Microsoft hires Ali Farhadi as Nvidia launches Nemotron 3

Key Takeaways

• Alibaba's Damo Academy launched the XuanTie C950, a 5-nm RISC-V CPU operating at 3.2 GHz, touted as the world's highest-performing RISC-V CPU for AI, tripling the speed of its predecessor.
• Microsoft hired top AI researchers, including Ali Farhadi, Hanna Hajishirzi, and Ranjay Krishna, to reduce reliance on OpenAI and develop internal advanced foundation models.
• NVIDIA introduced Nemotron 3 models (Ultra, Nano Omni, Super) for advanced AI agents, focusing on reasoning, multimodal understanding, and safety, alongside Nemotron 3 Content Safety and VoiceChat.
• The IAB Tech Lab launched Agentic Advertising Management Protocols (AAMP) to standardize agentic AI in advertising, with the AAMP Buyer Agent SDK integrated into Kochava's StationOne.
• Oscar-winning director Daniel Roher criticized the AI economy as a 'Ponzi scheme,' raising concerns about financial speculation and unchecked technological advancement.
• WebinarTV is reportedly recording Zoom meetings and converting them into AI-generated podcasts for profit without explicit permission, raising privacy concerns.
• AI is being used to detect subtle ground movements from satellite radar images, aiding in early warnings for landslides and avalanches.
• Relyance AI launched Lyo, a new tool to secure autonomous AI systems by monitoring data activity across code, cloud infrastructure, and AI agents.
• Alibaba aims to become a full-stack AI technology provider, using the XuanTie C950 to expand its hardware capabilities for agentic AI and inference computing.
• The League of Women Voters Copper Country highlighted gender bias, comparing it to AI training, emphasizing the need to address biases in both society and AI systems.

Alibaba unveils XuanTie C950 chip for AI demand

Alibaba's research arm, Damo Academy, has introduced the XuanTie C950, a new central processor unit based on the RISC-V architecture. This CPU is designed for cloud computing and allows customers to customize it for their specific AI inferencing needs. Alibaba aims to become a full-stack AI technology provider, including hardware, and this chip supports that goal. The XuanTie C950 is optimized for cloud environments and offers flexibility for various AI workloads.

Alibaba launches new 5-nm AI chip amid OpenClaw popularity

Alibaba Group has revealed its latest central processing unit designed for artificial intelligence applications. This new chip is built using the RISC-V open-source architecture. The release comes as demand for AI agents, particularly driven by the popularity of OpenClaw, continues to grow in China. Sources suggest that TSMC, a long-time partner, will manufacture the chip.

Alibaba's new RISC-V AI chip advances its strategy

Alibaba Group is expanding its AI hardware capabilities with the new XuanTie C950 chip, designed for agentic AI and inference computing. Developed by Damo Academy, this RISC-V based CPU is optimized for cloud environments and can be customized for specific AI tasks. This move supports Alibaba's vision to be a complete AI technology provider, covering both hardware and software. The company's chip subsidiary, T-Head, is also advancing its position in the competitive AI chip market.

Alibaba's new RISC-V chip boosts AI agent performance

Alibaba has developed a new server chip, the XuanTie C950, built on the open-source RISC-V architecture. This chip operates at 3.2 GHz and is described as the world's highest performing RISC-V CPU for AI. It is reportedly over three times faster than its predecessor, the XuanTie C920. The chip is designed for high-performance cloud systems and agentic AI, offering customization for specific AI workloads with low licensing fees. Alibaba's T-Head semiconductor arm is driving this in-house chip development.

AI offers benefits but not for film companies

New technology advancements in artificial intelligence may not benefit film studios as much as expected. While companies invest in new technologies, the advantages gained could be diminished by competition. This suggests that the film industry might not see the significant gains from AI investments that other sectors could experience.

Oscar winner calls AI economy a 'Ponzi scheme'

Oscar-winning director Daniel Roher has criticized the current artificial intelligence economy, calling it a 'Ponzi scheme' in his new documentary, 'The Age of AI.' Roher suggests that investors are pouring money into AI companies based on future promises without fully understanding the technology or its profitability. The film explores the hype surrounding AI and warns about the potential dangers of unchecked technological advancement and financial speculation, particularly highlighting the roles of Silicon Valley and Wall Street.

Microsoft hires top AI researchers for new team

Microsoft has hired several prominent AI researchers, including former Allen Institute for AI (Ai2) CEO Ali Farhadi, Hanna Hajishirzi, and Ranjay Krishna. They will join Mustafa Suleyman's AI organization within Microsoft. These hires aim to reduce Microsoft's reliance on OpenAI for advanced AI models. The researchers will also maintain their faculty positions at the University of Washington. This move strengthens Microsoft's efforts in developing advanced foundation models.

