What changed in AI, and what to do about it.
A weekly briefing on AI developments and what they mean for your business. Browse past issues below.
Karpathy joins Anthropic
Anthropic just made recursive AI research an explicit org structure, with Andrej Karpathy running it.
Gemini moved into every surface Google owns
Google made agents a default surface across Search, Android, Workspace, and shopping, and METR put real numbers on what those agents still can't verify on their
The strongest agent finished under 4% of real work last week
Six vendors said agents are the primary user this week. A new benchmark says agents finish under 4% of real workflows. Both are true.
Mintlify just killed seat pricing in one sentence
Three same-week moves admit that the pricing model SaaS was built on no longer fits when agents are the users.
Mercedes cut 8 months of work to 8 days
Coding agents stopped being a developer experiment yesterday. Plus Codex on Windows, a Claude Code console, and a quiet supply-chain attack on npm.
Microsoft has made $30B from OpenAI
The Microsoft-OpenAI revenue split just shifted. Plus Google's first AI-generated zero-day and Mira Murati's debut.
30 days of AI training is free (limited time)
Exclusive for newsletter subscribers: A 30-day personalized AI training program, free for readers who complete the extended assessment. AI Ready Team is open too.
Only 19% of AI users work in a company built for it
Microsoft surveyed 20,000 workers, and the gap they found is not the tool, it is the org chart around it.
Anthropic just leased Elon's Memphis supercluster
Anthropic gets Colossus 1, OpenAI's depth gap widens to 3.5x, and Wendy's keeps scaling while Taco Bell pauses.
The 3.5x intelligence gap inside enterprises is widening
OpenAI's enterprise data shows depth of use, not access, is now the divide. Plus IBM's Think keynote, GPT-5.5 Instant, and CPC ads in ChatGPT.
Anthropic just put Claude inside Excel
Anthropic shipped ten ready-to-use finance agents and put Claude inside Excel and PowerPoint. Here's what changed.
The AI labs just became consulting firms
Anthropic and OpenAI both launched enterprise services firms yesterday, hours apart. The implementation layer for enterprise AI just changed hands.
Big Tech is spending $725B and supply is still the bottleneck
Microsoft, Alphabet, Amazon, and Meta all said capacity, not demand, is the constraint. The procurement implications start now.
Your Word doc is now an agent
Agentic AI inside Word and Excel went from feature flag to default this week, which changes who has it from the early adopters to everyone with a 365 license. Plus four shorter reads on Big Tech earnings, GM and Gemini, the largest seed in European history, and an open model crossing a frontier coding threshold.
OpenAI is now a Bedrock vendor
Microsoft ended OpenAI's cloud exclusivity, and within 48 hours, OpenAI was live on AWS. Plus five shorter reads on Musk, Manus, NVIDIA, Joby, and Google's Pentagon deal.
Your AI stack has a guardrail problem
An AI agent deleted nine seconds of work that took three months to build. Your stack might be running the same access pattern.
The week AI stopped being a product
GPT-5.5 set a new bar, the coding wars got ugly, and Sullivan & Cromwell showed operators exactly how AI velocity breaks.
Your AI vendor supply chain just became your attack surface
Today's signal: the AI tools you connected to your systems over the last 18 months are now a supply-chain attack surface, and three breaches in one week are the evidence. Plus, what to do about it, OpenAI going after enterprise, and five other things that moved.
Anthropic just tied itself to AWS for a decade
Today's operating signal: the compute layer is becoming a lock-in tier of its own. Plus four shorter reads on Apple, Chrome, NSA, and a new Anthropic Labs product.
Switching AI foundations is about to get expensive
The case for one vendor, and when to break your own rule.
Why the Same Model Performs Differently
Why the same model can vary by 16 points, and how harness design determines agent performance in production
Anthropic Unveiled Its Most Powerful Model and Decided Not to Release It
What Claude Mythos found in the world's most critical software and why it changed everything
Anthropic just passed OpenAI in revenue. Here's what's behind it.
$30 billion in run-rate revenue, 1,000 enterprise customers spending $1M+ each, and a 3.5 gigawatt compute deal.
What the Claude Code Leak Actually Tells Us
500,000 lines of source code, 44 hidden feature flags, and a clearer picture of where AI agents are heading
What Stripe Actually Built
1,300 pull requests a week, zero human-written code. Here's what the deployment looked like, and what it means for the rest of us.