Navigating LLM Deployment: Tips, Tricks and Techniques by Meryem Arik at Qcon London
Initial proofs of concept benefit from hosted solutions, but self-hosting is necessary for scaling models to cut costs, enhance performance, and meet security needs.
Using quantization and optimizing inference can help maximize GPU resources and efficiency in deploying Large Language Models. [ more ]
Clock is ticking on AI Act compliance as EU law progresses
Organizations using AI technology in Europe may have a limited time frame of six to 12 months to comply with upcoming legislation.
The EU's AI act, which sets regulations for AI systems, could be published and adopted in May, with compliance requirements staggered based on AI categories. [ more ]
California's Workplace Violence Prevention Plan Deadline is in Less Than Two Weeks
California employers must have a Workplace Violence Prevention Plan in place by July 1, 2024, with specific training, inspection, and incident log requirements. [ more ]