Extropic’s revolutionary probabilistic hardware is set to challenge Nvidia and other chipmakers, promising a game-changing boost in computing efficiency and sustainability.
Challenging the world’s most successful chipmaker with an entirely new type of computer chip may seem absurd, but it is no more ridiculous than the AI race itself.
Artificial intelligence (AI) has a rich history dating back to the 1950s.
The term 'computer science' was first coined in 1956, marking the beginning of AI research.
Since then, significant advancements have been made in machine learning, natural language processing, and neural networks.
Today, AI is integrated into various industries, including healthcare, finance, and transportation.
According to a report by Gartner, global AI spending is expected to reach $190 billion by 2025.
Extropic is not a normal startup. But then, these are hardly normal times. The company is developing a radical new kind of computer chip that harnesses the thermodynamic fluctuations that naturally occur within electronic circuits—and which are normally a headache for engineers—using them to perform highly efficient calculations with probabilities.
Extropy is a term coined by Max More in 1983 to describe the concept of self-improvement and enhancement through technology.
It emphasizes the idea that individuals can increase their intelligence, lifespan, and overall well-being through the use of advanced technologies such as artificial intelligence, biotechnology, and nanotechnology.
The extropian movement advocates for a future where humans can transcend their biological limitations and become something more than human.
Extropy is often associated with transhumanism and the development of life extension therapies.
Harnessing Thermodynamic Fluctuations
This chip might well find some takers as AI giants search for ever more computer power to build AI models that perform artificial reasoning. As we all worry about AI’s incredible energy demands, Extropic’s innovative approach could be a game-changer. The company aims to deliver a chip that is three to four orders of magnitude more efficient than today’s hardware, a feat that would make a sizable dent in future emissions.
The Science Behind Probabilistic Computing
A technical document provided to me by Extropic includes a signal from an oscilloscope showing a probabilistic bit or p-bit in action. A conventional computer bit is fixed as either a one or a zero. A p-bit has a certain probability of being in either state, and the oscilloscope image shows a p-bit flipping between 1 and 0. The crucial thing is that Extropic can control the probability that the bit will be in either state at any point in time.
By engineering interactions between several such p-bits, it is possible to perform more complex probabilistic computations. This innovation enables the development of a scalable, mass-manufacturable, and energy-efficient probabilistic computing platform.
Probabilistic computing is a paradigm that deals with the use of probability theory to perform computations.
It involves representing uncertainty and ambiguity in calculations, rather than relying on precise values.
This approach has gained popularity in machine learning and artificial intelligence applications, where dealing with uncertain data is common.
Probabilistic algorithms can handle noisy or missing data, making them suitable for real-world problems.
They also enable the modeling of complex systems and decision-making under uncertainty.
Controlling Thermodynamic Effects
One of Extropic’s innovations is a way of controlling thermodynamic effects in conventional silicon to perform calculations without extreme cooling. Efforts to compute thermodynamically have traditionally relied on superconducting electronic circuits, but ‘Verdon and his cofounder, Trevor McCourt, are using fluctuations of electric charge in regular silicon instead.’
Applications for Monte Carlo Simulations
Extropic’s hardware is perfect for running Monte Carlo simulations, a class of computation that involves sampling probabilities. These computations are widely used in areas like finance, biology, and AI. The company aims to deliver a chip that can support the most computationally-hungry workloads, including those used in building reasoning models like ‘OpenAI o3 and Gemini 2.0 Flash Thinking from Google.’
A New Era for Computing
The founders of Extropic concede that the idea of taking on Nvidia and other chipmakers might seem, on the face of it, absolutely insane. Nvidia’s chips are still the best for training AI, and switching to a completely alien architecture would be costly and time-consuming. However, we are at a unique moment when AI companies need so much computer power for AI that they are building datacenters next to nuclear power stations.
Perhaps, given all this, it is more nuts not to try to reinvent how computers work. With the technology’s environmental impact only getting worse, Extropic’s innovative approach could be just what we need to revolutionize computing and create a more sustainable future.
- wired.com | How Extropic Plans to Unseat Nvidia
