The Omniscient Bulletin · 2026-06-05
The Omniscient Bulletin — June 5, 2026
Safety warnings, national security directives, and a labor market reckoning arrive in the same week
Thursday's AI news cut across every layer of the stack simultaneously: Anthropic disclosed that Claude now authors more than 80% of its own codebase and called for a conditional global pause; lab CEOs jointly warned Congress that the same models could lower the barrier to bioweapons; and Trump signed a memo directing the Pentagon to accelerate AI adoption in warfighting. Underneath it all, tech workers absorbed a year of AI-cited layoffs totaling 123,000 jobs. The technology is reshaping its own development, its own governance, and the workforce around it - all at once.
Frontier
Anthropic: Claude wrote 80% of its own codebase in May, doubling task-completion horizon every four months
A new Anthropic Institute report reveals that as of May 2026, Claude authored more than 80% of code merged into Anthropic's codebase - up from low single digits before Claude Code launched in early 2025. Engineers now ship 8x as much code per day as in 2024. On open-ended tasks, Claude's success rate hit 76% in May, up 50 percentage points in six months. Anthropic warns full recursive self-improvement could arrive sooner than most institutions are prepared for and proposes a pause - but only if multiple frontier labs across multiple countries agree to stop under the same conditions.
Policy
Altman, Amodei, Hassabis, Suleyman and dozens of co-signers urge Congress to mandate bioweapon screening
The CEOs of OpenAI, Anthropic, Google DeepMind, and Microsoft AI - joined by Meta's Chief AI Officer Alexandr Wang, Paul Graham, Patrick Collison, and scores of life-sciences researchers - signed an open letter calling on Congress to require all synthetic DNA and RNA providers to screen customers and orders for sequences that could enable biological weapons. The letter warns that AI is eroding the knowledge barriers that have historically kept bioweapons out of reach of bad actors. A bipartisan Senate bill tracking the same goal is already in committee.
Policy
Trump signs NSPM-11, directing the Pentagon and intelligence agencies to accelerate AI adoption
President Trump signed National Security Presidential Memorandum-11 on June 5, ordering the military, intelligence community, and related federal agencies to rapidly adopt frontier AI models across warfighting and intelligence domains. The memo directs the Secretary of War to update autonomous weapons policy within 90 days, rescinds the Biden-era NSM-25, and instructs agencies to terminate contracts with vendors that repeatedly violate the memorandum's principles. A classified annex covering sensitive applications is due within the same window.
Industry
UK CMA issues world-first binding order: Google must give publishers an opt-out from AI scraping
The UK Competition and Markets Authority issued a binding order requiring Google to give publishers effective tools to opt out of having their content used in AI Overviews, AI Mode, and AI model training - the first such regulatory mandate anywhere in the world. Google has nine months to comply, though it has already begun testing opt-out controls in the UK. The CMA said the move is designed to restore bargaining power for news organizations that have seen traffic collapse since AI summaries displaced organic search results.
Labor
US tech sector cut 123,000 jobs in 2026 - its heaviest layoff year since 2023 - with AI the leading cited cause
Outplacement firm Challenger, Gray & Christmas reported that US tech companies announced 38,242 job cuts in May alone, the sector's worst single month since August 2024 and more than any other industry. That pushed the 2026 running total to 123,653, up 66% on the same period in 2025. AI was the most-cited reason for cuts across all sectors for a third consecutive month - even as the same companies plan a combined $725 billion in AI capital spending this year. Unemployment claims have not risen in lockstep, leaving the displacement case contested.