Tech & AI Newsfeed: June 3, 2026 Digest

Google DeepMind Launches Gemma 4 12B: A Unified Open-Weights AI Model

Google DeepMind has officially announced its latest machine learning breakthrough with the release of Gemma 4 12B. This highly anticipated 12-billion parameter open-weights AI model leverages a unified neural network architecture designed to seamlessly process both text and image data, setting a new standard for multimodal open-source AI development.

Autonomous Vehicle Safety: New Causality Algorithms Explain Self-Driving Car Crashes

In a major leap forward for autonomous vehicle safety, researchers at King’s College London have developed innovative causality algorithms. These advanced algorithms can automatically analyze and explain the root causes of self-driving car crashes using actual causality, providing critical insights for improving future autonomous driving systems and AI policy.

Sources:

NVIDIA Expands Global AI Cloud Infrastructure with New Exemplar Partners

To meet surging global AI compute demand, NVIDIA is rapidly expanding its AI cloud infrastructure ecosystem. Notably, YTL Artificial Intelligence Cloud has recently achieved the prestigious NVIDIA Exemplar Cloud status, highlighting the growing scale and capabilities of enterprise AI cloud solutions worldwide.

Sources:

Quantum Computing Breakthroughs: Microsoft Majorana 2 Chip and IBM’s $10B Investment

The quantum computing race is accelerating with two major industry announcements. Microsoft has officially unveiled the Majorana 2, a next-generation topological quantum chip designed for enhanced reliability using agentic AI. Simultaneously, IBM has announced a massive $10 billion strategic investment to scale its enterprise quantum computing initiatives and hardware development.

Sources:

Cybersecurity Alert: Critical VS Code Vulnerability Exposes GitHub OAuth Tokens

Software engineers and developers face a severe cybersecurity threat due to a newly discovered one-click vulnerability in Visual Studio Code (VS Code). This critical security flaw allows malicious actors to steal GitHub OAuth tokens, granting them full read and write access to compromise private and public user repositories.

Sources: