In partnership with
What Shipped This Week
Monday: OpenAI pushed a major Codex update positioning it as a tool for everyone, not just developers, with role-based onboarding for finance, marketing, and data science, a faster UI, and a one-click importer that pulls your entire setup from other tools.
Tuesday: Anthropic launched Claude connectors for Blender, Autodesk Fusion, Adobe Creative Cloud, Ableton, Splice, Canva, SketchUp, and Resolume, letting creative professionals work with Claude inside the tools they already use. Claude Code also added mobile push notifications so developers know when a long task finishes without sitting at their desk.
Wednesday: Gemini added the ability to generate and download files directly from chat, including Google Docs, Sheets, PDFs, Word, and Excel. Prompt and download, no copying or reformatting required.
Thursday: Anthropic launched Claude Security into public beta for Enterprise customers, scanning codebases for vulnerabilities without requiring custom tooling. Hours later, Sam Altman announced GPT-5.5-Cyber is rolling out to critical infrastructure defenders.
Friday: GPT-5.5 became OpenAI's biggest model release ever, with API revenue growing more than twice as fast as any prior launch and Codex doubling revenue in under seven days. OpenAI added virtual pets to Codex and announced a custom dictionary for voice dictation.
What Went Wrong This Week
Nothing catastrophic this week, but a few stories carried implications worth paying attention to.
The Musk vs Altman trial is the obvious one. Musk told a jury that AI could kill humanity, called himself a fool for donating $38 million to what became an $800 billion company, and said accepting equity from Altman after the for-profit restructuring would have felt like a bribe. The legal outcome is genuinely uncertain with prediction markets near 50-50, but the real damage is reputational. Every day of testimony puts OpenAI's founding promises back under a spotlight the company would clearly rather avoid, right as they are preparing for a public stock listing.
The Microsoft deal restructuring is not a crisis, but it is a signal. Microsoft spent $13 billion to lock in exclusive access to OpenAI's technology, and this week that exclusivity quietly ended. Microsoft keeps the equity and the preferred partnership, but OpenAI can now work with Amazon, Google, and anyone else. For Microsoft shareholders, the moat they paid for just got a lot smaller.
The Owl Alpha model that appeared on OpenRouter with no attribution is less a problem and more an uncomfortable reminder. A frontier-capable model showed up from nowhere, with no announcement and no accountability. That kind of thing is becoming routine, and the gap between what capable AI can do and who is actually watching is not closing.
Arnold Schwarzenegger has a newsletter.
Yeah. That Arnold Schwarzenegger.
So do Codie Sanchez, Scott Galloway, Colin & Samir, Shaan Puri, and Jay Shetty. And none of them are doing it for fun. They're doing it because a list you own compounds in ways that social media never will.
beehiiv is where they built it. You can start yours for 30% off your first 3 months with code PLATFORM30. Start building today.
What It All Means
The week's most telling detail is not any single product launch. It is the scoreboard. OpenAI had arguably their best seven days in company history: GPT-5.5 revenue records, Codex doubling, Images 2.0 pulling in millions of new users, a cybersecurity model rolling out to critical infrastructure, and virtual pets shipped on a Friday just because they could. All of this while their CEO was being cross-examined in a federal courthouse about whether he stole a charity.
That tension is worth sitting with. The trial is not just a legal dispute between two billionaires. It is a public reckoning with a question the AI industry has been avoiding: when you raise tens of billions at a nonprofit and restructure into one of the most valuable companies in the world, what exactly did you promise and to whom? The founding charter is in evidence. The text messages are in evidence. The jury is listening.
Meanwhile the competitive landscape keeps shifting faster than any single story can capture. Anthropic quietly became the connective tissue inside creative software used by millions of professionals. A model nobody has claimed appeared on a public platform and performed at frontier level. The pace at which capable AI is proliferating, from known labs, unknown labs, open source releases, and stealth drops, is accelerating faster than any governance framework is moving to meet it.
It’s been an eventful week.