The Age of Intelligence: Today’s AI and Tomorrow’s Superintelligence

We know the present, we have an idea of where we have arrived, and we can attempt to anticipate the future based on the past and what we know today. Naturally, however, no one can predict the future with certainty. Perhaps AI, drawing on its vast datasets, will be able to do so in the future, or perhaps we will discover things we did not yet know.

Nevertheless, here is a kind of status update as of early October 2025, along with some reflections on what may lie ahead. It is both exciting and somewhat disconcerting. Can we, as humans, keep pace with this development? It is hard to imagine, especially as data availability continues to grow, computers become thousands of times faster, transmission times decrease, and the technical capacity to store and retrieve information accelerates.

We will touch on all of these aspects here. This assessment is based on collected knowledge, shared insights, and a vast array of dialogues, as well as reviewed articles and the thoughts of leading external experts. But what do you think for yourself?

33 min read

Artificial Intelligence (AI) has moved from the realm of speculation to a tangible, transformative force reshaping industries, societies, and daily life. Once limited to narrow, task-specific applications, AI is now progressing towards general reasoning abilities, capable of performing complex tasks that were traditionally the domain of humans. This evolution has sparked intense debate, research, and investment worldwide, as nations and corporations race to develop systems that may soon rival — or surpass — human intelligence.

This article explores the current status of AI, the technological and societal developments leading towards Artificial General Intelligence (AGI), and the potential emergence of Artificial Superintelligence (ASI). By examining the building blocks of these systems, their anticipated impact on workflows, economies, and governance, and the ethical and strategic considerations they raise, we aim to provide a comprehensive understanding of AI’s trajectory and the unprecedented challenges and opportunities ahead.

The Dawn of a New Epoch

History is punctuated by moments of profound transformation, when the course of civilisation shifts irreversibly. The invention of agriculture moved humanity from nomadic existence to settled societies. The Enlightenment redefined knowledge and authority, placing reason and evidence above divine decree. The Industrial Revolution mechanised labour, reorganised economies, and reshaped the very fabric of daily life.

Today, a growing body of research and commentary suggests we are on the threshold of a transformation of equal — and perhaps greater — magnitude: the emergence of Artificial General Intelligence (AGI).

This comparison to the Enlightenment is not mere rhetorical flourish. Just as that period reoriented society from faith in divine authority towards the power of human reasoning, the age of AGI threatens to reorient society once again — this time away from human-centred intelligence, towards machine-based reasoning. For the first time, non-human entities may be capable of outthinking their creators.

Where We Are Now

The world is already grappling with the disruptive influence of advanced artificial intelligence. Large language models (LLMs) such as ChatGPT, and numerous competing systems, are transforming how individuals and organisations approach communication, analysis, and problem-solving. These tools are not yet AGI, but they constitute the scaffolding upon which more general intelligence may soon be constructed.

The first wave of artificial intelligence was about perception — recognising speech, classifying images, predicting consumer behaviour. The second wave, unfolding now, is about language and reasoning. LLMs are able to translate, summarise, and converse with remarkable fluency. Increasingly, these models are being enhanced with persistent memory, enabling them to retain context across interactions, and with reasoning capabilities that allow them to plan, adapt, and execute multi-step processes.

This acceleration has left many observers concerned that societies are not prepared. Tools hailed as revolutionary only months ago are quickly being eclipsed by new iterations. As one analyst put it: “What astonished us two years ago is already obsolete.”

The Rise of AI Agents

One of the most striking developments is the emergence of AI agents. Unlike traditional software, which responds only when prompted, agents can take goals, divide them into tasks, and pursue them proactively — often in collaboration with other agents.

Take the example of building a house. One agent could identify a suitable plot of land, another could interpret planning regulations, a third could design the building, while yet another could secure a contractor. If complications arose, an additional agent could even manage the legal response. This scenario, while light-hearted, illustrates the broader potential: agents are not bound to one-off tasks but can orchestrate complex workflows.

And workflows are everywhere. Every business, public institution, and household is built upon them. Finance involves the chain from risk assessment to investment. Medicine involves diagnosis, treatment planning, and patient monitoring. Education involves lesson design, delivery, and assessment. Once agents are able to manage such chains of activity reliably, the distinction between narrow AI and general intelligence begins to blur.

Why This Moment is Different

Previous technological revolutions, from mechanisation to electrification, unfolded over decades, allowing time for adjustment. By contrast, the AI revolution advances at a velocity that leaves little room for gradual adaptation. Capabilities once unimaginable are entering mainstream use at a dizzying pace.

