Apple's 'Minions' research turns idle Mac GPUs into a virtual GPU cloud

appleinsider.com Oct 15, 2026

Apple’s ‘Minions’ research turns idle Mac GPUs into a virtual GPU cloud

Curated by Shubham · TechCapsules · Oct 15, 2026

AI-summarised brief · reviewed before publication

Apple unveiled Minions, a system that aggregates idle Apple Silicon GPUs across Macs to form a virtual GPU cloud for AI inference. The architecture includes a central Minions Server, PostgreSQL metadata store, Minion Executors on macOS devices, and a watchdog for fault tolerance. Using Thunderbolt 5 RDMA, the setup combines compute and memory from multiple machines, reducing long inference jobs from over ten hours to minutes without extra hardware. Apple’s researchers already employ Minions internally, showcasing its scalability and efficiency.

💡 Why It Matters

· By turning everyday Macs into a high‑performance AI cluster, Minions eliminates the need for costly GPU farms, offering a cost‑effective, scalable solution that could reshape how businesses deploy machine‑learning workloads.

AI Apple Capsules

Read full article on appleinsider.com

Sweden Joins European Quantum and AI Skills Academies

Sweden has joined the European Quantum and AI Skills Academies, with Nordita leading the Nordic‑Baltic Quantum Academy and Swedish institutions spearheading AI education for healthcare and social care. The European Quantum Skills Academy (EQA) brings together 73 partners and over 100 affiliates across 23 countries, backed by a €19.8 million budget, half funded by the EU. The AI initiative, supported by Region Värmland, Experio Lab, and universities, emphasizes user participation, democratic principles, and responsible technology use.

💡 The move positions Sweden as a key player in Europe’s quantum workforce, ensuring the country’s expertise aligns with EU sovereignty goals. It also expands AI training in critical public sectors, fostering inclusive, ethically grounded innovation.

news.microsoft.com · Oct 15

Missed a lecture? This AI agent at Bar-Ilan University helps students catch…

Bar-Ilan University piloted an AI learning agent on its Moodle platform in February, allowing students to query and summarize course materials in Hebrew, Arabic and English. Built with Azure AI Services by KPMG and OpenApp using Microsoft Copilot Studio, the bot draws exclusively from faculty‑uploaded content—recorded lectures, transcripts, slides and articles—and can generate topic summaries, quizzes and answer checks. The trial involved 71 courses and roughly 2,400 students, and the university plans to extend it to another 100 courses for the 2026/27 academic year. Undergraduate Yarden Shay, who previously missed lectures, now uses the tool to create notes, test herself and reduce daily study time from five‑six hours to three, reporting greater self‑reliance and improved academic experience.

💡 By turning fragmented lecture assets into an on‑demand tutor, the system cuts study time and eases exam anxiety, demonstrating how campus‑wide AI can transform student support beyond mere content delivery.

appleinsider.com · Oct 15

Joswiak: Siri AI is for everyday, not the AI ‘lunatic fringe’

Apple senior vice president Greg Joswiak told Time that the company’s new Apple Intelligence and upgraded Siri are designed for everyday users, not for AI enthusiasts he described as the “lunatic fringe.” He emphasized that Apple will support users but its goal is to make the assistant useful. Joswiak said the integration capabilities of Apple’s ecosystem give it an advantage over competitors such as Claude or Grok. The comments came in a cover story about incoming CEO John Ternus, who spent 15 years developing the iPhone Duo announced on September 9, 2026, with pre‑orders starting October 16 and shipments on October 23. Apple has faced criticism for lagging in AI, and Joswiak’s remarks aim to reassure consumers that focus is on improvements than demonstrations.

💡 By prioritizing mass‑market usability, Apple aims to embed its AI deeply across devices, turning integration into a competitive moat.

techcrunch.com · Oct 15

AI agents are coming to Rivian’s cars

Rivian unveiled an AI assistant that lets occupants control vehicle functions and access information with voice commands. At Disrupt, CEO RJ Scaringe said upcoming models will support agents like Anthropic’s Claude, OpenAI’s Dots or Meta’s Muse, calling the prospect “magical.” The company built its own software stack instead of using Apple CarPlay or Android Auto, creating an internal “agentic framework” that routes requests to local edge models or cloud services as needed. As of May, the assistant can pull data from apps such as Google Calendar, and Scaringe hinted at deeper integration with external context‑rich agents, though he warned safety‑critical controls will remain isolated from AI. Rivian will officially detail the rollout at its AI & Autonomy Day in December.

💡 Enabling third‑party AI agents could turn Rivian vehicles into personalized digital hubs, differentiating them in a market where infotainment is a key buying factor.

thequantuminsider.com · Oct 15

Riverlane and NVIDIA Use AI to Speed Up Quantum Decoder Calibration

Riverlane and NVIDIA have demonstrated that AI‑driven calibration of quantum error‑correction decoders can be accelerated by using NVIDIA H100 GPUs. In tests, the Local Clustering Decoder’s surface‑code training cycle dropped from three minutes on a single CPU core to 2.8 seconds on an H100, while the GPU processed over four million syndromes per second versus about 59,000 on a CPU. This rapid recalibration could allow quantum systems to adapt to hardware drift more frequently, reducing idle time and improving logical error rates.

💡 Faster decoder calibration means quantum computers can maintain optimal error‑correction performance with less downtime, a critical step toward scalable, practical quantum applications.

appleinsider.com · Oct 15

OpenAI swears it won’t use the Apple trade secrets it seemingly admits…

A federal judge has indicated a strong likelihood of granting Apple an injunction against OpenAI over alleged trade‑secret theft. In response, OpenAI has pledged not to use any proprietary Apple information and will provide forensic images of drives from 14 former Apple employees. The move aims to limit potential damages while the case proceeds. OpenAI also countersued Apple, claiming the tech giant stole its own secrets, though Apple denies the allegations. The dispute remains pre‑trial, with discovery requests largely satisfied.

💡 By preemptively restricting its use of Apple data, OpenAI seeks to mitigate the impact of a possible injunction that could halt its AI device development. The case underscores the escalating legal battles over intellectual property in the rapidly evolving AI sector.