Google Cloud debuts G4 VMs optimized for large-scale AI inference

Google Cloud's new G4 virtual machines, powered by eight NVIDIA RTX PRO 6000 Blackwell GPUs and AMD Turin CPUs, offer quadrupled computing and memory capacity over G2 VMs, targeting AI, simulation, and graphics-heavy enterprise workloads.

🗞 #ai

Dia browser debuts in public beta for Arc members with AI-powered tools

Dia, the new browser from the Browser Company, builds on Arc with AI-driven browsing, a chat-based interface, and personalized tools for tasks like coding or writing. It emphasizes context-aware support and a polished visual design.

🗞 #dia

OpenAI prepares model selection feature for custom GPTs and Codex upgrade

OpenAI is introducing model recommendations for custom GPTs and multiple code output options in Codex, offering developers improved control over performance, speed, and cost without needing to manually adjust or rerun workflows.

🗞 #chatgpt

Leak reveals Gemini 2.5 Flash Lite and Pro GA with June release targets

Google is preparing to launch new Gemini 2.5 models, with the Flash Lite Preview expected around June 17 and the Pro version near June 19, targeting both lightweight inference tasks and advanced enterprise applications.

🗞 #aistudio

xAI tests Voice Mode and Tasks on Grok web with signs of Grok 3.5 integration

xAI is testing Voice Mode and Tasks for Grok on the web, featuring spoken input, audio replies, and scheduled deep searches. Limits on usage hint at upcoming tier changes, while Grok 3.5 integration appears key to both additions.

🗞 #grok

Perplexity adds scheduled Tasks feature for Pro and Enterprise users

Perplexity has introduced a Tasks feature for Pro and Enterprise users, allowing scheduled prompts to automate searches or research. While browser access is limited, the tool aligns with emerging trends in AI-driven workflow automation.

🗞 #perplexity

Experimental Copilot Actions lets testers control a virtual browser

Microsoft's Copilot Actions feature launches in early testing, enabling cloud-based browser control with a virtual desktop and manual input. Though still slow, its browser-based model suggests potential for broader platform support.

🗞 #microsoftcopilot

Microsoft develops Copilot 3D to turn images into 3D objects

Copilot 3D is an early-stage Microsoft project that aims to convert flat images into 3D models, featuring a viewer and galleries for object manipulation. Though core tools are inactive, it points to broader efforts in AI-driven visual creation.

🗞 #microsoftcopilot

Gemini Code Assist gets latest Gemini 2.5 Pro with context management and rules

Google upgraded Gemini Code Assist with the Gemini 2.5 Pro model, enabling 1 million-token contexts, deeper code insight, and advanced review tools. New IDE features support reusable prompts and project rules, with higher task completion rates reported.

🗞 #gemini

Experimental Gemini 2.5 Pro Deep Think update nears release

Google's upcoming Deep Think mode for Gemini 2.5 Pro boosts computing for complex tasks like advanced math, coding, and multimodal queries, offering around 15% higher performance, with likely initial access limited to Ultra-tier users.

🗞 #gemini

Google Gemini 2.5 out of preview across AI Studio, Vertex AI, and Gemini app

Google released stable versions of Gemini 2.5 Pro and Flash, launched a Flash-Lite preview, and expanded access across major platforms. The update introduces cost savings, native audio, and advanced math and code capabilities for varied workloads.

🗞 #aistudio

ICYMI: OpenAI enables shared prompts for API and Playground

OpenAI now supports prompts as an API primitive, allowing centralized management, version control, and optimization across its tools. This update addresses developer needs for structured prompt configuration and deployment.

🗞 #chatgpt

Adobe puts AI art in your pocket with Firefly mobile release

Adobe released the Firefly mobile app globally, allowing users to generate and refine visuals or videos from text and sync with Creative Cloud. It supports multiple AI models and includes a $10 token-based plan for premium features.

🗞 #ai

Midjourney launches its first video model V1 with 5 to 20-second clips

Midjourney launched its V1 video model on June 18, 2025, allowing users to animate images into short 5–20 second clips. While lacking audio and advanced video fidelity, it offers a low-cost entry amid rising competition.

🗞 #ai

xAI tests screen sharing on Voice Mode for Grok web

xAI is testing screen sharing and advanced X search functions in Grok, expanding its toolset for web use. These early-stage features point to tighter X platform integration ahead of the delayed Grok 3.5 release.

🗞 #grok

Perplexity developing AI try-on tool to preview clothes before buying

Perplexity is developing an AI-powered “Try On” tool for online shopping that generates user images wearing selected items, aiming to support personalized purchase decisions. The feature is in testing with no confirmed release date.

🗞 #perplexity

Jules agent becomes more context-aware with AGENTS.md support

The updated Jules agent brings improved performance, smarter task execution, consistent environment setup, and context awareness through AGENTS.md, aligning with unique repository workflows and supporting automated test writing and execution.

🗞 #jules

Upcoming Grok update will introduce centralized Files tab

xAI is adding a Files tab to Grok’s web interface, centralizing file management with filters and built-in tools for viewing, editing, and executing content, aligning with efforts to position Grok as a full productivity platform.

🗞 #grok

Anthropic developing Memory and AI-powered Artifacts for Claude

Anthropic is expanding Claude's artifacts to support embedded AI capabilities, enabling users to build functional, shareable mini-apps without coding. A new memory feature is also in development to allow Claude to recall past conversations.

🗞 #claude

AI Studio revamp under development with MCP support and Jules tie-in

Google is preparing updates to AI Studio, including a unified build section linked to Jules SWE Agent, support for Imagen 4, a modular tools menu, and daily usage limits, signaling broader capabilities and platform management.

🗞 #aistudio

Google launched Imagen 4 and Imagen 4 Ultra on AI Studio and APIs

Google released Imagen 4 via the Gemini API with limited free access starting June 24, 2025. It generates four 1024×1024 images per call, includes SynthID watermarking, and supports up to 2K resolution through Vertex AI.

🗞 #aistudio