🤖 Microsoft Copilot Studio evolution 🤖
Multi-agent orchestration changes the game

Microsoft has unveiled powerful new agent capabilities at Build 2025, headlined by multi-agent orchestration in Copilot Studio. This groundbreaking feature allows organizations to build connected systems where agents delegate tasks to one another across Microsoft 365, Azure AI Agents Service, and Microsoft Fabric. Rather than relying on disconnected agents in silos, these multi-agent systems can collaborate to complete complex, business-critical tasks spanning multiple systems and workflows. For example, a Copilot Studio agent can pull CRM data, hand it to a Microsoft 365 agent to draft a proposal, then trigger another to schedule follow-ups—all working in sync. The feature is currently in private preview with public preview coming soon.

🖥️ Intel Arc Pro GPUs for AI developers 🖥️
Affordable hardware for AI workloads

Intel has launched two new GPUs targeting AI developers: the $299 Arc Pro B50 with 16GB VRAM and the $500 Arc Pro B60 with 24GB VRAM. These cards are specifically designed for memory-intensive AI workloads, not gaming. The B60's 24GB VRAM at just $500 represents a significant price breakthrough compared to NVIDIA alternatives, making it particularly attractive for running large language models at a fraction of the cost. A dual-GPU B60 configuration could offer 48GB of total VRAM for under $1,000. Initial availability will be through OEM workstation vendors in Q3 2025..

🎧 UW's spatial speech translation system 🎧
Preserves voice and spatial location

University of Washington researchers have developed an AI-powered headphone system that can translate multiple speakers simultaneously while preserving their spatial location and unique voice characteristics. The "Spatial Speech Translation" system uses modified noise-canceling headphones with extra microphones to detect surrounding conversations in a 360-degree field. It then separates individual speakers, translates speech in real-time, and plays it back while maintaining both voice qualities and spatial positioning. Currently supporting Spanish, German, and French with a 2-4 second delay, the system runs locally on Apple M2 chips without requiring an internet connection.

INTERESTING NEWS

🎬 Hedra raises $32M for AI character videos 🎬

Hedra has secured $32 million in Series A funding led by Andreessen Horowitz, bringing its total funding to $44 million since 2024. The company's Character 3 model allows users to create lifelike AI character videos from text, audio, and images via Hedra Studio. With 3 million users and over 10 million videos created to date, Hedra's technology powers viral content like AI talking babies. The company plans to triple its 20-person team and increase research efforts to enhance video quality and control capabilities.

🎭 SAG-AFTRA files complaint over AI-recreated Darth Vader voice 🎭

The actors' union SAG-AFTRA has filed an unfair labor practice complaint against Epic Games for using AI to recreate James Earl Jones' voice as Darth Vader in Fortnite without proper union notification or bargaining. This marks a significant escalation in the ongoing tension between creative professionals and companies using AI to replicate performers' likeness or voice without adequate compensation or consent, potentially setting precedents for how AI-generated content is handled in entertainment.

H-FARM EVENTS ON AI

FutureShots25 event this weekend!

Join us May 23-24 at FutureShots25 where our CEO, Diego Pizzocaro, will be speaking! This event brings together innovators who are transforming prompts into voice, designing humanoid prototypes, creating systems used by millions, restoring art with AI, and sending creativity to space. Will we lead the way or just watch? Stop by our stand for "FROM IDEA TO APP WITH AI" where we'll transform your concept into a working prototype in minutes - something you can see, touch, and share. Have an idea that could be your next billion-dollar business? We have the talent and execution to help make it real.