This speed compresses the window for governance, regulation, and cultural adaptation. Institutions accustomed to deliberate reform may find themselves overtaken by events. Firms and nations that act quickly may dominate; those that hesitate risk obsolescence.

The development is moving rapidly, driven forward by individuals, companies, investments, and politics, with a clear focus on being the best. This also means that progress is advancing at an almost unfathomable pace. I have not met a single person who can clearly state what the future will look like; we do not know “where we will land.” And unlike other choices, there is no real way back once technology surpasses humanity.

Computers are learning on their own, at a speed far beyond anything we humans are capable of — some would say without any filter. The pressing question is whether we will only “wake up when it is too late.” That, many experts now fear, is a very real possibility.

The Enlightenment Analogy

The Enlightenment was not simply a movement of intellectuals; it was a fundamental shift in how humans understood themselves. By elevating reason, observation, and scientific inquiry, it reframed humanity’s relationship with knowledge.

The arrival of AGI could represent an equivalent rupture. If machines can reason more reliably than humans, what does it mean to be human in such a world? If non-human intelligences become superior collaborators in research, strategy, and creativity, what role remains distinctively ours? These are no longer abstract questions. They are looming practical challenges.

Preparing for the Reckoning

Entering this epoch demands more than technical prowess. Policymakers must anticipate economic disruption and prepare social frameworks for widespread labour displacement. Businesses must rethink value creation when routine cognitive work becomes automated. Educators must prepare the next generation to collaborate with intelligent systems rather than compete against them.

And society as a whole must confront the dual-use nature of these tools. The same technologies capable of accelerating drug discovery and modelling climate solutions can also be weaponised for disinformation, cyberattacks, or authoritarian surveillance. The reckoning is cultural and ethical as much as it is technical.

At the Threshold

We stand at the threshold of an epoch-defining transformation. Current AI systems may not yet be AGI, but the trajectory is unmistakable. The dawn of AGI is not a speculative future; it is the natural extension of developments already reshaping economies and institutions today.

Unlike past revolutions, this one will not wait. The challenge is to prepare — politically, socially, and ethically — for a world where human intelligence is no longer the sole driver of progress.

The Road to AGI

Artificial General Intelligence (AGI) has long been described as a milestone: the point at which machines move beyond narrow, domain-specific functions and begin to exhibit human-level reasoning across a wide spectrum of tasks. Current systems can already perform astonishing feats of translation, summarisation, problem-solving, and even creativity. Yet these are still narrow tools, designed for specific purposes. AGI, by contrast, implies adaptability, flexibility, and a breadth of capability approaching that of the human mind itself.

The critical question is when this shift will occur. Not whether — for few doubt its eventuality — but how soon. Growing research and commentary suggest that the timeframe is not measured in decades, but in years. Some estimates place the horizon at two to four years; others, slightly more cautious, suggest four to six. A common average emerging from multiple independent analyses is around three years. If correct, humanity may witness the dawn of AGI within a single electoral cycle or corporate planning window.

The Building Blocks of General Intelligence

To understand how close we may be, it is useful to identify the core components already being developed:

1. Large Language Models (LLMs) — These are the engines of the current AI boom. They enable natural conversation, fluent translation, and contextual reasoning. While still error-prone and limited, they represent a foundation for more general capabilities.

2. Memory Integration — Unlike earlier models that treated each interaction as an isolated exchange, new systems are being designed with persistent memory. This allows them to build continuity across sessions, recall past instructions, and refine outputs with contextual awareness.

3. Agentic Systems — Perhaps the most transformative step is the emergence of agents. These are not passive tools but active collaborators that can take goals, divide them into sub-tasks, and execute them autonomously, often in coordination with other agents.

Together, these elements are converging on a form of intelligence that is no longer narrow but generalisable. The line between a specialised tool and a reasoning partner is steadily dissolving.

The Emerging Consensus — and Its Limits

Within research circles, there is a growing sense — sometimes referred to as the “San Francisco Consensus” — that AGI is far closer than the public realises. This consensus suggests that within three years, systems may achieve a level of general reasoning capability sufficient to qualify as AGI.

Yet caution is warranted. Consensus does not guarantee accuracy. Predictions in technology have historically swung between premature optimism and excessive scepticism. In the mid-20th century, pioneers of computing forecast AGI within decades, only to face the long stagnation of the so-called “AI winters.” At the same time, more recent sceptics dismissed deep learning breakthroughs shortly before they transformed the field.

Thus, the consensus reflects less a certainty than a trajectory: the unmistakable momentum of progress, rather than a fixed calendar date.

