She Quit for Health... Then Rejoined OpenAI in a Day
Grab the popcorn, because the AI industry just served up a plot twist that reality TV producers would reject for being too on the nose. Lilian Weng, co-founder of Thinking Machines Lab, quit her startup citing burnout and health issues. Then, within roughly 24 hours, OpenAI confirmed she's coming back. Yes, really. We are not making this up.
Let's rewind the tape, because this saga deserves a slow-motion replay. Weng, the former VP of AI Safety Research at OpenAI and one of the most recognizable faces in the AI world, announced her departure from Thinking Machines in a painfully candid note shared on X. No corporate fluff, no "pursuing new opportunities" filler. She straight-up said the startup pace was wrecking her.
"I don't feel I'm able to continue at the pace a startup requires," she wrote, explaining that months of consistent stress and workload had pushed her "beyond what my health can sustain physically." She even floated taking a narrower role, then decided that wasn't who she is. "If I cannot give it 100% for what I should be responsible for, I fail my job."
Inspirational stuff. Vulnerable, honest, deeply human. The AI community flooded the replies with hearts, support, and "take care of yourself" energy. Soumith Chintala, Miles Brundage, and a whole roster of industry heavyweights showed up to cheer her on.
And then... she joined OpenAI. Within a day. The universe has a sense of humor, and it is apparently a chaos goblin.
Here's where it gets spicy. OpenAI told TechCrunch that Weng will lead a top-level team focused on accelerating internal research, with a special emphasis on recursive self-improvement. That's the technical term for AI systems that iterate on themselves to become more powerful. So she quit a startup to protect her health... and her new job is literally making AI evolve itself at an even more frantic pace. Peak "I'm leaving to spend more time with my family" energy, and we are living for it.
Meanwhile, Thinking Machines co-founder and ex-OpenAI CTO Mira Murati responded to Weng's post with all the grace of someone who has definitely seen this movie before: "We'll miss you, it's been wonderful building Thinky together. I'm glad that you're putting your health first. Thank you for everything."
Plot twist within the plot twist: Weng is the FOURTH co-founder to exit Thinking Machines since Murati launched the lab in early 2025 with a team of six. Andrew Tulloch bounced to Meta back in October. Then Barret Zoph and Luke Metz both returned to OpenAI in January in a coordinated announcement that had everyone checking their calendars. And now Weng makes it three co-founders who went straight back to OpenAI specifically.
At this point, OpenAI isn't just competing with Thinking Machines. It's running a buy-one-get-one-free loyalty program on its own alumni, and honestly? Kind of iconic.
The timing is brutal, too. Thinking Machines just released Inkling, a 975-billion-parameter model with a million-token context window that rocketed to the top of rankings for US-built open models. A career-defining week... immediately buried under another co-founder exit story. You can't write this stuff.
For the rest of us, this is a masterclass in reading between the lines. "I need a calmer place" is now officially AI-speak for "I have accepted a job leading AI self-improvement research at the most intense lab on Earth." File that one under phrases that mean the opposite of what they say.
Weng knows exactly what she's signing up for this time, though. She spent nearly seven years at OpenAI before leaving in late 2024, so she's not walking in blind. Maybe that's the point: at a startup you're a co-founder carrying the whole world on your shoulders, while at OpenAI you're a top-tier researcher with an actual support system.
Either way, keep the snacks ready. The OpenAI-Thinking Machines saga now has more episodes than a K-drama, and casting isn't done yet. The future worth building is human -- but the present worth watching? Absolutely unhinged.
Comments