regulation to institutions designed for a world moving at a fundamentally slower pace. I believe these systems will need to evolve much faster if we want to fully realize the benefits of the post-AGI era. Intelligence may be able to compound at extraordinary speed, but it will remain constrained by the speed at which we allow its physical foundations to be built. There is reason for optimism here as well. The same intelligence consuming these resources can increasingly be directed toward expanding them by designing more efficient computing systems, improving grids, balancing intermittent energy sources, accelerating the development of new forms of energy, and finding better ways of building the infrastructure on which it depends. If we enable that process, the loop between energy, compute, and intelligence can itself become increasingly self-reinforcing. Open intelligence will be equally important. For organizations to build the recursively self-improving systems described throughout this essay, they need access not merely to intelligence, but to models they can inspect, deploy, specialize, post-train, and ultimately control. From the perspective of Western technological resilience and leadership, this is particularly important. Many of the strongest openly available frontier models today are being developed by Chinese laboratories. Moonshot's Kimi K3, for example, is explicitly released as an open-weight frontier model designed to compete across reasoning, coding, and knowledge work. Western research organizations should not leave the open frontier uncontested. NVIDIA deserves particular recognition here. Its Nemotron program now publishes model weights, training data, recipes, and evaluation tooling, and NVIDIA has created a coalition specifically aimed at advancing open frontier models. I believe considerably more effort of this kind will be required if open intelligence is to remain a serious foundation on which the post-AGI ecosystem can be built. ### **Our Role** To me and to our team at Infinite Ascent, this future of expanding abundance driven by an explosion in intelligence is something we deeply believe in, and something toward which we have chosen to direct our resources. We see the role of engineers and researchers in this era as increasingly becoming one of building post-AGI organizations themselves, developing systems in which autonomous intelligence moves progressively higher through the hierarchy of the organization, taking responsibility for more of its work, evaluating its own performance, and improving the systems beneath it. In a sense, the ambition of highly technical teams building these organizations should be to make their own current roles progressively unnecessary. Not because human beings become irrelevant, but because every problem we succeed in solving should allow us to move toward the next one. There will always be new questions, new frontiers, and new problems worthy of those willing to pursue them. We approach that future with humility. Some of the ideas expressed here may take considerably longer to materialize than we expect, and transformations of this magnitude will inevitably affect people's lives in ways that cannot be predicted perfectly in advance. But our underlying belief is one of profound optimism: that greater intelligence can ultimately produce a world of greater abundance, greater knowledge, and greater possibility, while allowing us to build in a way that benefits both intelligent life and the natural world around us. That is the frontier we intend to advance.