Nobody Aces the AI Safety Exam
The Future of Life Institute released its 2026 AI Safety Index this week, and the results should give anyone building on frontier AI models pause. The highest grade any major lab received was a C+, awarded to Anthropic. OpenAI and Google DeepMind each earned a C, Meta landed at D+, and xAI, DeepSeek, and Mistral effectively failed the assessment.
The index evaluates companies across risk management practices, transparency, and — crucially — whether they actually follow through on their own stated safety commitments. The report’s authors put it bluntly: “Even the most safety-focused frontier lab is doing a mediocre job by its own stated standards.”
The timing is notable. As models grow more capable and agentic — with OpenAI’s GPT-5.6, Anthropic’s Claude Sonnet 5, and Google’s upcoming Gemini 3.5 Pro all pushing the boundaries of autonomous tool use — the gap between capability and safety governance keeps widening. These results add fuel to growing calls from regulators on both sides of the Atlantic for more rigorous oversight.
Andrej Karpathy Joins Anthropic in Major Talent Coup
In one of the most significant talent moves of the year, Andrej Karpathy — the influential AI researcher, former Tesla AI director, and OpenAI founding member — has joined Anthropic. He arrives alongside Tom Blomfield, co-founder and former CEO of fintech unicorn Monzo, who is joining Anthropic’s AI compute team.
The hires extend what has been an aggressive 2026 recruiting run for Anthropic, which earlier this year brought Nobel laureate John Jumper over from Google DeepMind. The company, now valued at $965 billion after its $65 billion Series H round in May, is clearly investing in top-tier talent as it prepares for a potential IPO as early as October 2026.
Karpathy’s move is particularly symbolic. Having co-founded OpenAI and later led Tesla’s Autopilot AI team, his decision to join Anthropic signals a broader shift in where the industry’s best researchers see the most promising work being done.
South Korea Bets $880 Billion on AI Dominance
South Korean President Lee Jae-myung announced a sweeping ten-year, $880 billion national AI strategy — one of the largest government-backed technology investments in history. The plan allocates roughly $518 billion for memory chip factories and approximately $550 billion for AI data centers, with a target of 8.4 gigawatts of data-center power capacity by 2029.
The strategy also includes an ambitious push into humanoid robotics, aiming to grow South Korea’s global market share from 1% to 20% by 2028. The plan positions the country to compete directly with the United States, China, and the EU in the increasingly geopolitical race for AI infrastructure supremacy.
The announcement comes as TSMC continues to post record revenues — June was up 68% year-over-year, with Q2 hitting $39.6 billion — underscoring the real and compounding demand for AI compute hardware worldwide.
Google’s Gemini 3.5 Pro Expected to Launch Tomorrow
All eyes are on July 17, when Google is widely expected to launch Gemini 3.5 Pro. The model reportedly comes after a full architectural rebuild — Google DeepMind scrapped the original base model after engineers found structural failures in recursive tool-calling and SVG generation, pushing the release date back.
Unconfirmed specs suggest a 2-million-token context window, a new Deep Think reasoning layer for improved multi-step problem-solving, and autonomous workflow capabilities. The launch is set to coincide with Shanghai’s World AI Conference, where Chinese President Xi Jinping is expected to attend in person for the first time since the event began in 2018.
However, as of today, no model card, pricing page, or API listing has appeared in Google’s official documentation. Developers planning around the date are, as one commentator noted, “planning around a leak, not a signed launch post.” If the specs hold up, Gemini 3.5 Pro could significantly reshape the competitive landscape just a week after OpenAI’s GPT-5.6 public rollout.
Over 200 Experts Demand Action on AI’s Economic Impact
More than 200 economists and AI researchers, including 16 Nobel laureates, have signed an open letter demanding that policymakers and technology leaders “must act now” to prepare for the economic impact of artificial intelligence.
The letter comes as real-world examples of AI-driven workforce displacement are becoming harder to ignore. Hotel-software company Mews this week cut about 15% of its staff — roughly 170 jobs — and explicitly attributed the reduction to AI efficiency, saying that “individuals can now do work that once needed teams.” It is one of the clearest examples yet of a company openly naming AI as the reason for layoffs.
Meanwhile, US venture funding data tells its own story: 86% of the $412.7 billion raised in H1 2026 went to AI companies, representing an unprecedented concentration of capital that raises questions about both the sustainability of the AI boom and its impact on the broader tech ecosystem.