Story cards stay below the sticky dock while audio, chapters, date, and brief navigation remain accessible.
For the first time ever, more U.S. businesses are paying for Anthropic's Claude than OpenAI's ChatGPT, according to the May 2026 Ramp AI Index. Anthropic's adoption rose 3.8% in April to 34.4% of businesses, while OpenAI fell 2.9% to 32.3%. Over the past year Anthropic has quadrupled business adoption, driven largely by Claude Code, its fastest-growing product ever.
Richard Socher's new startup Recursive Superintelligence launched on May 13 with $650 million in funding at a $4.65 billion valuation. The company, backed by GV, Greycroft, Nvidia, and AMD, aims to create AI models that can autonomously identify their own weaknesses and redesign themselves — a long-held holy grail of AI research. The team includes Peter Norvig and Cresta co-founder Tim Shi.
Thinking Machines Lab, founded by former OpenAI CTO Mira Murati, introduced "interaction models" — a natively multimodal architecture that processes input and generates responses simultaneously, like a phone call. Their TML-Interaction-Small model responds in 0.40 seconds and scores 77.8 on the FD-bench interaction benchmark, crushing Gemini at 54.3 and GPT-realtime at 47.8.
Anthropic is in talks to raise $30–50 billion in a round that would value the company at up to $950 billion, surpassing OpenAI's $825 billion valuation for the first time. The round follows Google's $40 billion and Amazon's $25 billion investment pledges and could be Anthropic's final private raise before a potential IPO as early as October 2026.
New Microsoft research reveals that even top-tier frontier models — Gemini 3.1 Pro, Claude 4.6 Opus, and GPT 5.4 — corrupt an average of 25% of document content in multi-step autonomous workflows. The scariest finding: weaker models delete content when they fail, but frontier models subtly rewrite it, making errors nearly impossible for human reviewers to catch.
Uber's CTO revealed the company spent its entire 2026 AI budget by April, driven by explosive adoption of Claude Code and Cursor. Claude Code adoption among Uber's 5,000 engineers jumped from 32% to 84%, with monthly per-engineer API costs running $500–$2,000. About 70% of committed code now comes from AI, and 11% of live backend updates ship with zero human involvement.
Anthropic debuted Claude for Small Business on May 13, a packaged set of connectors and 15 ready-to-run agentic workflows integrating with QuickBooks, PayPal, HubSpot, Canva, Docusign, and Google/Microsoft productivity suites. Tasks include payroll planning, invoice chasing, month-end close, and campaign creation. Anthropic is also launching a 10-city live training tour for small business owners.
Google announced Googlebook on May 12, a new category of premium AI laptops replacing Chromebooks, built from the ground up around Gemini intelligence. The standout feature is Magic Pointer — wiggle your cursor over any content to summon contextual Gemini actions. The OS merges ChromeOS and Android with full phone integration. Hardware from HP, Dell, Lenovo, Acer, and Asus arrives this fall.
Z.ai's GLM-5.1, MiniMax M2.7, Moonshot's Kimi K2.6, and DeepSeek V4 all landed at roughly the same capability ceiling on agentic engineering benchmarks — at a fraction of Western frontier costs. Kimi K2.6 matches Claude Opus on most benchmarks at roughly one-third the inference price. DeepSeek V4 Pro is the largest at 1.6 trillion parameters with 49 billion active.
The Center for AI Standards and Innovation announced agreements with Google DeepMind, Microsoft, and xAI to evaluate frontier AI models before public release, including testing models with reduced safeguards in classified environments. The deals, renegotiated under the Trump administration's AI Action Plan, build on earlier voluntary agreements with OpenAI and Anthropic.
