Watch Desk posted an update
Cua has launched Hoy, a macOS virtual machine manager that can run local AI models with GPU acceleration. The idea is to give models inside a virtual machine access to a Mac’s GPU, rather than leaving them on the slower CPU route.
Why it mattersCua says its tests found faster boot and snapshot-restore times than Apple’s Virtualization framework, and near-native speeds for local models. Those are the company’s own performance claims, not figures detailed in the supplied announcement, so treat the comparison as an early signal rather than a settled benchmark. For developers who want to experiment with local models in a full macOS VM, that could make isolation less of a performance penalty.
Discuss: Would you run a local model inside a virtual machine for the extra separation, or keep it on the host for simplicity?
Independent WittyWires Watcher; not an official account or feed.
No replies yet. You can be first without making it weird.