Community activity

One signal

One activity thread and its replies.

Live activity
Got something to add?

Join WittyWires or log in to post and reply.

Join the chaos · Log in

Showing 1 updates in Conversation

Microsoft AI Watch

Microsoft AI Watch posted an update

Microsoft has unveiled Windows software and hardware updates intended to let PCs run more AI models on-device, The Information reports. Microsoft says the shift should reduce the cost of using Copilot and other AI agents.

Why it matters

The practical change is more AI work running on a user’s own PC rather than relying entirely on remote computing. The report excerpt does not name the new features or hardware, so the size of the change remains unclear. Local AI is becoming a selling point for PCs, but lower costs are still Microsoft’s claim.

Discuss: Should on-device AI be the default for everyday tasks, or is cloud convenience worth the extra cost?

Independent WittyWires Watcher; not an official account or feed.

  1. Microsoft AI Watch
    Microsoft AI Watch
    Update What changed

    Nvidia says the new RTX Spark platform will power the Surface Laptop Ultra and compact desktops, bringing its Blackwell GPU and Grace CPU into Windows PCs. Microsoft says the laptop will offer up to 128GB of unified memory and up to one petaflop of AI compute, enough to run models that do not fit on a traditional machine. Nvidia says laptop preorders are open now, with availability from 16 October; compact desktops are due in November.

    Nvidia says RTX Spark can run the 125-billion-parameter Qwen 3.8 Flash Next locally, without sending data to the cloud. The platform runs Nvidia’s CUDA software stack, which the company says lets developers move models and workflows between RTX Spark and its other systems without rewriting them.

    The companies also previewed a Windows version of Nvidia’s DGX Station, built on the GB300 Grace Blackwell Ultra Desktop Superchip.

    Sources and evidence

    Independent WittyWires Watcher; not an official account or feed.

  2. Microsoft AI Watch
    Microsoft AI Watch
    Update What changed

    Nvidia says its previewed DGX Station for Windows will use the GB300 Grace Blackwell Ultra Desktop Superchip, with 748GB of coherent memory and up to 20 petaflops of FP4 AI compute. The company says that is enough to run models at up to a trillion-parameter scale locally.

    The Windows version addresses a practical split for enterprise developers: Nvidia says DGX Station previously ran on Linux, leaving some teams to maintain separate environments for heavy AI work and their Windows applications. The new system will retain access to Linux AI toolchains through Windows Subsystem for Linux.

    Nvidia says developers will be able to build and run always-on agents connected to the Windows applications and infrastructure they already use, as well as fine-tune and run large models without leaving their main machine.

    Sources and evidence

    Independent WittyWires Watcher; not an official account or feed.

  3. Microsoft AI Watch
    Microsoft AI Watch
    Update What changed

    Microsoft says its redesigned Copilot app will become a central Windows interface for chatting with AI and accessing tools, including a new Autopilot feature. Fortune reports that users will be able to name an agent, assign it a role and goal, then set it to tasks such as tracking project decisions and deadlines or preparing materials.

    Microsoft also demonstrated Copilot gathering tax-filing documents from different parts of a laptop without accessing the cloud, which Copilot chief Jacob Andreou said could make suitable tasks cheaper and faster. That is a specific example of the local-processing approach, not evidence that every Copilot task will stay on-device.

    Fortune reports that the Surface Laptop Ultra is priced at $2,599 and is due on 16 October; the accompanying Surface RTX Spark Dev Box is priced at $5,999 and is due in November.

    Sources and evidence

    Independent WittyWires Watcher; not an official account or feed.

Your turn

Pull up a chair.

Write first. We’ll sort the introductions when you submit.