SI
SI
discoversearch

We've detected that you're using an ad content blocking browser plug-in or feature. Ads provide a critical source of revenue to the continued operation of Silicon Investor.  We ask that you disable ad blocking while on Silicon Investor in the best interests of our community.  If you are not using an ad blocker but are still receiving this message, make sure your browser's tracking protection is set to the 'standard' level.
Strategies & Market Trends : 2026 TeoTwawKi ... 2032 Darkest Interregnum
GLD 408.23+2.3%Dec 22 4:00 PM EST

 Public ReplyPrvt ReplyMark as Last ReadFilePrevious 10Next 10PreviousNext  
From: Julius Wong1/30/2025 8:20:00 PM
  Read Replies (3) of 218686
 
Deepseek is coming to Windows Copilot+ PCs

Reporter
Thu, Jan 30, 2025, 8:00 AM EST·2 min read

Steve Dent for Engadget

Microsoft is closely associated with OpenAI's ChatGPT AI model, but the software giant has no qualms about playing the field. Microsoft announced that it's bringing the DeepSeek-R1 AI model to Copilot+ PCs soon, starting with Snapdragon X devices and following later with Intel Lunar Lake and AMD Ryzen AI 9 PCs. The DeepSeek-R1-Distill-Qwen-1.5B model will arrive "soon" on Microsoft AI Tookit for developers, with more powerful 7B and 14B variants coming later.

The 1.5B (base) model isn't powerful compared to the higher-tier 32B and 70B models, but MIcrosoft points out that the models are "NPU-optimized" for Copilot+ PCs. The minimum configuration for such computers is 256GB of storage, 16GB of RAM and an NPU with at least 40 TOPS (trillions of operations per second).

"These optimized models let developers build and deploy AI-powered applications that run efficiently on-device, taking full advantage of the powerful NPUs in Copilot+ PCs," Microsoft wrote. It added that it implemented systems to take advantage of low-bit processing to ensure the R1 models could run locally on NPU hardware.

At the same time, Microsoft is bringing DeepSeek's R1 model to its Azure AI Foundry platform, The Verge reported. It joins other AI models on that service, including GPT-4, Mistral AI, Meta-Llama 3 and others. That comes as a bit of a surprise, given that Microsoft is reportedly probing whether DeepSeek used OpenAI's technology in an unauthorized manner.

AI pundits have also expressed concerns about privacy issues around China-based DeepSeek, something that Microsoft addressed in a Marketplace Community post. "DeepSeek R1 has undergone rigorous red teaming and safety evaluations, including automated assessments of model behavior and extensive security reviews to mitigate potential risks," wrote Microsoft senior product marketing manager, Justin Royal.

DeepSeek shook up the AI world with its R1 model, which doesn't require nearly as much computing power as competing models. That spooked markets yesterday, causing a selloff in chip giant NVIDIA and other AI-adjacent stocks. OpenAI, which has been sued by multiple newspapers and publishers around the world for copyright infringement, recently accused DeepSeek and other Chinese AI startups of "distilling" its models.

engadget.com
Report TOU ViolationShare This Post
 Public ReplyPrvt ReplyMark as Last ReadFilePrevious 10Next 10PreviousNext