NVIDIA CEO Jensen Huang projects $1 trillion in revenue by 2027 from its Blackwell and Vera Rubin platforms, a 200% increase from previous forecasts. This confidence is rooted in the belief that demand for agentic AI will rise exponentially.

Huang detailed a new token-driven AI economy where the importance of inference surpasses that of training due to the proliferation of agentic AI models. He highlighted Anthropic's Claude Code AI agent as a revolution in software engineering, noting that all NVIDIA engineers are now assisted by AI agents. Huang referred to agentic AI as "the new computer" and predicted that every SaaS company would become an "AgaaS" (agentic as a service) company.

NVIDIA's strategic moves include a significant push into CPUs, the shipment of its first Groq chips (from an acquisition specializing in inference), and a collaboration with OpenClaw, an open-source AI agent software. Huang likened OpenClaw to Windows for personal computers, suggesting companies will need an "OpenClaw strategy." However, security concerns arise from OpenClaw's requirement for full computer access, leading NVIDIA to introduce NemoClaw for more secure enterprise use.

Further announcements included the development of a Vera Rubin computer for space-based AI data centers and partnerships with Hyundai, Nissan, BYD, and Geely to build 18 million robotaxis annually. Despite these ambitious projections, investor skepticism regarding AI investments is growing, with NVIDIA's shares experiencing a decline post-announcement.

👉 Read the full article on Gizmodo

This study investigates the skills that best predict success in LLM-driven programming, known as "vibe coding." Vibe coding is a technique that allows users to specify programs in natural language and iterate from observed behavior without directly editing source code.

A preregistered cross-sectional study involving 100 tertiary-level students measured computer-science achievement, domain-general cognitive skills, written-communication proficiency, and a vibe-coding assessment. The findings indicate that both writing skill and CS achievement are significant predictors of vibe-coding performance.

Notably, CS achievement remained a significant predictor even after controlling for domain-general cognitive skills. These results can inform tool and curriculum design, helping to determine when to emphasize prompt-writing versus CS fundamentals to support future software creators.

👉 Read the full article on arXiv

Senator Elizabeth Warren is pressing the Pentagon over its decision to grant xAI, Elon Musk's AI company, access to classified networks. This move highlights significant concerns regarding national security, data privacy, and the involvement of a private AI entity with access to sensitive information.

Senator Warren's inquiry likely focuses on the potential risks associated with a private company like xAI having access to potentially delicate defense-related data. The article details the reasons behind Warren's scrutiny and the broader implications of the Pentagon's decision.

👉 Read the full article on TechCrunch