{"id":21408,"date":"2026-10-14T12:09:56","date_gmt":"2026-10-14T12:09:56","guid":{"rendered":"https:\/\/news.theck1.no\/?p=21408"},"modified":"2026-10-14T12:09:56","modified_gmt":"2026-10-14T12:09:56","slug":"apple-researchers-turn-idle-macs-into-a-virtual-gpu-cloud-with-minions","status":"publish","type":"post","link":"https:\/\/news.theck1.no\/?p=21408","title":{"rendered":"Apple researchers turn idle Macs into a virtual GPU cloud with \u2018Minions\u2019"},"content":{"rendered":"<p style=\"margin:0 0 1em; padding:0.6em 0.9em; border:1px solid #d0d7de; border-radius:6px; background:#f6f8fa; color:#444; font-size:0.9em;\"><strong>AI-rewritten:<\/strong> This is a summary of an article from Tom&#8217;s Hardware, rewritten by AI (Qwen, running locally) to make it easier to read. The facts come from the original article &ndash; read it for the full story.<\/p>\n<div style=\"margin-bottom:1em; color:#666; font-size:0.9em;\"><strong>Tom&#8217;s Hardware &bull;  Shane Downing  &bull; October 14, 2026<\/strong><\/div>\n<hr\/>\n<p>Apple researchers have developed an internal system called &quot;Minions,&quot; which turns idle Mac computers into a virtual GPU cloud to manage artificial intelligence workloads. As described in a recent arXiv paper, this system helps reduce wait times for stalled AI jobs that previously sat in queues of over ten hours. By distributing these tasks to networked Macs, the researchers report that request wait times can now drop to as little as minutes. The process is triggered by requests sent through company Slack channels, though there are concerns regarding data privacy when teams use devices outside their immediate group.<\/p>\n<p><!--more--><\/p>\n<p>The architecture consists of three main components: a Minion Server, a Minion Executor, and a Minion Watchdog. The server splits tasks into smaller subtasks stored in a PostgreSQL database, while the executor on each Mac pulls these tasks via REST APIs without a central scheduler. Each Mac runs specific jobs entirely on its own hardware, handling either embeddings with open-source Hugging Face transformer models or text generation using Ollama. A watchdog monitors heartbeats sent every five minutes to catch failures caused by sleep modes, network drops, or power loss, returning unfinished tasks to the queue for retry.<\/p>\n<p>The paper notes that this approach avoids new GPU requests and eliminates infrastructure costs, shifting scaling from manual to organic processes. While participation is informal, Apple mitigates confidentiality risks through team-scoped deployments with separate databases and Mac pools. This development addresses growing compute bottlenecks acknowledged by Apple in its July quarterly filing, which cited constrained supply and high costs for AI resources. Future work aims to extend Minions to distributed model training and potentially federate idle machines across organizations or the global community.<\/p>\n<div style=\"margin-top:2em; padding:1em; border-left:4px solid #0073aa; background:#f5f7fa;\">\n<p style=\"margin:0;\"><strong>Source:<\/strong> Tom&#8217;s Hardware &bull;  Shane Downing  &bull; October 14, 2026<\/p>\n<p style=\"margin:0.5em 0 0;\"><a href=\"https:\/\/www.tomshardware.com\/tech-industry\/artificial-intelligence\/apple-researchers-turn-idle-macs-into-a-virtual-gpu-cloud-with-minions-researchers-say-the-distributed-scheme-could-scale-globally-and-support-model-training\" target=\"_blank\" rel=\"noopener\">Read the original article at Tom&#8217;s Hardware &rarr;<\/a><\/p>\n<\/div>\n","protected":false},"excerpt":{"rendered":"<p>AI-rewritten: This is a summary of an article from Tom&#8217;s Hardware, rewritten by AI (Qwen, running locally) to make it easier to read. The facts come from the original article &ndash; read it for the full story. Tom&#8217;s Hardware &bull; Shane Downing &bull; October 14, 2026 Apple researchers have developed an internal system called &quot;Minions,&quot;<\/p>\n<p class=\"more-link\"><a href=\"https:\/\/news.theck1.no\/?p=21408\" class=\"themebutton2\">READ MORE<\/a><\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[7],"tags":[],"class_list":["post-21408","post","type-post","status-publish","format-standard","hentry","category-it-hardware"],"_links":{"self":[{"href":"https:\/\/news.theck1.no\/index.php?rest_route=\/wp\/v2\/posts\/21408","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/news.theck1.no\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/news.theck1.no\/index.php?rest_route=\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/news.theck1.no\/index.php?rest_route=\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/news.theck1.no\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=21408"}],"version-history":[{"count":0,"href":"https:\/\/news.theck1.no\/index.php?rest_route=\/wp\/v2\/posts\/21408\/revisions"}],"wp:attachment":[{"href":"https:\/\/news.theck1.no\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=21408"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/news.theck1.no\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=21408"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/news.theck1.no\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=21408"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}