The Jolting Mindset: Why Exponential Thinking No Longer Explains Our Reality

The Jolting Technologies Hypothesis is a complete framework for interpreting an increasingly complex and rapidly changing world

What if the world isn't changing exponentially, but superexponentially? What if the acceleration itself is accelerating? What if the mathematical models we use to predict technological progress, guiding trillion-dollar investments, inform startup strategies, and shape government policies, are fundamentally wrong? These are not hypothetical questions. The evidence suggests we're living through a phase where progress is more than exponential. If true, then every framework based on constant doubling times is already obsolete. The difference between exponential and superexponential isn't subtle: it's the difference between having years to adapt and having months. This is the message of the Jolting Technologies Hypothesis.

When Ideas Meet Action at Zero Latency

Our tools are getting so good that the gap between having an idea and executing it is shrinking toward zero.

In the 1980s, if you wanted to test a business idea, you might spend months on market research, business plans, and funding rounds. Today, you can spin up a landing page, run ads, and validate demand in hours. If you wanted to create music, you needed instruments, recording studios, and distribution deals. Now, your laptop contains an entire production studio, and global distribution is one click away.

This shrinking gap has an important secondary effect. As we execute ideas faster, we can measure results and iterate more quickly. Each iteration teaches us something, leading to new ideas. This creates a feedback loop where opportunities accelerate. The cumulative value created over time is a function of how rapidly we iterate, and how many opportunities we test per iteration.

We are watching AI tools transform from curiosities to creative partners in real-time, the speed of change feels different, but accelerating in a way that defies conventional models.

From Exponential to Superexponential

Moore's Law Isn't Enough In A World Of Shrinking Doubling Times

The world is now increasingly shaped by the superexponential progression in the power of AI. Understanding exponentials like Moore's Law is not enough. This is why experts have to constantly revise their predictions, when they are surprised an AI capability arrived much sooner than they expected: they are only forecasting at an exponential rate of progress.

We're comfortable with exponential growth. Moore's Law taught us that computing power doubles every two years, and we've built our expectations around such predictable expansion. But what I am observing in AI feels fundamentally different.

Exponential growth is fast, but the rate of growth stays constant. If something doubles every two years, it will keep doubling every two years. Predictable, if dramatic already in its consequences.

But AI capabilities are not following this pattern. The doubling times themselves are shrinking. What took two years to double is taking one year, then six months, then three months. The acceleration is accelerating.

It is a fundamentally different growth pattern. When doubling times compress, progress blindsides us. Even the experts. A technology that seems decades away can arrive in years.

I coined the term "jolting technologies" because these sudden lurches forward feel like jolts to the system, and also in analogy to the name of the third derivative of position, the jolt. They're not the smooth curves we're used to planning around. They're discontinuous leaps that remake entire industries, or the world, overnight.

Are AI Agents Jolting?

The Jolting Technologies Hypothesis Predicts Superexponential Improvements.

The mathematics of Jolting Technologies starts from a simple generalization. If position is where you are, velocity is how fast you're moving, and acceleration is how quickly your speed changes, then jolt is how quickly your acceleration changes. It's the third derivative.

When jolt is positive and sustained, you're in a "jolting regime" of a particular phenomenon. The Jolting Technologies Hypothesis postulates that we are currently in a jolting regime for Artificial Intelligence.

To compare jolts across different scales, I define a normalized Jolt magnitude. This tells us the jolt intensity relative to current capability, making it meaningful whether we're talking about language models or image generators.

There is plenty of anecdotal support for the Jolting Technologies Hypothesis, with the CEOs of Microsoft, OpenAI, Anthropic, xAI, NVIDIA, all talking about shrinking doubling rates. It is however harder to collect more granular data.

One of the promising sources is the compilation of the ability of AI agents to complete tasks successfully. The research organization METR mapped and analyzed this, coming to the conclusion that the length of the tasks doubles every 7 months. If they have to revise their conclusion, shortening this period, it will be a strong experimental indication that the Jolting Technologies Hypothesis is currently valid.

The Complete Mathematical Framework

The Jolting Technologies Hypothesis a a rigorous mathematical framework for understanding how AI and other transformative technologies evolve. Here are the seven core equations that define this paradigm.

The Foundation: From Ideas to Infinite Opportunities (Equations 1-4)

1. The Zero-Gap Principle As our tools become infinitely powerful, the time from conception to execution vanishes. This is the seed of all acceleration.

2. The Opportunity Explosion When ideas execute instantly, opportunities don't just grow—they explode toward infinity. Each executed idea spawns new ideas in an accelerating cascade.

3. Cumulative Value Creation Total value isn't only about speed, but derives from the integral of iteration frequency times opportunities realized. Faster cycles compound into massive value creation.

