India's AI talent shortage hits 82%, OpenAI's custom chip is here, and the biggest names in tech pool $500M to help workers keep pace.
Google launches real-time Live Translate across 70 languages, wider than any rival at launch
Google announced a Live Translate feature that handles 70 languages in real time, with no additional app required. The coverage is wider than comparable features from OpenAI or Anthropic at their respective launches, and it works directly in calls and conversations on Android devices.
Why it matters: AI is quietly becoming a universal translator in your pocket. For students, professionals, and communities communicating across language barriers, this is genuinely useful, right now, not in some future rollout.
Raise Us: Anthropic, OpenAI, Amazon, Microsoft, and IBM pool $500M to retrain AI-displaced workers
Former US Commerce Secretary Gina Raimondo and former Indiana Governor Eric Holcomb launched Raise Us, a $500 million public-private initiative backed by Anthropic, the OpenAI Foundation, Amazon, Microsoft, IBM, Bank of America, and Eli Lilly. The fund is specifically focused on helping workers adapt to AI-driven changes in the labour market.
Why it matters: The companies building AI are now co-funding the effort to prepare workers for what that AI displaces. That's a significant shift from "AI creates more jobs than it removes" rhetoric to actual capital going toward workforce transition, worth watching for how it's implemented.
Anthropic accuses Alibaba of the largest-ever AI model distillation attack
Anthropic wrote to the US Senate Banking Committee accusing operators affiliated with Alibaba and its Qwen AI lab of conducting what it called the largest known distillation attack on Anthropic to date, training rival AI systems on Claude's outputs to replicate its capabilities without building them from scratch.
Why it matters: AI intellectual property theft is moving from a technical concern to a geopolitical flashpoint. If it's possible to copy a model's "knowledge" by running millions of queries against it, every frontier lab has a new kind of security problem, and lawmakers are being asked to respond.
82% of Indian employers can't fill AI roles, even as AI-linked hiring jumps 32%
A new survey finds 82% of Indian employers are struggling to hire people with AI skills, with AI Model Development and AI Literacy now ranking as the two hardest skills to find. At the same time, AI-linked hiring is projected to grow 32% in 2026 to nearly 3.8 lakh new roles. The government's SOAR (Skilling for AI Readiness) initiative has so far enrolled 1.34 lakh students and teachers in partnership with Microsoft, HCL, and NASSCOM.
Why it matters: The gap between available roles and available people is India's defining AI challenge right now. If you're building skills in this space (even as a non-engineer) you're positioning yourself in a market with far more demand than supply.
OpenAI's first custom chip "Jalapeño" is real, and it could cut AI costs in half
OpenAI and Broadcom officially unveiled Jalapeño, OpenAI's first internally designed AI accelerator chip. Initial prototype data centre deployment is targeted by end of 2026, with full production scale expected in 2028. OpenAI spent approximately $14 billion running ChatGPT on third-party GPUs in 2025, a 50% inference cost reduction at that scale would fundamentally change its economics.
Why it matters: Right now, OpenAI pays enormous rent to run its own models on someone else's hardware. A proprietary chip is a long-term bet on self-sufficiency. It's also a sign of how mature AI infrastructure investment has become, this is now a hardware race, not just a software one.
EU AI Act content disclosure rules arrive August 2: 36 days away
The EU AI Act's Article 50 takes effect on August 2, 2026. From that date, providers must ensure AI-generated content is watermarked in machine-readable format, and deployers must disclose that content is AI-generated at the first point of interaction. New York's law requiring disclosure of synthetic performers in advertising is already in force as of June 9.
Why it matters: The era of unmarked AI content is ending in Europe, and global companies serving European users will have to comply whether or not they operate locally. For educators, marketers, and content creators, this changes what "using AI" publicly is going to look like.
Four senior Google DeepMind researchers defect to Anthropic in six days
Jonas Adler and Alexander Pritzel are the latest senior researchers planning to leave Google DeepMind for Anthropic, bringing the total to four high-profile departures from a single team in under a week. This follows Anthropic's simultaneous accusation of IP theft by Alibaba and its active recruitment of top technical talent. The model race and the talent race are now running at the same speed, and they're not always going in the same direction.
A24 + Google DeepMind back film AI; Microsoft ships coding and reasoning models; xAI releases Grok 4.20
A24 and Google DeepMind announced a $75M partnership to build AI tools for film production and distribution, the first major Hollywood-AI lab deal of its kind. Microsoft launched MAI-Code-1-Flash (a code-generation model) and MAI-Thinking-1 (a reasoning model built for efficiency), both aimed at reducing developer reliance on OpenAI. xAI released Grok 4.20, including multi-agent variants. More model options, more competition, lower prices, the pattern continues.