Jack Hidary, CEO of SandboxAQ, argues that the current focus on Large Language Models (LLMs) in artificial intelligence is too narrow, overlooking the broader potential and critical applications of AI. Speaking with CNBC's Sarah Eisen on "Squawk on the Street," Hidary emphasized that while LLMs are certainly impactful, a more expansive view of AI's capabilities is essential for investors and industry leaders alike.
Hidary, whose company focuses on both quantum and AI software, highlighted a significant transformation occurring globally. He noted that demand for AI is not confined to the digital realm but is also profoundly impacting the physical world. This shift is driving the need for advanced computational power, with AI for the physical world requiring solutions that go beyond traditional digital processing. "We're seeing a transformation of the Gulf region with AI, both AI for the digital world and AI for the physical world," Hidary stated.
A key insight from Hidary is the inefficiency in current AI energy consumption. He pointed out that in the United States, a substantial portion of electrons generated at night are effectively wasted. "We in the United States throw away 40% of the electrons that we create at night. Instead of throwing it away, we could have newer kinds of battery chemistries, we could have catalysts, we could have new kinds of materials science," Hidary explained. This inefficiency presents a significant opportunity for AI to optimize energy usage and resource management in critical industries.
The discussion also touched upon the role of hardware in advancing AI. Hidary specifically mentioned NVIDIA, acknowledging its significant contributions. He noted that NVIDIA's GPUs are crucial for the computational demands of AI, but also highlighted the emerging importance of quantum computing. "The AI for the digital world, like ChatGPT, Gemini, these are great tools, tools that are super useful, and that has created this massive challenge of creating all these data centers," Hidary elaborated. However, he stressed that the future of AI lies in a more hybridized approach.
"We've got to widen the aperture," Hidary urged, advocating for a broader perspective on AI's applications. He believes that focusing solely on LLMs misses the potential for AI to revolutionize sectors like materials science, drug discovery, and energy. The argument is that AI can be instrumental in discovering new materials, optimizing battery chemistries, and developing new catalysts, which are vital for addressing global energy challenges and advancing manufacturing processes.
Hidary’s perspective is that while LLMs are a powerful tool for digital applications, they are only one facet of AI's potential. The real transformative power of AI, he suggests, will come from its application in the physical sciences and industries. This requires not only advanced software but also specialized hardware, including quantum processors, to tackle complex problems that are currently intractable.
He further elaborated on the concept of "AI for the physical world," explaining its potential to solve real-world problems. "We can have newer kinds of battery chemistries, we could have catalysts, we could have new kinds of materials science," Hidary stated, underscoring the tangible impact of this branch of AI. This involves using AI to accelerate discovery and innovation in areas that directly affect our physical environment and infrastructure.
The interview also touched upon the ongoing debate about AI valuations. Hidary expressed concern that an over-reliance on LLMs might lead to an unsustainable bubble. "If we're just going to talk about getting returns on our investment based on people using it for document summaries and customer service, then I think we're not going to be happy. The bubble will not end well," he warned. This underscores his belief that the long-term value and sustainability of AI investments depend on their ability to solve complex, real-world problems across various sectors, not just digital ones.
Hidary also highlighted the potential for quantum computing to complement AI, particularly in areas like materials science and drug discovery. He pointed out that quantum computers can simulate molecular interactions and chemical reactions with unprecedented accuracy, a task that is computationally prohibitive for even the most powerful classical computers. By integrating quantum capabilities with AI, companies can accelerate the discovery of new materials, optimize chemical processes, and develop more effective drugs and therapies.
"If we're thinking about AI in bigger terms, in terms of the energy sector, in terms of the bio-pharma sector, in terms of logistics, in terms of human robotics, then I think we can actually make this AI trade work," Hidary contended. This broader vision of AI’s application is what he believes will drive sustainable growth and innovation, moving beyond the current hype cycle around LLMs.
The conversation with Hidary provides a critical perspective for investors and professionals in the AI space, urging a more holistic and forward-thinking approach. The emphasis is on recognizing AI's potential to solve fundamental challenges in the physical world, which will require a combination of advanced digital and quantum computing solutions.
