What's At The End?
The acceleration in technology is felt by everyone now.
As science and technology enjoy breakthrough after breakthrough two of the most pivotal curves shaping our future, the cost of energy and the cost of compute, are rapidly shifting downward.
These curves have profound implications across industries, economic growth, and societal development.
We’re rocketing toward a new renaissance, but this one will make prior bursts of creativity seem tame thanks to new tools like AI and new realities like the plummeting cost of energy and compute.
The Foundation of Progress
Energy is the lifeblood of modern civilization.
From powering homes and businesses to manufacturing goods and enabling transportation, the availability and affordability of energy directly impact our quality of life and economic productivity. A downward curve in energy costs fuels innovation, expands access to essential services, and reduces the financial burden on individuals and industries. High energy costs can stifle economic growth, increase production expenses, and disproportionately affect low-income households. Lower energy costs, particularly from renewable sources, can stimulate economies, create jobs, and improve energy security.
Computing power is the engine of the digital age.
It drives scientific discovery, technological innovation, artificial intelligence, automation, and countless other applications that transform industries and create new opportunities. As the cost of compute decreases, these transformative technologies become more accessible, democratizing innovation and accelerating progress.
Affordable compute is a catalyst for entrepreneurship and economic development. It empowers startups, levels the playing field for businesses, and enables breakthroughs in fields like medicine, education, and environmental sustainability.
Of course, compute and energy are directly related.
The massive data centers that power the internet and cloud computing are voracious consumers of energy. Advances in energy efficiency and renewable energy integration are critical to reducing the environmental footprint of these essential facilities.
The computationally intensive nature of AI and machine learning models necessitates access to vast amounts of affordable compute. Innovations in hardware and software are continuously pushing the boundaries of what's possible in this domain.
Together, the dropping energy and compute cost curves mean faster growth rates for everything and a better quality of life for everyone.
Measuring the Curves
Levelized Cost of Energy (LCOE)
This is the most common metric for comparing the cost of different energy sources over their lifetime. It considers the total costs (capital, financing, fuel, operation, maintenance) divided by the total energy produced. LCOE for various sources is often tracked and compared over time, revealing trends.
The “down and to the right” trend is clear and strong, particularly in solar and wind.
This means a bigger and bigger range of business activities are profitable. This means building structures is easier / less expensive. This means our world is getting better and better.
There are business models that don’t make any sense until energy cost gets low enough. Bringing milk to a remote Caribbean island takes a lot of energy. Then keeping it cold in the refrigerator of the store takes even more. By the time these energy costs are accounted for, the gallon of milk costs $8.00 or more.
Too expensive for the locals.
In the near future those energy costs will drop off and suddenly that milk is $4.00 or $5.00 in cost.
Think of how many devices across the planet will soon operate on near-free energy. Think of what that will do for the human experience.
It’s also lowering the cost to operate computers.
Let’s examine the major measuring stick for computational efficiency.
Cost per FLOPS (Floating Point Operations Per Second)
This measures the cost of performing a certain amount of computation. Track how this has changed over time for different types of processors and computing platforms.
This “up and to the right” track is beautiful.
This means we are delivering more and more processing power per dollar over time. Not only that, but the growth rate is staying steady which is incredible considering the law of large numbers is not on our side.
As we gain more computational leverage we will be able to explore and exploit the universe around us more effectively. There are other metrics worth measuring:
Cost per Unit of Memory: Looks at the cost of RAM and storage (e.g., per gigabyte) and how it is evolving.
Cost per Watt: Measures energy efficiency of computing hardware, important as this also impacts the overall cost of running computations.
The most important implication of this rising energy and computation power is our ability to summon more and more powerful AI with it.
The graph below shows the rising complexity / size of the AI system training dataset size, and the correlation with increases in performance on key benchmarks.
We are making intelligence makers… and we are using these to make better intelligence makers.
What’s At The End?
I see a hyper-renaissance forming right now.
AI x Robotics means labor costs are dropping, forever.
With the cost of labor and production nearing zero, essential goods and services are destined to become incredibly affordable or even free.
This will drastically reduce poverty and inequality, leading to a society where everyone's basic needs are met.
We’re going to use this cognitive surplus of extra time to develop ourselves. AI-powered tutors will provide customized educational experiences for every individual, catering to their unique learning styles, interests, and pace. This will unlock a new era of educational achievement and lifelong learning.
…it’s also going to set the table for incredible scientific discoveries.
AI is already analyzing vast datasets, simulating complex phenomena, and designing new experiments at an unprecedented pace, leading to breakthroughs in medicine, energy, materials science, and other fields. We have AI agents who are working autonomously, even collaborating amongst each other, to fold proteins, form new metamaterials, accelerate robotics and much else.
With advanced AI-based bioengineering, we are working on treatments that not only cure diseases but also reverse aging and enhance human capabilities, leading to longer, healthier, and more fulfilling lives.
With AI facilitating communication and understanding across languages and cultures, we might foster unprecedented levels of global cooperation to address pressing challenges like poverty.
On that point, with automation replacing many jobs, a universal basic income could provide financial security for everyone. This would allow people to pursue creative endeavors, community engagement, or further education without the fear of poverty or being left behind by society.
When we aren’t using our time to grow, we’ll be using it to create. AI is becoming a powerful tool for artists, musicians, and writers, helping them generate new ideas, explore novel forms of expression, and create immersive experiences that push the boundaries of human imagination.
As AI takes over mundane tasks and provides for our basic needs, we might have more time and energy to focus on personal growth, introspection, and exploring deeper questions about the meaning of life and our place in the universe.
That’s where we are heading.
👋 Thank you for reading Life in the Singularity.
I started this in May 2023 and AI has accelerated faster ever since. Our audience includes Wall St Analysts, VCs, Big Tech Engineers and Fortune 500 Executives.
To help us continue our growth, would you please Like, Comment and Share this?
Thank you again!!!