4. The Jolt Definition The third derivative (the rate of change of acceleration) defines a jolt. When positive, technology isn't just accelerating; the acceleration itself is increasing. In technological terms we can think of an increasing rate of innovation.

The Advanced Framework: Real-World Complexity (Equations 5-7)

5. Multiple Interacting Jolts Reality is messier than single technologies. AI hardware jolts forward. Algorithms jolt independently. Data availability jolts. These interact amplify each others, through their respective breakthroughs. When language models improve, they accelerate research in vision. When chips get faster, they enable new algorithms. The total jolt exceeds the sum of parts.

6. Resource-Constrained Reality Infinite acceleration meets finite reality. As resource consumption approaches maximum capacity, even powerful jolts get dampened. This equation captures why progress eventually hits walls: energy limits, data exhaustion, talent shortages. Understanding these constraints helps predict when jolting periods will plateau.

7. Paradigm Shifts and Discontinuities Progress isn't smooth. When paradigms shift, everything changes. Think transformers replacing RNNs, or potentially AGI replacing narrow AI. The pre-paradigm trajectory gives way to an entirely new curve. These discontinuities make long-term prediction nearly impossible: you can't extrapolate across a paradigm shift.

Why This Matters Now

These equations are tools to guide us navigating our turbulent times.

• Equations 1-4 explain why AI progress keeps surprising us

• Equation 5 shows why AI advances compound across domains

• Equation 6 tells us which resources to watch as limiting factors

• Equation 7 prepares us for sudden, discontinuous leaps

We're living through a jolting period. Understanding these dynamics is essential for navigating a world where decades of progress compress into years, where multiple technologies jolt simultaneously, and where paradigm shifts lurk just beyond our prediction horizons.

The mathematics of jolts gives us a lens to see what's coming, even if we can't predict exactly when.

The Jolting Technologies Paper

After the development phase and feedback through peer review, I'm sharing the complete framework for understanding AI's superexponential growth.

You can read the academic paper about the mathematical foundations of Jolting Technologies on https://arxiv.org/abs/2507.06398

What's in the Paper

• Complete mathematical framework with formal proofs establishing when and how technologies exhibit jolting behavior

• Monte Carlo validation methods demonstrating robust detection even under high-noise conditions

• Hybrid detection algorithms combining peak analysis, pattern matching, and duration metrics

• Analysis of implications for AGI timelines, including why traditional exponential models systematically underestimate progress

What's in the Code Repository

The paper is complemented by a repository on GitHub: https://github.com/davidorban/jolting

• Python implementation of the complete jolt detection framework

• Simulation code for generating synthetic jolting time series

• Monte Carlo experiment suite with reproducible results

• Example notebooks for analyzing your own time-series data

• Visualization tools for derivatives and jolt magnitude

Key Insights for Different Communities

For Researchers: New mathematical tools for analyzing technological acceleration beyond simple exponential models. The third derivative provides a quantitative measure of "acceleration of acceleration."

For Policymakers: Traditional forecasting fails for AI because it assumes constant or smoothly changing growth rates. Jolting technologies require adaptive governance frameworks that can respond to sudden capability leaps.

For Technologists: We now have tools to quantify when incremental progress might suddenly become discontinuous. This helps identify which technologies are approaching phase transitions.

Abstract

This paper investigates the Jolting Technologies Hypothesis, which posits superexponential growth (increasing acceleration, or a positive third derivative) in the development of AI capabilities. We develop a theoretical framework and validate detection methodologies through Monte Carlo simulations, while acknowledging that empirical validation awaits suitable longitudinal data. Our analysis focuses on creating robust tools for future empirical studies and exploring the potential implications should the hypothesis prove valid. The study examines how factors such as shrinking idea-to-action intervals and compounding iterative AI improvements drive this jolting pattern. By formalizing jolt dynamics and validating detection methods through simulation, this work provides the mathematical foundation necessary for understanding potential AI trajectories and their consequences for AGI emergence, offering insights for research and policy.

I welcome feedback, collaborations, and applications of this framework to new domains. What technologies beyond AI might be jolting? Test the detection tools on your own data and share what you find.

The Choice of Recognizing Our Jolting Reality

Besides papers, math, or code, the real transformation happens in your mind. Stop thinking in smooth curves and start thinking in jolts. Prepare for compressions of time that will make decades happen in years, and years in months. The exponential mindset was adequate until recently, but it could become a liability. Every strategic decision, every investment, every policy you craft from just an exponential worldview is already obsolete. The jolting mindset is the model that matches reality. Those who make this cognitive leap will navigate the turbulence ahead. Those who don't will wonder why the future keeps arriving before they're ready. (I am not even considering those who are still around withe a linear mindset!)

The choice is yours: think exponentially and be perpetually surprised, or think in jolts and be prepared for the astonishing leaps that define our era.