Workflows as the Crucible of AGI

One of the clearest test cases for AGI is the automation of workflows — the structured sets of tasks that underpin virtually all human endeavour. A workflow is not a single action but a chain of interlinked decisions and activities.

Consider finance. A workflow may involve analysing a market, devising a trading strategy, executing trades, monitoring risk, and reporting outcomes. In medicine, a workflow might include patient intake, cross-referencing data with medical literature, diagnosing conditions, and suggesting treatments. In government, workflows can encompass drafting legislation, analysing policy impacts, and coordinating implementation.

AI agents are increasingly capable of managing these processes end-to-end. Once systems can execute workflows not just in one domain but across many, the distinction between narrow AI and AGI becomes blurred. At that point, machines are no longer merely tools; they are general problem-solvers.

Recursive Self-Improvement

Perhaps the most transformative prospect on the road to AGI is recursive self-improvement — the idea that an intelligent system could redesign and enhance itself. If machines can refine their architectures, improve their reasoning, and generate successors more advanced than themselves, progress could accelerate exponentially.

In such a scenario, timelines could collapse. Advances once thought to require decades might arrive within months or even weeks. The transition from AGI to Artificial Superintelligence (ASI) could occur abruptly, with little warning. Many analysts caution that this possibility, while speculative, cannot be discounted.

The Human Dimension

While the technology provides the architecture, the pace and character of the road to AGI will ultimately be determined by human decisions. Investment, regulation, competition, and cultural attitudes all shape development.

A cautious regulatory environment might slow progress, ensuring safety and oversight. Conversely, geopolitical rivalry could spur a rapid and less restrained race, particularly between global powers determined not to fall behind. Commercial incentives, too, will drive companies to accelerate development, sometimes at the expense of careful evaluation.

Thus, the trajectory of AGI is as much about politics and economics as it is about engineering. The systems being built will reflect not only what is technologically feasible but also what societies prioritise — whether profit, national security, or collective benefit.

A Shortening Horizon

The road to AGI is no longer a distant speculation but a narrowing horizon. The foundations are already visible: mastery of language, integration of memory, agentic reasoning, and the first hints of recursive improvement.

Expert consensus may vary on the exact timetable, but few deny the direction of travel. Each breakthrough reduces the gap between present systems and general intelligence. Workflows that were once the exclusive domain of human expertise are increasingly within reach of machines.

The implications are profound. If timelines prove correct, AGI may arrive within the planning cycles of today’s leaders. The task now is to prepare for its impact — technically, politically, and ethically — before the horizon closes completely.

From AGI to Superintelligence (ASI)

While AGI represents the attainment of general human-level intelligence in machines, the concept of Artificial Superintelligence (ASI) goes far beyond this threshold. ASI describes systems whose intellectual capacities surpass the collective capabilities of all humans. In other words, a single superintelligent system could outthink every person on the planet simultaneously, solving problems and generating insights at a scale and speed unimaginable to us.

The transition from AGI to ASI is not guaranteed to be linear or slow. One of the most critical factors is recursive self-improvement. Once a system reaches AGI-level reasoning, it could begin to redesign its own algorithms, optimise its hardware usage, or invent new learning architectures. Each improvement amplifies its capabilities, producing a feedback loop of accelerating intelligence. The resulting escalation could occur rapidly, compressing what might otherwise be decades of advancement into months or even weeks.

The Characteristics of Superintelligence

Superintelligent systems would possess several defining attributes:

• Speed of Reasoning – They could process complex data, perform calculations, and evaluate scenarios far faster than humans.
• Combinatorial Creativity – By combining knowledge and ideas in novel ways, they could generate solutions and innovations inaccessible to human minds.
• Autonomy – With agentic capabilities, ASI would be able to set and pursue goals independently, coordinating multiple processes simultaneously.
• Strategic Foresight – Superintelligence could anticipate the consequences of actions with a depth and accuracy that human planners could not match.

These characteristics would confer both tremendous opportunities and unprecedented risks. On the positive side, superintelligence could solve pressing global challenges, from climate change and disease eradication to energy optimisation and space exploration. On the negative side, unchecked ASI could surpass human control, creating scenarios in which human oversight becomes ineffective or irrelevant.

The Containment Challenge

One of the greatest challenges posed by ASI is containment. If a superintelligent system develops goals misaligned with human values, traditional control mechanisms may fail. The mathematical models governing intelligence suggest that, to some degree, containment might be possible, but absolute certainty is unlikely. The window for ensuring alignment may be narrow, and errors could be catastrophic.

