Macro • Foundations

The Law of Accelerating Returns

Why technological progress is exponential, not linear — and how understanding this creates asymmetric advantage

By Marc Theiler 25 min read March 2026

Most people dramatically underestimate the pace of technological change. Not because they're unintelligent, but because human intuition is fundamentally wired for linear thinking. We evolved in a world where change happened slowly, predictably, and within the span of a human lifetime. A hunter-gatherer could expect the world to look largely the same when they died as it did when they were born.

That world no longer exists. And the gap between how we intuitively expect change to unfold and how it actually unfolds is perhaps the greatest source of mispriced opportunities in markets today.

Key Insight

The Law of Accelerating Returns states that the rate of change in evolutionary systems (including technology) increases exponentially over time. Each advancement creates the foundation for faster subsequent advancement.

The Exponential Blindspot

Here's a thought experiment that reveals the power of exponential thinking:

Imagine taking 30 linear steps. You'd end up about 30 meters from where you started — roughly the length of a basketball court. Now imagine taking 30 exponential steps, where each step doubles the previous one (1, 2, 4, 8, 16...). Where would you end up?

One billion meters. You'd circle the Earth 26 times.

This is not an abstract mathematical curiosity. This is literally how technology progresses. And the inability to internalize this reality is why most people consistently underestimate what's coming — and why those who understand it have asymmetric advantage.

Ray Kurzweil and the Pattern

Ray Kurzweil, the inventor and futurist, first articulated the Law of Accelerating Returns in 2001. But his insight came from decades of tracking technological progress across multiple domains and recognizing a consistent pattern: technology doesn't progress linearly — it compounds.

Consider computing power. In 1965, Gordon Moore observed that the number of transistors on integrated circuits doubled approximately every two years. This became known as Moore's Law. But Kurzweil recognized something deeper: Moore's Law was just one instance of a broader phenomenon. Before integrated circuits, there were transistors. Before transistors, relays. Before relays, vacuum tubes. Before vacuum tubes, electromechanical calculators.

Each paradigm reached its limits and was replaced by a new one that continued the exponential trajectory. The exponential isn't tied to any single technology — it's a property of information-based systems themselves.

The Data Speaks

When you plot technological progress on a logarithmic scale, the pattern becomes unmistakable:

These aren't coincidences. They're manifestations of the same underlying dynamic: when a domain becomes information-based, it begins following an exponential trajectory.

Why Our Intuition Fails

The human brain evolved to track change in a world where most things moved slowly. Our ancestors needed to predict where a prey animal would be in five seconds, not where technology would be in five years. Linear extrapolation served survival well.

But apply linear thinking to exponential systems and you get systematic errors:

This creates a predictable pattern: new technologies are initially dismissed as toys or impractical, then suddenly seem to take over the world "overnight." The iPhone launched in 2007 to mockery from established players. Twelve years later, smartphones had fundamentally restructured human civilization.

"We always overestimate the change that will occur in the next two years and underestimate the change that will occur in the next ten." — Bill Gates

The Investment Implications

Understanding exponential dynamics creates three distinct advantages for investors:

1. Earlier Recognition of S-Curves

Technologies follow S-curve adoption patterns: slow initial growth, rapid acceleration, then saturation. The exponential framework helps you recognize where on the curve a technology sits.

Bitcoin in 2010 looked like a toy. Bitcoin in 2013 looked like a bubble. Bitcoin in 2017 looked like mania. Bitcoin in 2024 looks like infrastructure. Each stage was predictable to those who understood the adoption curve of transformative technologies.

2. Conviction Through Volatility

Exponential technologies are volatile precisely because most market participants don't understand them. They oscillate between irrational dismissal and irrational exuberance. Understanding the underlying trajectory provides conviction to hold through drawdowns that shake out linear thinkers.

Amazon dropped 95% from its 2000 peak. Those who understood that e-commerce was on an exponential trajectory — and that Amazon was building the infrastructure layer — held through a decade of mockery to see 100x returns.

3. Identification of Convergence Points

The most powerful opportunities emerge when multiple exponential technologies converge. AI plus cheap compute plus abundant data creates something qualitatively different than any component alone.

We're currently witnessing the convergence of:

Each of these alone would be transformative. Together, they're restructuring civilization.

The Skill Investment Corollary

Here's what most macro analysis misses: the same exponential dynamics that create financial opportunities also create skill opportunities.

When a technology is on an exponential trajectory, the skills required to work with that technology become exponentially more valuable. But there's a lag — the market for skills is even less efficient than financial markets.

The Skill Investment Principle

In an era of accelerating returns, your learning rate must exceed the rate of technological change. The most valuable skill is the ability to rapidly acquire new skills as paradigms shift.

Consider the evolution of AI skills over just the past three years:

Those who understood the exponential trajectory invested in learning the next paradigm before it became obvious. By the time everyone agrees a skill is valuable, the arbitrage is gone.

Practical Application

How do you actually use this framework? Three practices:

1. Track Exponential Metrics

For any technology you're evaluating, find the key metric and plot it on a log scale. Is it following an exponential trajectory? What's the doubling time? Where are we on the S-curve?

For AI, track parameters in frontier models, cost per inference, and capability benchmarks. For crypto, track active addresses, transaction volume, and developer activity. For energy, track cost per kilowatt-hour and deployment rates.

2. Discount "Expert" Skepticism

Experts in established fields systematically underestimate disruptive technologies. They're anchored to linear extrapolation from current capabilities. When domain experts dismiss an exponential technology as "impossible" or "decades away," that's often a buy signal.

In 2014, experts said autonomous vehicles were 20+ years away. In 2016, experts said large language models would never have general capabilities. In 2020, experts said mRNA vaccines would take a decade to develop.

3. Position for the Next Curve

The current exponential becomes the foundation for the next one. The question is always: what does this enable that wasn't possible before?

Cheap computing enabled the internet. The internet enabled mobile. Mobile enabled the app economy. The app economy enabled the gig economy. Each wave was predictable from the previous one if you asked the right question.

Today: what does cheap AI inference enable? What does programmable money enable? What does abundant clean energy enable? The answers to these questions are the investment themes of the next decade.

The Meta-Skill

There's one skill that compounds faster than any other in an era of accelerating returns: the ability to learn rapidly and adapt to new paradigms.

This isn't about being a generalist or avoiding specialization. It's about building a cognitive infrastructure that can absorb new information quickly, recognize patterns across domains, and update beliefs when evidence demands it.

The specific skills you need will change. The meta-skill of rapid learning remains valuable forever.

Conclusion: The Asymmetric Advantage

Understanding the Law of Accelerating Returns isn't just intellectually interesting — it's a practical edge. While most market participants are anchored to linear extrapolation, those who understand exponential dynamics can:

The world is not changing at a constant rate. It's accelerating. And in an accelerating world, the gap between those who understand this and those who don't will only widen.

The question isn't whether you believe in exponential change. The question is whether you're positioned for it.