Microsoft has just committed to buy 25 years of electricity from a recommissioned nuclear power station and in doing so has signalled something profound about our future. Technological progress is accelerating exponentially, and yet we live in a linear world. In this newsletter, I explore what Microsoft sees that others don't, and what it means for individuals, businesses, and policymakers.
Linear Thinking Doesn't Work In An Exponential World
Three Mile Island energy plant, the site of the worst nuclear accident in US history, is about to re-open with a sole customer.
Three Mile Island
Microsoft's nuclear power play seems like an outlier. A tech giant investing in old-school nuclear energy? But this bold move is just one example of how differently exponential companies think about the future. They see what others are missing: an approaching wave of energy-hungry artificial intelligence (AI) that will transform your life, career, or business.
Consider these rapid developments in AI:
In 2018, Google DeepMind's AlphaFold barely won a competition to predict 3D protein structures—a problem that had stumped scientists for 50 years.
By 2020, they won by a massive margin, effectively solving the problem.
A year later, they published predictions for all 200 million proteins known to science.
Two years later, their new AlphaProteo model can design novel proteins, transforming drug development.
This breakthrough earned DeepMind co-founder Sir Demis Hassabis a Nobel Prize.
The acceleration is even more dramatic in AI gaming. When Sir Demis, a former junior chess champion, played Google DeepMind's AlphaZero model—designed to self-learn two-player games—he could initially compete. Four hours later, it had become the world's best chess player, demonstrating capabilities beyond human comprehension.
The pace of technology adoption tells an equally striking story. While landline phones took nearly 25 years to reach widespread adoption in the UK, mobile phones needed only 10 years. ChatGPT? It reached a million users in five days and now boasts 200 million global active users. OpenAI's recent valuation exceeded £150 billion.
Just 12 months ago, generative AI video was a novelty. Now, major studios like Lionsgate are integrating generative AI into both pre-production and post-production. The democratisation of AI capabilities is equally remarkable. Until 2021, training specialised machine learning models for object recognition required significant resources. Today, general-purpose AI models like ChatGPT can analyse images with remarkable sophistication—identifying not just objects but understanding context and historical significance. I recently watched it analyse a photo of a room of lawyers, accurately describing not just the event but also detailing the historical heraldic imagery on the ceiling. This breakthrough means robots can now understand and navigate our world using generalist models.
The Great Acceleration
For most of human history, change was imperceptible from generation to generation. Homo sapiens lived relatively unchanged lives for 200,000 years. Even revolutionary inventions like the wheel, agriculture, and the steam engine gave societies generations, or at least decades, to adapt.
That era of gentle progress is over. Consider this accelerating pattern of technology adoption:
Power looms (1785): 50 years to industry-wide use
Telephone (1876): 50 years to widespread use
Television (1940s): 25 years to mass adoption
Personal computers (1980s): 15 years to mass adoption
The internet (1990s): 7 years to widespread use
Smartphones (2007): 5-10 years to global ubiquity
Facebook (2004): 4 years to reach 100 million users
Instagram (2010): 2 years to reach 100 million users
TikTok (2017): 3 years to reach 1 billion users
ChatGPT (2022): 5 days to reach 1 million users
AlphaZero: 4 hours from beginner to world's best chess player
Yet even these examples understate what's coming. For the first time in human history, we're not just creating tools that augment our abilities—we're creating intelligence that will be orders of magnitude more capable than ours.
This intelligence can now use our tools, from driving automobiles to operating software using a keyboard and mouse.
Intelligence: The Ultimate Multiplier
We are heading towards artificially general intelligence (AGI) that can autonomously learn and improve itself. AGI will quickly surpass humans in all cognitive tasks and design machines to exceed human physical capabilities. As it improves exponentially, artificial super intelligence (ASI) will emerge—a new species operating and designing solutions at a pace beyond human comprehension.
Consider the efficiency gap: A human researcher needs 20+ years of education to reach research capability, years to publish findings, and can only focus on a handful of problems simultaneously. It takes one-third of a human lifetime just to understand our scientific inheritance before building upon it. An AGI can instantly replicate knowledge across millions of instances, process thousands of research paths simultaneously, and share discoveries instantaneously.
The Intelligence Explosion
The transition from today's AI tools to AGI and then ASI won't follow comfortable timelines. As Oxford's Nick Bostrom notes, once AI can enhance its own intelligence, improvements will compound exponentially.
What makes this transition unprecedented?
Recursive Self-Improvement: AI systems designing better AI systems
Global Knowledge Integration: Instant sharing of all human knowledge
Parallel Processing: Millions of simultaneous improvements
No Biological Constraints: Development unlimited by human limitations
Exponential Convergence
While individual technologies advance exponentially, their convergence creates even more dramatic acceleration. Quantum computing and AI amplify our computational power and with that problem-solving capabilities. AI-optimised fusion research could revolutionise energy production. Biotechnology combined with AI and robotics is transforming everything from drug discovery to agriculture.
Today, drug development takes years and billions in investment, energy grids waste power during peak shifts, and education struggles with standardisation. Tomorrow, AGI could design affordable, personalised medicines in hours, quantum-AI systems could optimise energy grids in real time, and AI tutors could deliver customised education to every child globally—all simultaneously.
