Where AI Will Be in 1, 2, or 3 Years: The Future of Our Civilization
- Tinka C. Muhwezi
- 1 hour ago
- 8 min read
Elon Musk recently shared remarks on X from billionaire investor Ken Griffin that raised a fundamental question: where will AI be in 1, 2, or 3 years? Reflecting on a sudden leap in productivity after seeing autonomous AI agents operating inside Citadel, CEO Ken Griffin explained that tasks traditionally performed by highly specialized teams with advanced academic training were increasingly being completed by AI systems at dramatically faster speeds.
According to Griffin, work that previously required teams of experts with master’s degrees and PhDs over weeks or months was increasingly being completed by AI agents in hours or days. What struck him most was not the automation of simple office work, but the growing ability of AI to enter domains long associated with expertise, analysis, and highly skilled knowledge professions.
These were analytical functions associated with elite knowledge work such as finance, advanced modeling, research, and decision support. Activities that once depended on years of education and institutional experience were increasingly being accelerated by systems capable of compressing months of work into dramatically shorter timelines.
Griffin later admitted the experience left him deeply reflective because it raised a larger question. If AI can already accelerate highly specialized work at this pace today, what happens when these systems become more autonomous, more integrated, and more capable over the next one, two, or three years?
That question may become one of the defining issues of this decade.
For years, the automation conversation focused mainly on repetitive labour and routine processes. The assumption was that machines would first replace manual work while knowledge professions remained insulated. Yet AI is increasingly moving into domains built around expertise itself.
Finance is adapting to it. Programming is reorganizing around it. Legal drafting is experimenting with it. Research institutions are accelerating because of it. Medicine and scientific discovery are beginning to integrate it.
The transition ahead may therefore affect more than labor.
It may reshape the world of expertise, data, and strategy itself.
Where AI Will Be in 1, 2, or 3 Years
Public debate around artificial intelligence still revolves around models. Conversations often focus on ChatGPT versus Gemini, Claude versus Grok, and open systems versus closed ecosystems. Yet beneath this visible competition sits a deeper layer that may ultimately shape the next phase of AI: infrastructure.
Artificial intelligence increasingly depends on systems that remain largely invisible to the public. Semiconductors, GPUs, data centers, cooling networks, cloud providers, energy grids, fiber infrastructure, and satellites now form the foundation beneath every model people interact with.
This reflects themes explored earlier in FTN’s The Invisible Infrastructure: How Satellites, Ocean Sensors, and AI Are Turning Earth into a Real-Time System, which examined how unseen systems increasingly influence economies, governance, and geopolitics.
The next generation of AI leaders may therefore not simply be those building the most advanced models. Instead, the ultimate edge may belong to those controlling the infrastructure that supports intelligence itself—the gigawatts of nuclear energy, the high-density cooling facilities, access to large volumes of rare earth and the global chip supply chains.
AI to Recover Lost Engineering Knowledge of Ancient Civilizations
Across centuries and continents, societies created structures that still challenge modern imagination because they demonstrate ambition at extraordinary scales.
The pyramids of Egypt remain among humanity’s most recognizable engineering achievements, while India’s Kailasa Temple represents another remarkable example. Unlike conventional architecture, where materials are assembled upward, Kailasa was carved downward from a single rock mass.
Historians estimate that hundreds of thousands of tons of stone were removed to create the structure, and it remains one of the most extraordinary engineering feats ever attempted.
Similar fascination surrounds the Rock-Hewn Churches of Lalibela, as well as Al-Khazneh and Ad Deir, whose monumental forms continue to raise questions about planning, labor organization, and construction capabilities. Even later structures such as St. Peter’s Basilica and the United States Capitol demonstrate civilizations that viewed architecture not merely as function, but as identity, continuity, and permanence.
Around some of these sites, alternative communities continue to discuss lost knowledge theories and Tartarian narratives. Mainstream archaeology does not support those interpretations, yet their persistence reveals something important: humanity remains captivated by the possibility that earlier civilizations possessed organizational or engineering capabilities that modern societies still underestimate.
Artificial intelligence may reopen this discussion, not by validating myths, but by reconstructing possibilities.
Future systems could simulate ancient construction methods, reverse engineer geometries, test engineering hypotheses, optimize structural designs, and digitally recreate lost techniques. Construction itself may increasingly become computational.
The next architectural revolution may therefore begin not only with new materials, but with intelligence.
AI May Become the Greatest Architectural Force Since Industrialization
Construction may emerge as one of AI’s least discussed yet most transformative frontiers.
For centuries, architecture relied primarily on human imagination working within engineering limits. Artificial intelligence introduces a third force: computational exploration.
Future systems may analyze thousands of years of architectural history while testing millions of design variations in hours rather than months. AI-guided systems could optimize materials, improve earthquake resilience, redesign urban layouts, model climate adaptation, and generate structures beyond conventional workflows.
As a result, construction may evolve from static planning toward adaptive intelligence. Entire cities could eventually operate as responsive systems where infrastructure continuously adjusts to population movement, energy demand, transportation flows, and environmental conditions.
This possibility overlaps with FTN’s earlier analysis in The New World Order Is Not Political It Is Systemic: How Energy, Data and Trade Form the Real Power Map, where infrastructure itself emerged as a strategic layer shaping the future. AI may now accelerate that transition.
Energy May Become AI’s Greatest Constraint
Every major AI breakthrough carries an invisible cost: electricity.
The expansion of frontier models, data centers, and agentic systems is driving extraordinary demand for power infrastructure. Training and operating advanced AI increasingly requires industrial-scale energy consumption.
