Anthropic has announced a breakthrough in its technology transition, utilizing its in-house AI model, Mythos, to port its Claude model from Nvidia’s CUDA architecture to AMD GPUs. This process, which traditionally took months, was completed in just three days with minimal human intervention.
This rapid switch is part of Anthropic’s strategy to diversify its hardware usage, aiming to run Claude on four different chip families. By reducing dependency on Nvidia, the company seeks to cut costs and enhance operational flexibility in a competitive AI landscape.
Anthropic’s achievement highlights the growing capabilities of AI in automating complex processes, potentially reshaping the dynamics of the AI hardware market.
Compiled automatically by the Tech AI Newsdesk from public AI-news sources and summarised in our own words.
