Tech News
← Home  ·  All topics

Compute

24 GoKawiil briefs on this topic

Cornelis Networks raises $205M to challenge Nvidia's AI networking dominance

Cornelis Networks, an Intel spinoff, secured $205 million in funding led by IAG Capital Partners. The company unveiled its Active Compute Fabric, a networking technology designed to reduce GPU idle time by letting chips process and transmit data simultaneously, and it has already begun shipping the product.

SmartGym 8 adds AI-generated workouts via its own private cloud system

Fitness app SmartGym has released version 8, adding a feature called SmartGym Intelligence that lets users build workout routines by typing natural-language requests, such as specifying exercise types or restrictions. Because SmartGym isn't yet eligible for Apple's Private Cloud Compute, developer Matt Abras built a similar anonymized cloud infrastructure that routes requests to frontier AI models without logging or linking them to user identities. The update also deepens Siri integration and adds features tied to iOS 27, iPadOS, and watchOS 27.

Apple details usage caps and paid tiers for server-based Apple Intelligence features

Apple published a new support document clarifying that Siri AI, Image Playground, Clean Up, Extend, Spatial Reframing, and certain Shortcuts and developer tools relying on server-side models will face usage limits. The company says limits will vary by feature, complexity, and demand, and that users who hit them will temporarily lose access before it's restored. Apple confirmed increased access can be purchased, though it hasn't disclosed exact thresholds or a rollout timeline.

Apple launches Reference Image feature to verify iPhone 18 Pro photo authenticity

Apple unveiled Apple Reference Image at its Surprise and Shine event, a new iPhone 18 Pro feature designed to confirm whether a photo is unaltered. The camera captures signed sensor data that Apple's Private Cloud Compute converts into an unchangeable reference version stored in the Photos app, which can be compared against edited copies to spot manipulation. Apple is also opening APIs for third-party app integration and adopting the SynthID standard to flag AI-generated or altered images.

Apple's iPhone 18 Pro adds Reference Image tool to verify unedited photos

Apple has unveiled a Reference Image feature on the iPhone 18 Pro that creates a cryptographic signature for photos at the moment of capture, allowing later verification of whether an image has been altered by AI. Using Apple Private Cloud Compute, the system lets users compare an edited photo against its original 'digital negative' and can share these signed authentication records with others. The feature will not roll out in the European Union or China, while Apple's own AI-generated images will instead carry SynthID watermarks.

Apple adds hardware-based 'Reference Image' mode to iPhone 18 Pro to verify authentic photos

Apple has introduced a new camera mode called Reference Image, launching with the iPhone 18 Pro and Pro Max, that cryptographically signs sensor data at capture to prove a photo hasn't been altered. Using Private Cloud Compute, the phone generates an unchangeable reference version of the image that users can compare against edited copies in the Photos app. Apple is also opening a Reference Image API for developers across iOS, iPadOS, and macOS, though the feature won't be available at launch in the EU.

Applied Compute launches AC2, a cloud platform for training open-weight AI models

Applied Compute has introduced Agent Cloud (AC2), a platform designed to give researchers a unified environment for training, running inference on, and continually updating open-weight AI models. The system includes a CLI, SDK, and MCP server, letting developers set it up through coding agents like Claude Code, Cursor, or Codex, or manually via a quickstart guide.

Perplexity launches Hybrid Compute to mix local and cloud AI processing

Perplexity has introduced Hybrid Compute, a feature that divides AI tasks between cloud-based models and models running locally on a user's device. It flags files or data containing personal information and lets users choose to process those locally while sending the rest of the task to the cloud, or send everything to the cloud. The feature currently works only in the Perplexity app on Apple Silicon Macs, supporting local models like Gemma 4 E4B and Qwen 3.6 alongside cloud options such as Claude Opus 5 and GPT 5.6 Sol.

Perplexity adds Hybrid Compute privacy feature to its Mac app

Perplexity has rolled out Hybrid Compute in its Mac app, letting its Computer assistant start tasks in the cloud and shift sensitive steps, like those touching personal files, to a local model running on the Mac itself. An on-device classifier detects personal details such as names and account numbers, replaces them with placeholders before any cloud processing, and restores them once results return. The feature requires Apple silicon, macOS 15 or later, and at least 24GB of unified memory, with 32GB recommended, and uses no cloud credits or API key.

Perplexity launches Hybrid Compute to split AI tasks between cloud and local models

Perplexity has introduced Hybrid Compute, a new feature within its Perplexity Computer platform that divides a single task between a cloud-based frontier model, such as Opus 5 or GPT-5.6 Sol, and a smaller model running locally on a user's Mac. The system automatically flags sensitive files or data and routes them to the local model, while less sensitive parts of the task go to the cloud for stronger reasoning power. Users can review and adjust which files are kept local before the task runs, and choose from local options including Gemma E4B and two Qwen 3.6 variants.

Android 17 QPR2 Beta 4 adds data logging viewer for Private Compute Core

Google's Android 17 QPR2 Beta 4 introduces a new 'Data logging' section inside Privacy Dashboard that lets users monitor data leaving the Private Compute Core, the isolated sandbox behind features like Live Caption and Smart Reply. Users can toggle logging on or off, choose between a limited three-day/100MB mode and a more exhaustive 12-hour/100MB detailed mode, and download or delete the resulting log files.

Nvidia disputes WSJ report of paused deals under its AI Compute Partnership program

The Wall Street Journal reported that Nvidia had frozen some transactions under its two-month-old 'take or pay' AI Compute Partnership after partners objected to Nvidia dictating which customers could lease its GPUs. Nvidia pushed back, saying the program remains active and is simply evolving amid strong demand, without confirming or denying specific paused deals.