This isn't science fiction; it's the inevitable result of exponential technologies combining and compounding each other's capabilities.
The Productivity Quantum Leap
The numbers tell a compelling story. Historical productivity growth shows an accelerating trend:
First Industrial Revolution: 0.3% annually
Second Industrial Revolution: 1.2% annually
Information Age: 3% annually
Current AI Applications: 25-40% in specific domains
Near-Future AI: Potential 50%+ gains in knowledge work
But even these figures may be conservative. When AI can improve itself and work continuously, traditional productivity measures become meaningless.
What is clear is that the cost of human cognitive labour is going to zero.
The Energy Parallel
Even our energy systems demonstrate this pattern. While the transition from coal to electricity took a century and nuclear power needed decades, solar energy costs dropped 90% in just ten years. Now, AI-optimised renewable systems threaten to make traditional grids obsolete in sunny regions.
Yet even this revolution could be disrupted. AI-accelerated fusion research might leap ahead, potentially making both fossil fuels and current renewables obsolete within years, not decades.
Within one human lifetime, the cost of energy will tend towards zero.
Standing on an Exponential Curve
OpenAI CEO Sam Altman recently said that, when standing on an exponential curve, the line always looks flat when looking at the past, but always appears vertical when looking towards the future.
This creates a collective challenge: individuals, organisations, and policymakers base strategies on a past that increasingly has no relevance to our collective future. We expect linear progress from our current position on the exponential curve, consistently underestimating the pace of breakthrough achievements.
Exponential future, linear past.
As a qualified lawyer, I understand that law and regulation always lag behind societal reality. In the UK, effective consumer and financial services protection legislation arrived 50-75 years late in the 1980s. While AI-related legislation like the EU's Artificial Intelligence Act (2021) exists, it focuses on ethics and risk rather than keeping pace with technological advancement.
Of course, this unprecedented acceleration raises legitimate concerns about technological unemployment, economic inequality, and ethical governance. However, these challenges make proactive engagement more crucial, not less. The question isn't whether to embrace this future, but how to shape it responsibly.
What Must Be Done
The challenges of exponential change demand proactive action. This isn't a future problem—it's a now problem. Whether you're an individual navigating your career, a business leader shaping strategy, or a policymaker guiding society, today's actions will determine who thrives and who falls behind.
For Individuals
The era of mastering a single skill for a lifetime career is over. To stay relevant:
Embrace Continuous Learning: Make learning a lifelong habit, not a phase. Unlearn outdated practices and explore areas where AI and technology are reshaping industries.
Develop Adaptability as a Core Skill: Adaptability will be the new measure of intelligence. Those who can pivot quickly and reinvent themselves will flourish.
Focus on Uniquely Human Capabilities: Cultivate emotional intelligence, creativity, and strategic thinking—areas where humans currently excel over machines.
Learn to Collaborate with AI Systems: See AI as a partner, not a threat. Integrate AI into your personal workflow, from decision-making to creative projects.
Read my newsletter, 'Charting Your Career Path in the Age of AI'.
For Businesses
Exponential technologies demand exponential strategies. Linear thinking and rigid structures won't survive.
Shift from Linear to Exponential Planning: Prepare for rapidly escalating disruptions. Use scenario planning to anticipate and capitalise on shifts before competitors.
Build Flexible, AI-Ready Infrastructure: Invest in modular, scalable systems that can integrate evolving AI tools. Legacy systems will cripple innovation.
Invest in AI Integration Now: Delay means falling behind. Make AI central to business strategy, not just a pilot project. Develop your own AI OS for become 'AI-first'.
Create Adaptive Organisational Structures: Flatten hierarchies, empower agile teams, and foster rapid experimentation and learning.
Develop or Apply Exponential Technologies: Focus on AI and machine learning, biotechnology and genomics, clean energy and sustainability, quantum computing, autonomous systems and robotics, and blockchain/Web3.
For Policymakers
Stop assuming that the future will resemble the near past. Be brave and accept that policy-making must evolve to match technological disruption's pace.
Develop Anticipatory Regulation: As a former trustee of Nesta, that has looked at this, I know that it is not straightforward, but attempt to create evolving regulatory frameworks alongside technology, incorporating input from technologists, ethicists, and industry leaders.
Re-imagine Education Systems: Shift from rote memorisation to teaching adaptability, systems thinking, and digital literacy.
Create Flexible Economic Frameworks: Support new forms of labour and value creation as traditional jobs evolve or disappear.
Accept that employers will have an alternative: AI-driven cognitive and physical labour will soon be better, faster, cheaper and safer (read my newsletter on this).
Foster International Cooperation: Technology transcends borders; governance frameworks must too. Address AI ethics, cybersecurity, and global equity collaboratively.
A Call to Exponential Action
Exponential change isn't just reshaping industries—it's redefining what it means to be human. Preparing for this future requires collective action. We all shape this future. Individuals must cultivate adaptability, businesses must embrace bold innovation, and policymakers need to rethink their approach fundamentally.
The time for linear thinking and gradual adaptation is over. The time to start planning for an exponential, uncertain, but distinctly different future is now.
Choose wisely. Choose quickly. Choose exponentially.
Are you ready for the exponential future? Is your organisation adapting fast enough?
Thanks for reading.
Follow me on Bluesky.