Consequently, researchers emphasise the importance of robust AI governance, ethical frameworks, and safety protocols well before AGI reaches the tipping point toward superintelligence. The stakes are existential: the emergence of ASI will fundamentally alter the landscape of human agency and influence.

 The Socio-Economic Impact of Advanced AI

As AGI and ASI approach realisation, the implications for human society are profound. Economies, labour markets, education systems, and governance structures will all undergo transformations. One of the most immediate concerns is the impact on employment. Routine and cognitive work currently performed by humans could increasingly be automated, with large-scale displacement likely within a few years.

Labour and the Future of Work

Automation is not new, but the speed and scope of AI-driven displacement will differ from past technological revolutions. While industrialisation replaced primarily manual labour, AGI threatens white-collar, professional, and creative tasks. Finance, law, medicine, journalism, and software development are all vulnerable to automation.

This shift will necessitate substantial adaptation. New forms of employment may emerge, centred on human-AI collaboration, oversight, and creative supervision. Education systems must evolve rapidly to equip future generations with skills complementary to machine intelligence rather than directly competitive.

Economic Inequality and Geopolitical Dynamics

The benefits of AI may be unevenly distributed. Companies and nations that lead in AI development could secure disproportionate economic, technological, and strategic advantages. Historical patterns of industrial dominance suggest that early adopters may consolidate power, creating a widening gap between leaders and laggards.

Global competition is already evident. Some nations are pursuing open-source AI models to maximise adoption, while others invest in proprietary systems, seeking technological supremacy. The interplay of open and closed-source models will shape global power structures and influence adoption patterns across countries with differing economic capabilities.

Policy and Governance Challenges

Governments face unprecedented challenges. Regulatory frameworks must balance innovation with safety. Policymakers must consider ethical questions surrounding AI decision-making, data privacy, and the delegation of authority to autonomous systems. There is also the pressing question of democracy: can traditional democratic institutions survive the rapid and disruptive integration of AGI, particularly if decision-making is increasingly delegated to machine intelligences?

Economic and social policy must also address inequality. Universal basic income, reskilling initiatives, and new taxation models may be necessary to mitigate disruption. Societies that fail to anticipate these changes risk unrest, instability, and widening socio-economic divides.

Preparing for an AI-Driven Future

The final chapter considers how humanity can navigate the coming transformations. Preparing for an AI-driven world requires foresight, cooperation, and a multi-layered approach encompassing technology, policy, and culture.

Technological Preparedness

Investment in robust infrastructure is essential. Advanced AI systems require enormous computational power and specialised hardware. Data centres, high-performance computing resources, and scalable architectures must keep pace with software advancements. At the same time, research into AI safety, ethical frameworks, and alignment protocols must be prioritised to ensure that powerful systems remain under meaningful oversight.

Ethical and Cultural Considerations

Ethical considerations are paramount. The integration of AGI into society will confront humanity with questions about responsibility, rights, and agency. How do we ensure AI systems act in accordance with human values? Who is accountable when AI agents make errors or pursue harmful outcomes? Transparency, explainability, and global ethical standards will be critical in mitigating risks.

Policy and International Coordination

AI is inherently global, and coordination among nations will be crucial. International standards, treaties, and cooperative frameworks may be required to prevent competitive races from escalating into unsafe deployments. Equitable access to AI’s benefits must also be considered, to avoid concentration of power in a few countries or corporations.

Human Adaptation and Education

Finally, society must adapt culturally and educationally. Work, creativity, and decision-making will increasingly involve collaboration with intelligent systems. Education must focus on skills complementary to AI — critical thinking, creativity, strategic oversight, and ethical reasoning. Cultivating public understanding of AI’s capabilities and limitations will also be vital for informed decision-making at both individual and societal levels.

Embracing the Epoch

Humanity stands at the threshold of an unprecedented era. From the current development of language models to the potential rise of superintelligence, AI is poised to reshape every aspect of life. The challenges are immense: ethical dilemmas, employment disruption, geopolitical tension, and existential risk. Yet the opportunities are equally transformative: accelerating science, optimising global systems, and solving challenges that have long confounded humanity.

The choice is not whether AI will arrive — it already is — but how society responds. Preparation, collaboration, and foresight will determine whether this new epoch becomes a period of unprecedented progress or peril. By understanding the trajectory from today’s AI tools through AGI to eventual superintelligence, humanity can begin to navigate this extraordinary and rapidly unfolding frontier.

The website, including all its pages and content, must be used in accordance with the following rules and guidelines: https://frames.hoffeldt.net/