This helps explain renewed interest in nuclear power. Microsoft announced plans connected to the former Three Mile Island site to support future energy requirements, while Google and Amazon have explored partnerships involving small modular reactors and next-generation energy systems.
The logic is becoming increasingly clear. AI growth drives electricity demand, electricity demand places pressure on grids, grid pressure reshapes energy planning, and energy planning increasingly becomes geopolitics.
The AI race may therefore evolve into an energy race. Nations controlling reliable electricity, nuclear expansion, grid infrastructure, and energy security may gain structural advantages.
Data Centers Are Becoming Strategic Assets
Data centers are increasingly becoming strategic infrastructure.
Questions surrounding who builds them, powers them, secures them, and supplies their semiconductors now carry geopolitical significance.
Competition between the United States and China increasingly reflects this transition. Rivalry extends beyond tariffs and markets into semiconductors, rare earth minerals, compute capacity, energy systems, undersea cables, and industrial ecosystems.
The AI competition therefore intersects directly with broader geopolitical themes already explored across FTN, including infrastructure dependency, trade systems, and strategic power.
The emerging divide may increasingly separate nations possessing compute capacity from those that do not.
AI Agents May Transform Organizations by 2028
The next three years may normalize AI agents, but this transition differs from today’s chatbots.
Agents increasingly move beyond conversation into execution. They may research, organize information, summarize documents, automate workflows, negotiate tasks, and manage systems.
The implications extend beyond productivity. Small teams may begin operating with capabilities previously reserved for large institutions, while entrepreneurs gain leverage that once required entire departments.
This transformation reflects Ken Griffin’s broader observation because the issue is no longer automation alone. It is amplification.
AI may increasingly amplify individuals, and that shift could reshape organizational structures themselves.
AI and the Faster Architecture of Civilization
Perhaps the most important change ahead will not be intelligence alone.
It may be speed.
Artificial intelligence could accelerate the pace at which civilizations design, build, coordinate, and expand infrastructure.
China already offers examples of infrastructure execution at extraordinary scale.
Chongqing East railway station, China’s newest and largest high-speed rail hub, spanning approximately 1.22 million square meters, is a testament to integrated infrastructure execution, where construction phases, heavy machinery operations, completed terminal spaces, and high-speed rail arrivals converge to demonstrate the scale and velocity of China’s ever expanding infrastructure.
Public reactions reflected both admiration and debate. Many viewers highlighted the speed and ambition behind China’s infrastructure development, while others questioned whether some visual sequences appeared AI generated.
Comparisons also emerged with slower and more expensive infrastructure projects elsewhere, including California’s long-running high speed rail development.
The broader significance extends beyond transport.
China has spent decades building high speed rail systems, logistics corridors, ports, industrial zones, power infrastructure, and large scale construction ecosystems. In an AI era, this execution model becomes increasingly important because intelligence itself depends on physical systems.
Data centers require land, electricity, cooling networks, transmission lines, semiconductors, and compute facilities. AI infrastructure increasingly becomes construction infrastructure.
Western economies often operate differently. Major projects frequently move through environmental reviews, public consultations, regulatory approvals, legal challenges, financing cycles, and political debate. These processes provide oversight and accountability but can also slow deployment timelines.
The debate around nuclear power increasingly reflects this tension.
As AI drives electricity demand, governments are reconsidering nuclear expansion, small modular reactors, and grid modernization. Yet building new capacity often requires years of licensing, environmental assessments, financing negotiations, and approvals.
The challenge ahead may therefore extend beyond building better AI models.
It may involve building civilization infrastructure faster.
Because the future AI race may not only reward intelligence.
It may reward execution.
A Civilization Shift Reshaping Our World
It’s wild to think about, isn’t it?
We spent the last few years treating AI like a cool productivity tool or a quirky parlor trick—worrying about ChatGPT writing student essays or generating funny images. But when you zoom out, you realize the tech companies aren't actually bracing for a software war. They are bracing for a heavy-industrial land grab. They are buying up nuclear power plants, cornering the market on copper and transformers, and treating data centers like sovereign territory.
Where AI will be in one, two, or three years remains fundamentally uncertain. The code is evolving too fast for anyone to map its destination. Yet, several tectonic trajectories are already becoming visible, proving that the next few years won't just change how we search the web or write emails—it is going to fundamentally rewire the physical grid of the planet.
AI is evolving into invisible infrastructure. Look closely, and you can see civilization itself beginning to embed intelligence directly into its physical foundations.
Construction is becoming computational, as generative algorithms optimize building shapes, material stress, and urban layouts before a single shovel hits the dirt.
Expert work is undergoing a major transformation; the era of the chatbot is giving way to autonomous AI agents that execute complex operations independently. Because of this, the traditional leverage of scale is evaporating, allowing small organizations to scale with capabilities once reserved exclusively for large institutions.
This level of artificial intelligence is not abstract. It requires a huge amount of power, and to support it, energy systems are rapidly changing. Nuclear power, once sidelined, has regained major strategic importance as tech companies bypass traditional grids to secure access to reactors.
Data centers are no longer just server warehouses; they are becoming high-density energy hubs and key assets in global geopolitics.
The future may therefore belong less to whoever builds the smartest model, and more to whoever builds the systems that sustain intelligence itself. The true competitive advantage is no longer just software code, it is industrial capacity.
Because the real AI competition is not about chatbots. It is about civilization infrastructure.
And infrastructure has always shaped history.
Comments