The latest episode of IBM Tech Now covers the launch of Meta’s Llama 3.1, featuring the largest open-source language model with 405 billion parameters, alongside the implementation of the EU AI Act aimed at regulating AI applications based on risk. Additionally, it introduces a new IBM Cloud VPC sandbox for performance testing in a non-production environment, offering businesses the chance to optimize their cloud hosting with potential cost reimbursements.
In the latest episode of IBM Tech Now, the focus is on three significant developments in the tech landscape: the launch of Meta’s Llama 3.1, the implementation of the EU AI Act, and the introduction of a new IBM Cloud VPC sandbox. On July 23rd, Meta announced the release of the Llama 3.1 collection, which includes multilingual large language models available in various sizes: 8 billion, 70 billion, and a groundbreaking 405 billion parameters. The 405 billion parameter model is touted as the largest and most powerful open-source language model to date, positioning it as a competitor to leading proprietary models. This release aims to provide the global AI community with advanced tools for experimentation and development. Meta’s Llama 3.1 release
The Llama 3.1 models are designed to be both pre-trained and instruction-tuned, allowing users to tailor the models to specific use cases. This flexibility is seen as a significant advancement in the generative AI space, enabling developers to explore new ideas and approaches. The collaboration between IBM and Meta, established through the AI Alliance in 2023, aims to guide the evolution of AI in a way that reflects societal needs and complexities. The introduction of Llama 3.1 is viewed as a crucial step in this direction, offering an open-source model family that fosters innovation.
In addition to the advancements in AI models, the episode discusses the EU AI Act, which came into effect on August 1st. This legislation positions the European Union as a pioneer in establishing regulatory frameworks for AI, with the goal of creating a global hub for trustworthy AI. The Act adopts a risk-based approach, categorizing AI applications based on their potential risks to fundamental rights and safety. Key provisions include the prohibition of certain high-risk AI practices and the establishment of standards for the development and deployment of high-risk AI systems. The EU AI Act
The EU AI Act is characterized as a form of regulatory pragmatism, aiming to align constraints with the level of risk associated with different AI applications. The legislation is set to follow a rapid implementation schedule, reflecting the EU’s commitment to ensuring safe and responsible AI usage. For those interested in the specifics of the Act, further details are available through provided links.
Lastly, the episode highlights the introduction of a new IBM Cloud VPC sandbox, which allows customers to conduct performance testing in a non-production environment. This sandbox utilizes IBM Cloud virtual servers and offers the opportunity to test various applications, including those using second and fourth generation Intel Xeon processors. Customers can target specific benchmarks or bring their own applications for testing. For a limited time, participants can also receive reimbursement for their costs through IBM Cloud credits, making this an attractive option for businesses looking to optimize their cloud hosting environments. IBM VPC sandbox