Originally aired on October 26, 2023
This episode of the Edge of AI covers several key topics in the world of artificial intelligence and technology. The episode starts with an exciting breakthrough from Primrose Metaverse, as they unveil an AI model that can decode brain signals to recreate images in real-time. The usefulness of this technology for clinical applications is discussed, particularly in helping individuals who have lost their ability to speak due to brain lesions. Next, the episode turns to Chinese tech giant Baidu, who announces the launch of their latest generative AI model, Ernie 4.0. This new version of the AI model is positioned as a competitor to OpenAI’s GPT-4 and offers capabilities such as crafting entire novels and generating videos. Additionally, Baidu introduces Dart bot, a chatbot powered by generative AI, which will handle complex accounting questions for their audit and assurance professionals. The focus then shifts to the collaboration between Foxconn and Nvidia, as they announce their plans to create AI factories that will utilize Nvidia chips and software for AI applications, including autonomous vehicles. IBM also extends their partnership with Amazon Web Services to develop generative AI solutions. The episode concludes with highlights from Stanford University’s transparency index, which raised concerns about declining transparency in AI models created by big tech companies. The rankings of various foundation models from companies like OpenAI and Google are discussed, pointing to potential risks in AI development.
Primrose Metaverse’s breakthrough in brain imaging AI
- Primrose Metaverse achieves a significant milestone with their Naro Imaging AI model that can decode brain signals to recreate images in real-time
- The AI system aims to aid individuals who have lost their ability to speak due to brain lesions
- The research marks a step forward in understanding human intelligence and has potential applications in clinical settings
Baidu’s launch of Ernie 4.0 generative AI model and Dart bot
- Baidu launches Ernie 4.0, a generative AI model that can design advertising materials, generate videos, and craft entire novels
- Dart bot, powered by generative AI, will handle intricate accounting questions for Baidu’s audit and assurance professionals, enhancing workforce productivity
Foxconn and Nvidia’s collaboration on AI factories
- Foxconn and Nvidia join forces to create AI factories using Nvidia chips and software for various AI applications, including autonomous vehicles
- The AI factories will collect and process data from autonomous electric vehicles in real-time to enhance decision-making capabilities
IBM and Amazon Web Services partnership for generative AI solutions
- IBM extends their partnership with Amazon Web Services to develop generative AI tools for critical business use cases
- The collaboration aims to enhance generative AI capabilities and train 10,000 consultants by the end of 2024
Middleware’s launch of AI-powered Cloud observability platform
- Middleware launches its Cloud observability platform utilizing OpenAI’s GPT-4 for AI-driven anomaly detection and resolution
- The platform’s AI adviser can pinpoint infrastructure and application problems, propose efficient solutions, and predict errors
Stanford University’s transparency index rankings
- Stanford University releases a transparency index ranking AI foundation models created by big tech companies
- Meta’s AI model (Llama 2) scores 54 out of 100, while OpenAI’s GPT-4 scores 48 and Google’s Palm 2 scores 40
- Amazon’s Titan receives the lowest rank with a score of 12, raising concerns about declining transparency in AI models