Israel Lags in AI Infrastructure Investment, Entrepreneurs Seek US Funding
Translated & summarized from Globes by baba
Professor Amnon Shashua, founder of Mobileye and DoubleAI, stated that Israel is struggling to invest in the necessary AI infrastructure, causing entrepreneurs to seek funding in the US. His new company, DoubleAI, recently solved a 38-year-old mathematical problem for under $2,000, showcasing a rigorous AI verification process. Shashua also discussed advancements and challenges in physical AI, including autonomous vehicles and human-like robotics. The company's AI has demonstrated significant capabilities in software optimization, quantum computing, and hardware design.
The story in 6 lines · by baba
- Professor Amnon Shashua warns Israel is underinvesting in AI infrastructure, pushing entrepreneurs to the US.
- DoubleAI, founded by Shashua, solved a 38-year-old mathematical problem for under $2,000.
- Shashua's AI approach emphasizes rigorous verification, differentiating it from other AI models.
- The company's AI achieved significant performance gains in rewriting Nvidia's CUDA software.
- Challenges remain in human-like robotics, particularly regarding AI generalization and safety.
- DoubleAI's AI has also shown prowess in designing quantum circuits and hardware.
Professor Amnon Shashua, the founder of Mobileye and a new AI company called DoubleAI, stated that Israel is falling behind in the artificial intelligence era due to a lack of significant infrastructure investment. Unlike the traditional high-tech industry, which thrived on human intellect and innovation without massive capital, AI demands substantial computing infrastructure. Shashua noted that while Israel excels at creating solutions with limited resources, the current AI revolution requires a different scale of investment, prompting many Israeli entrepreneurs to establish their companies in the United States where funding is more readily available.
Shashua made these remarks at the annual Journey conference hosted by EY, where he discussed the need for Israel to develop a smart strategy for resource allocation in AI. He highlighted that despite various initiatives, concrete steps are lacking, and AI necessitates massive investments. Yair Shanir, a partner at Michael Dell's DTC fund, echoed Shashua's concerns, sharing his initial frustration with Chinese AI advancements but finding hope in Shashua's new venture.
DoubleAI recently announced a breakthrough in solving a 38-year-old mathematical problem, the "longest path problem" in graph theory, for less than $2,000. Shashua explained that his company's approach involves a rigorous verification process for AI agents, distinguishing it from other AI models that may optimize for a target metric without ensuring correctness. This method, he believes, is crucial for AI to tackle complex real-world problems beyond mathematics.
Shashua also addressed the "physical AI" domain, distinguishing between autonomous vehicles and human-like robotics. He noted progress in autonomous vehicles, citing Mobileye's advancements, but pointed out that human-like robotics faces a significant challenge in generalization, teaching robots complex tasks from simple demonstrations that can be applied in diverse environments. He also touched upon the safety concerns associated with integrating robots into human environments, emphasizing the need for careful regulation.
DoubleAI's success in solving the mathematical problem involved several milestones, including rewriting Nvidia's CUDA software library, which resulted in a 3.6-fold performance increase. The AI also designed a quantum circuit capable of breaking Bitcoin encryption and excelled in hardware design, demonstrating its broad capabilities. The company's paper on the graph theory problem was written entirely by AI, with only the agent's name listed as a co-author.
Mentioned
