Everybody wants to rule the AI world

What It Covers

The Vergecast covers week two of the Elon Musk vs. OpenAI trial, revealing internal texts and journal entries from Sam Altman, Greg Brockman, and Mira Mirati during the 2023 CEO firing. Hosts also analyze OpenAI's rumored ChatGPT phone, Apple's $250M Siri lawsuit settlement, Google's Fitbit Air rebrand, and Nintendo's Star Fox remake.

Key Questions Answered

• •AI Discovery Risk: Every executive using AI tools daily — Claude, ChatGPT, Copilot — is creating discoverable records. Courts have already ruled that sharing attorney-client privileged legal advice with an AI chatbot destroys that privilege, treating it as disclosure to a third party. Lawyers are now actively warning clients about this. Executives wearing recording bracelets or dumping context into AI systems are generating the next generation of lawsuit evidence without realizing it.
• •OpenAI Governance Failure: The OpenAI board fired Sam Altman without ever publicly stating a clear reason, communicating the decision through Mira Mirati as an intermediary to both sides simultaneously. Internal texts show Altman repeatedly asking basic questions — "am I still fired?" and "more time for what?" — revealing a governance structure so dysfunctional that the CEO learned his own termination status through a series of lowercase text messages from his CTO.
• •Platform Control Forces Hardware: OpenAI's rumored phone strategy is structurally necessary, not aspirational. Apple and Google will never grant third-party AI systems the hardware-level access — persistent camera recording, background processing, app control — required to deliver the always-on contextual AI experience OpenAI needs. Building proprietary hardware is the only path around App Store restrictions, even though competing with the iPhone has a near-zero historical success rate.
• •Musk's OpenAI Exit Pattern: Trial evidence shows Elon Musk demanded unilateral control of OpenAI, attempted to fold it into Tesla, and offered Sam Altman a Tesla board seat in exchange. When the other founders collectively refused, Musk left and immediately threatened to start a competitor and poach key researchers including Andrej Karpathy and Ilya Sutskever. He then hired Karpathy to Tesla while still nominally involved with OpenAI, a direct conflict he disclosed as a confession.
• •Apple's AI Liability: Apple settled a $250M class action lawsuit over Apple Intelligence advertisements that depicted Siri features — including cross-app memory and contextual meeting recall — that did not exist at launch and still do not fully exist. Eligible iPhone buyers from June 2024 through May 2025 can claim approximately $25. The settlement confirms Apple shipped a marketing narrative around AI capabilities years ahead of actual product delivery, a strategy that generated legal exposure while buying time.
• •Google Health Rebrand Backfire: Google rebranded the Fitbit app to Google Health to position its AI health coaching across all devices, including Apple Watch. The strategy creates an immediate trust problem: the Google brand carries data-collection associations that Fitbit did not, making users more likely to assume health metrics feed into ad targeting. Fitbit's community goodwill built over a decade gets replaced with Google's existing baggage around surveillance and data monetization.
• •Satya Nadella's Paper Trail Discipline: Throughout the OpenAI crisis, every party — Altman, Mirati, board members — attempted to loop Microsoft CEO Satya Nadella into written communications and failed. Nadella communicated exclusively by phone, leaving no discoverable text or email record. When he did act, he moved decisively by offering to hire the entire OpenAI team at Microsoft, using leverage without documentation — a model of crisis communication that stands in direct contrast to every other executive in the case.