October 21, 2023

October 20, 2023

October 22, 2023

  • Hardware

    Light Speed Ahead: 3D Photonic-Electronic Hardware Revolutionizes AI

    SciTechDaily, 10/21/23. Researchers from the University of Oxford and other universities have developed an integrated photonic-electronic hardware capable of processing 3D data. This breakthrough innovation significantly improves data processing parallelism for AI tasks. By using multiple radio frequencies to encode data, the approach enables multiple calculations to be carried out in parallel. This method shows promise for outperforming current electronic processors and offers the potential for further enhancements. The researchers anticipate a 100-times enhancement in energy efficiency and compute density, providing new possibilities for AI and machine learning applications. READ THE ARTICLE

  • Healthcare

    AI startup helps insurers spot cognitive decline in elderly drivers

    Japan Times, 10/21/23. Mind Foundry, an artificial intelligence startup, has successfully raised $22 million in funding, a testament to the growing demand for AI in critical sectors. The company’s AI tools are revolutionizing the insurance industry by detecting cognitive decline in older drivers, helping to predict and prevent accidents. Aioi Nissay Dowa Insurance Co., one of Asia’s largest insurance giants, has invested in Mind Foundry’s funding round, along with Parkwalk Advisors and the University of Oxford. This Series B funding brings Mind Foundry’s total funding to $44 million and showcases the increasing influence of AI on insurance coverage decisions. READ THE ARTICLE

  • IP

    Copyright-Free Sandbox: Implications of AI-Generated Content

    Patent, 10/21/23. The proliferation of synthetic works in our digital landscape is challenging the conventional wisdom of presuming all online content as protected by copyright. While some of these AI-generated works may lack originality, a significant portion could meet the originality standard if not for the absence of a human author. This emerging copyright-free realm presents opportunities for innovation and creation, but also challenges in discerning uncopyrighted works from copyrighted ones. It may be prudent to shift the burden back onto copyright claimants to explicitly mark their works. Additionally, legal mechanisms like contracts could play a role in safeguarding the interests of AI-generated works. READ THE ARTICLE

  • Nvidia

    Nvidia unveils new AI agent that trains robots

    The Hindu, 10/21/23. Nvidia Research recently unveiled their AI agent, Eureka, which can generate algorithms to train robots. By leveraging GPT-4’s natural language capabilities and reinforcement learning, Eureka enables robots to learn complex skills. The agent successfully trained a robotic hand to perform pen-spinning tricks, open drawers, and perform various other tasks without human prompting. Eureka represents a significant advancement in integrating generative and reinforcement learning methods to solve challenging tasks. Nvidia has open-sourced Eureka’s code and benchmarks for developers to build upon. This development showcases the potential for autonomous AI agents to expand beyond chatbots and contribute to various fields. READ THE ARTICLE

  • OpenAI

    Just how restrictive is OpenAI’s DALL-E 3 on ChatGPT?

    Mashable, 10/21/23. OpenAI has introduced DALL-E 3 for ChatGPT Plus subscribers, allowing users to generate images based on their ideas. This advanced image generator is capable of rendering hands, feet, and faces, expanding creative possibilities. However, there are strict ethical and safety guidelines in place to prevent the generation of harmful, violent, or adult content. While these safeguards ensure responsible use, they can sometimes limit creative expression. Despite this, some users have found clever ways to navigate the restrictions. OpenAI’s proactive stance in addressing ethical concerns is commendable in the evolving landscape of AI-generated content. READ THE ARTICLE