The courtroom receipts just revealed something more interesting than the lawsuit itself: what it actually looks like to work at a company building the agent economy in real time.
The Summary
- Apple sued OpenAI for trade secrets theft, naming two former Apple engineers now at OpenAI. The discovery phase is already leaking juice.
- Text messages from ex-Apple engineer Chang Liu reveal OpenAI employees "use AI all over the place" internally — a stark contrast to Apple's traditional engineering culture
- The real story isn't IP theft. It's the cultural gap between a 50-year-old hardware company and a company literally rewiring how work gets done.
The Signal
Apple's lawsuit against OpenAI accuses former employees Chang Liu and Tang Tan of stealing confidential information when they jumped ship. Standard Valley drama. But OpenAI's defense strategy is what matters: they published the actual text messages from Liu to his former Apple colleagues. And those messages tell a story Apple probably didn't want told.
"This place is crazy… hahahah we use Ai all over the place," Liu texted shortly after joining OpenAI. For context, this is a senior system electrical engineer with eight years at Apple describing his new workplace like he just discovered electricity. The implication is clear: at Apple, even the people building AI products weren't using AI internally. At OpenAI, it's ambient.
"The gap isn't about AI capability. It's about AI adoption culture inside the companies supposedly leading the AI race."
This is the real divide in tech right now, and it has nothing to do with model parameters or training compute. It's about whether your company uses its own tools to do actual work. Apple builds AI features for iPhones. OpenAI builds AI and then uses that AI to build more AI. One is product development. The other is recursive self-improvement.
Liu's other text is equally telling: "This startup style company is a funny and will be very hard in a couple months😁😁." Translation: I went from a structured engineering org to a place where agents are building the scaffolding while we're still standing on it. That's not chaos. That's what it looks like to work at Web4 velocity.
Key differences emerging:
- Apple: AI as product feature. Ship it to users. Internal tooling stays traditional.
- OpenAI: AI as work infrastructure. If we're not using it ourselves, why would anyone else?
- The second approach compounds. The first stagnates.
The lawsuit itself will probably settle or drag on for years. Trade secret cases usually do. But the discovery process just gave us something more valuable than a legal outcome: a window into how companies actually operate when AI stops being a product line and becomes the operating system for work itself.
Apple's case rests on the idea that these engineers carried proprietary knowledge about Apple's AI development in their heads when they left. Maybe they did. But OpenAI's counterargument is essentially: we're building something so different from what Apple does that their knowledge is barely transferable. And if Liu's texts are real, that's probably true.
The Implication
If you're trying to figure out where to work or what to build, this lawsuit just handed you a rubric. Ask one question: does this company use AI internally to do its actual work, or does it just build AI products for other people. The first kind of company is learning faster than the second kind can move.
The winner of this case matters less than the cultural gap it exposed. Apple thinks like a product company. OpenAI thinks like an AI-native organism. One of those models has a ceiling. The other doesn't.