+357 96 182730





A community calendar shared and populated by CBG members.



We list verified experts, funding options, and exlusive offers.



Free prizes offered up by magazines, radio stations, and more.


Coming Soon

Wouldn’t it be great to have a better way to network?

Discussion – 


Discussion – 


ChatGPT’s Energy Appetite May Spark a GPU Revolution in Business

Revolutionizing AI: The Innovations Shaping the Future of Computing

The landscape of artificial intelligence (AI) is evolving at an unprecedented pace, akin to the rapid hallucinations generated by ChatGPT. However, the soaring demand for Graphics Processing Units (GPUs) required for large-scale AI training has cast a spotlight on the escalating costs associated with pushing the boundaries of AI capabilities. OpenAI, the driving force behind ChatGPT, reveals that training the algorithm alone incurred a staggering cost exceeding $100 million. Additionally, the race to excel in AI is contributing to a concerning surge in energy consumption by data centers worldwide.

The GPU Conundrum

The AI gold rush has led to a surge in startups aiming to revolutionize the industry by creating innovative computational tools. While Nvidia’s GPUs currently dominate the AI hardware landscape, several up-and-coming ventures argue for a radical redesign of computer chips.

Normal Computing’s Paradigm Shift

One such startup, Normal Computing, founded by veterans of Google Brain and Alphabet’s X lab, presents a revolutionary prototype as the first step towards redefining computing from its foundational principles. Their stochastic processing unit (SPU) harnesses the thermodynamic properties of electrical oscillators, leveraging random fluctuations within circuits for calculations. Faris Sbahi, the CEO of Normal Computing, emphasizes the efficiency of their hardware and its suitability for statistical calculations. This novel approach holds promise for constructing AI algorithms capable of handling uncertainty, potentially mitigating the issue of large language models producing unreliable outputs.

Extropic’s Ambitious Thermodynamic Computing

Extropic, a stealth-mode startup founded by ex-quantum researchers from Alphabet, takes the idea of thermodynamic computing even further. Guillaume Verdon, Founder and CEO of Extropic, reveals their ambitious plan to integrate neural computing into an analog thermodynamic chip. Drawing inspiration from quantum computing, Extropic aims to usher in a new era of computing with a full-stack thermodynamic paradigm. This visionary approach seeks to push the boundaries of what AI can achieve, challenging the conventional norms of computing.

Rethinking Computing

The notion that a broader reconsideration of computing is imperative gains momentum as the industry grapples with the challenges of sustaining Moore’s Law. Even if Moore’s Law weren’t slowing down, the accelerating growth of model sizes, as witnessed in projects by OpenAI and others, poses a substantial hurdle. Peter McMahon, a professor at Cornell University specializing in innovative computing methods, underscores the urgency of exploring new computational avenues to keep the AI hype train on track.


As the pursuit of AI excellence continues, these groundbreaking ventures signal a paradigm shift in the way we approach computing. The convergence of thermodynamic principles and AI opens doors to unprecedented possibilities, challenging the industry to rethink its traditional models. The innovations discussed here not only address the current challenges but also pave the way for a future where AI capabilities transcend the limitations of existing hardware paradigms.


Submit a Comment

Your email address will not be published. Required fields are marked *

You May Also Like