IAB Tech Lab standardizes agentic AI for advertising

The IAB Tech Lab is working to make agentic AI more practical for the advertising industry by launching its Agentic Advertising Management Protocols (AAMP). This initiative aims to create standardized AI-driven processes for media buying, measurement, and compliance. AAMP Buyer Agent SDK has been integrated into Kochava's StationOne platform as an open-source workspace. This collaboration allows industry members to test and implement AAMP-based workflows, promoting interoperability and addressing the fragmentation in the emerging agentic advertising market.

NVIDIA Nemotron 3 models power advanced AI agents

NVIDIA has introduced a new generation of Nemotron 3 models designed to create advanced AI agents capable of reasoning, multimodal understanding, and safe operation. The Nemotron 3 Ultra offers high reasoning accuracy, while Nemotron 3 Nano Omni focuses on enterprise-grade multimodal understanding. These models, along with Nemotron 3 Super, are built for multi-agent systems with large context windows and enhanced efficiency. NVIDIA also released Nemotron 3 Content Safety for multimodal safety guardrails and Nemotron 3 VoiceChat for real-time conversational AI.

WebinarTV records Zoom meetings for AI podcasts

WebinarTV is reportedly scanning the internet for Zoom meeting links, recording these calls, and converting them into AI-generated podcasts for profit. The company claims to host over 200,000 webinars and may be recording calls without explicit permission. Some individuals have discovered their private Zoom calls were recorded and turned into podcasts only after WebinarTV contacted them to promote their services. This practice raises concerns about privacy and security.

Keweenaw group highlights Equal Pay Day bias

A Keweenaw-based organization is raising awareness for Equal Pay Day, explaining that women often have to work longer into the year to earn what men made the previous year. The League of Women Voters Copper Country compares gender bias to AI training, noting that just as AI can be influenced by biased data, people can unconsciously adopt discriminatory beliefs. They emphasize the need for deliberate action to address these biases in both society and AI systems. The group set up a booth at Michigan Tech University to inform students about these issues.

Penn State AI Hub seeks award nominations

The Penn State AI Hub is accepting nominations for its annual Artificial Intelligence Awards. These awards recognize faculty, staff, students, and alumni for significant contributions to AI research and applications made after May 31, 2024. Categories include AI Societal Impact, Best Undergraduate Thesis, Best Master's Thesis, and Best Doctoral Dissertation related to AI. The deadline for nominations is April 17, 2026. A panel of Penn State AI faculty will select the winners.

AI helps detect Earth's ground movement

Artificial intelligence is being used to detect subtle ground movements that could predict landslides and avalanches. Researchers are using AI to analyze satellite radar images, identifying unstable areas like the village of Kimtang in Nepal. This technology helps analyze vast amounts of data that would be impossible for humans to process manually. By understanding the physics of slope failure, AI can provide early warnings, allowing for potential evacuations and better risk management.

Relyance AI's Lyo secures autonomous systems

Relyance AI has launched Lyo, a new tool designed to address data security gaps in autonomous AI systems. Lyo continuously monitors and adds context to data activity across various platforms, including code, cloud infrastructure, and AI agents. This provides security teams with real-time visibility into how sensitive data is being used, a critical need as AI agents gain more access. Gartner predicts a significant rise in AI-related data breaches due to improper GenAI use.

Sources

• Alibaba Unveils New Chip Design to Meet Surging Demand for AI
• Alibaba unveils 5-nm AI chipset design amid OpenClaw frenzy
• Alibaba Unveils New RISC-V AI Chip, Advances Full-Stack Strategy
• Alibaba develops next-gen chip for agentic AI, Chinese media says
• AI could be a blockbuster — just not for film companies
• Oscar Winner Calls AI Economy a “Ponzi Scheme” in Explosive New Documentary
• Microsoft hires former Ai2 CEO Ali Farhadi and key researchers for Suleyman’s AI team
• Ad Tech Briefing: IAB Tech Lab accelerates push to make agentic AI more practical
• Building NVIDIA Nemotron 3 Agents for Reasoning, Multimodal RAG, Voice, and Safety
• This Company Is Secretly Turning Your Zoom Meetings into AI Podcasts
• Keweenaw organization holds Equal Pay Day event at MTU, likens bias to AI training
• Penn State AI Hub seeking nominations for annual Artificial Intelligence Awards
• Ground truth: When the Earth moves, AI can spot it
• Relyance AI’s Lyo addresses data security gaps in autonomous systems