TECH NEWS

This Minisforum Thunderbolt 5 dock lets a mini PC run a full desktop graphics card for video editing, and it’s…

https://cdn.mos.cms.futurecdn.net/FgVKHLk2MYTWrGJVAhr4BK-2000-80.jpg

If you edit video on a mini PC and are tired of slow exports and choppy timelines, Minisforum has a solution that lets you plug in a full-size desktop GPU without sacrificing your compact setup.

We've had excellent experiences with Minisforum's own mini PC range, and right now, its DEG2 external GPU docking station is down to $219 (was $399) at Newegg when you use code SPCN6.

• Best for: video editors working on a mini PC or laptop with a Thunderbolt 5, USB4, or OCuLink port who want desktop-class GPU acceleration.

• Why we like it: The device features a full-length PCIe x16 slot for a desktop graphics card, both Thunderbolt 5 and OCuLink connections with a physical switch between them, and a spare M.2 slot for extra project storage.

• What you need to know: You'll need a mini PC with an OCuLink or USB4 port for...

Copyright of this story solely belongs to www.techradar.com. To see the full text click HERE

Read more

https://d2908q01vomqb2.cloudfront.net/f1f836cb4ea6efb2a0b1b99f41ad8b103eff4b59/2026/10/14/ML-21817-featured-image.png

Beyond the KV cache: Training and serving i14’s Laplace 3B linear attention model on AWS Trainium2 with native PyTorch | Amazon Web Services

Most large language models (LLMs) use transformer architectures with softmax attention. During inference, their key-value (KV) cache grows with the conversation or document context. This can reduce concurrency, increase serving cost, and force agentic applications to discard useful history. i14 designed Laplace 3B as a linear attention alternative. Its recurrent