Intelligence Update
Open-weights models fuel appetite for on-premises AI
Last week, Meta released a new large language model (LLM) called Muse Glimmer, designed to run locally on devices such as laptops and desktops. The model can be downloaded at no cost and used for nearly any purpose. Larger and more capable versions of Muse for deployment on server hardware are expected in the coming months.
The announcement highlights the divergence between two approaches to AI that have a significant impact on where and how AI workloads are hosted. The key difference between them is whether the "weights" — the parameters of the model — are shared publicly so they can be used and even modified (open weights) or remain the property of the developer (closed weights).
Request an evaluation to view this report
Apply for a four-week evaluation of Uptime Intelligence; the leading source of research, insight and data-driven analysis focused on digital infrastructure.
Already have access? Log in here