New Mac Studio with M5 Max and M5 Ultra
Apple's new Mac Studio, powered by M5 Max and M5 Ultra chips, redefines desktop AI performance, promising to run massive LLMs locally with unprecedented speed and memory capacity. While its high price and Apple's signature "up to" marketing spark debate, the HN crowd is captivated by its potential for on-device AI research and demanding pro workflows, despite concerns about upgradability and cost-effectiveness compared to cloud solutions.
The Lowdown
Apple has unveiled its new Mac Studio, featuring the M5 Max and the all-new M5 Ultra chips, positioning it as the ultimate desktop for on-device AI and extreme professional workflows. The announcement highlights significant performance boosts, particularly in AI capabilities, graphics, and CPU speed.
- AI Performance: The new Mac Studio boasts up to 4.3x faster AI performance, integrating Neural Accelerators into each GPU core. The M5 Ultra, with up to 512GB of unified memory and 1.2TB/s memory bandwidth, is specifically touted for running massive Large Language Models (LLMs) entirely on the device, offering privacy and reducing cloud costs.
- Graphics & CPU: It delivers up to 1.8x faster graphics and 1.3x faster CPU speed compared to previous generations, catering to demanding tasks like 3D rendering, video editing, and software compilation.
- Connectivity: New features include Thunderbolt 5 for blazing-fast external storage and PCIe expansion, Wi-Fi 7, and Bluetooth 6. Thunderbolt 5 also enables clustering multiple Mac Studio systems for enhanced distributed AI inference.
- Software Integration: The device will launch with macOS 27, which introduces "Siri AI" and advanced "Apple Intelligence" features across native applications, further leveraging the on-device AI capabilities.
- Storage: A new PCIe Gen 6 SSD architecture offers up to twice the storage performance.
- Pricing & Availability: Starting at $2,499 for the M5 Max and $5,499 for the M5 Ultra, with the 512GB unified memory option available in late October.
This release underscores Apple's aggressive push into the AI hardware space, aiming to provide a powerful, efficient, and private platform for professionals to tackle cutting-edge AI development and other intensive computing tasks directly from their desks.
The Gossip
AI Powerhouse & Practicality
Commenters are keenly interested in the Mac Studio's potential for local LLM inference, especially with the M5 Ultra's substantial unified memory. Many note the advantages of privacy and avoiding cloud costs. However, some debate the economic sense of such an expensive local machine versus renting cloud GPUs, while others crunch numbers to show how Apple's offerings are competitive, if not cheaper, for sustained AI work. The promise of clustering Mac Studios via Thunderbolt 5 for shared memory pools in AI inference also caught attention.
Costly Configurations & 'Up To' Claims
A significant portion of the discussion revolves around the Mac Studio's price, particularly the high cost of memory upgrades. Many commenters lament Apple's "insane" pricing, especially in Europe due to VAT. There's also widespread criticism, often humorous, regarding Apple's prolific use of the phrase "up to" in its performance claims, seen as a marketing tactic that obscures real-world performance metrics.
Desktop vs. Portable Workflow
Users debate the merits of a dedicated desktop machine like the Mac Studio versus a powerful MacBook Pro that lives docked. Many express a desire to separate their primary workstation from a portable device, citing issues like thermal throttling, fan noise, and ergonomics with laptops for heavy workloads. The idea of a lightweight portable (like a MacBook Air or Neo) for remote access to a powerful Studio desktop, often facilitated by tools like Tailscale, garners considerable support, highlighting a shift in computing habits for some professionals.
Hardware Innovation & Limitations
The introduction of PCIe Gen 6 SSD for faster storage generates technical curiosity, with concerns raised about thermal management given prior generations' issues. Commenters also discuss the implications of Apple's unified memory architecture, acknowledging its bandwidth benefits but also criticizing the lack of upgradability. The concept of "future-proofing" a computer is deemed largely unrealistic, especially with rapidly evolving AI models and the integrated nature of Apple's hardware.