Solving the GPU shortage: software to 2-4X existing GPU utilization

<p>GPUs are rare and expensive right now. Every company doing AI model training needs more, and NVIDIA can&#39;t build enough, especially of the NVIDIA H100 GPU.</p> <p><br></p> <p>Run.ai CEO and co-founder Omri Geller says he has a software solution to this hardware problem.</p> <p><br></p> <p>The key: GPUs are mostly idle, even in high-demand settings. According to Geller, his software 2X to 4Xs your GPU capacity on existing hardware, simply by streamlining workloads and maximizing GPU usage time.</p> <p><br></p> <p>In this TechFirst, we chat about:</p> <p> - the GPU shortage</p> <p> - how many GPUs we need</p> <p> - what Open AI is using right now</p> <p> - whether OpenAI is getting dumber or not</p> <p> - and much more</p> <p><br></p> <p>As always, get a full transcript and subscribe to the audio podcast at my website:</p> <p>https://johnkoetsier.com/category/tech-first/</p>