• Uncovering AI
  • Posts
  • 🧠 A Nobel winner just left Berkeley for Beijing

🧠 A Nobel winner just left Berkeley for Beijing

OpenAI's number two steps back, and Meta wants your Instagram face

July 11, 2026

In partnership with

My fellow AI explorers

I want to start with a number: nearly half. That's roughly how many of Omar Yaghi's ~200 trained researchers at UC Berkeley have been Chinese nationals, and it explains a lot about why a Nobel Prize winner just packed up his lab for Beijing. This edition's two lead stories are both about people at the very top choosing to leave, one for China and one for her health. And both say something about where power and pressure sit in this industry right now.

In today’s edition:

Speak naturally. Send without fixing.

Wispr Flow turns your voice into clean, professional text you can send the moment you stop talking. Not rough transcription you have to clean up. Actual polished text — ready for email, Slack, or any app.

Speak the way you think. Go on tangents. Change your mind mid-sentence. Flow strips the filler, fixes the grammar, and gives you text that reads like you spent five minutes writing it.

89% of messages sent with zero edits. Millions of professionals use Flow daily, including teams at OpenAI, Vercel, and Clay. Works on Mac, Windows, and iPhone.

Talent War

A Nobel Prize Winner Just Chose China Over Berkeley

Omar Yaghi, who shared the 2025 Nobel Prize in Chemistry for his work on metal-organic frameworks, has left UC Berkeley to lead a brand new AI for Chemistry and Materials Science Research Center at Tsinghua University in Beijing.

Here's what stands out:

  • Yaghi, 61, was the James and Neeltje Tretter Professor of Chemistry at Berkeley. The announcement landed on July 4, an Independence Day detail nobody in China seems to have missed.

  • His new mandate: use AI to shrink materials discovery and synthesis cycles "by orders of magnitude," aimed at water scarcity, carbon neutrality, and clean energy: the same real-world problems his Nobel-winning MOFs already tackle at lab scale.

  • He's not a stranger to Tsinghua. He's held an honorary role there since 2022, and nearly half of the roughly 200 researchers he's trained were Chinese, a built-in pipeline that made this hire look less like a poach and more like a homecoming.

The context matters here. This isn't just a scientist chasing better lab funding. It's landing right as the Trump administration continues cutting US science spending and limiting international research partnerships, while China runs an aggressive, well-funded campaign to recruit exactly this kind of talent. France has already started offering relocation packages to US scientists. Chinese cities and provinces are handing out lump sums and monthly allowances to researchers willing to move. Yaghi isn't an isolated case either; researchers point to Harvard's Charles Lieber and statistician Liu Jun as part of the same pattern.

What makes this an AI story and not just a chemistry story is the framing Tsinghua itself is using: this is explicitly an AI-driven materials institute, positioned to "tackle complex problems beyond any single field" by bridging "Eastern and Western intellectual traditions."

Translation: China isn't just recruiting chip talent and LLM researchers anymore. It's recruiting the scientists whose work AI is about to supercharge in fields like materials science that touch batteries, semiconductors, and clean energy: the physical infrastructure the whole AI buildout depends on.

🔮 Prediction: This won't be the last high-profile AI-adjacent scientist to make this move. As long as US federal science funding stays unstable and China keeps writing blank checks paired with genuine research autonomy, expect a slow bleed of senior talent, especially in AI-for-science fields where China is explicitly trying to out-invest everyone else.

Do you think this is a one-off, or the start of a real brain drain in AI-adjacent sciences? Hit reply, I read every one.

Leadership Shakeup

OpenAI's Number Two Just Stepped Back, and the Bench Looks Thin

Fidji Simo, OpenAI's CEO of Applications and the executive who'd been running the company's entire business and product side, is stepping down from her full-time role. In a staff note, she said her medical leave for a relapse of a neuroimmune condition has "proven longer and harder than expected," and she'll transition to a part-time advisory position instead.

The details worth knowing:

  • Simo joined OpenAI's board in 2024, then took the newly created CEO of Applications role in May 2025, reporting directly to Sam Altman and absorbing oversight of COO Brad Lightcap, CFO Sarah Friar, and CPO Kevin Weil. At that time, Altman stepped back to focus on research, compute, and safety.

  • She first disclosed her health issues in April. That same memo announced Lightcap was shifting to a "special projects" role and CMO Kate Rouch was leaving to focus on her own cancer recovery.

  • The timing is brutal. Her announcement landed the same day OpenAI launched its GPT-5.6 model family (Sol, Terra, and Luna) and a new agent called ChatGPT Work. Both are explicitly positioned to compete with Anthropic. And the timing also comes on the same day OpenAI confirmed it's shutting down its Atlas browser.

Here's the part that should worry OpenAI watchers as much as the health news itself: this leaves a genuinely thin executive bench at a company valued at $852 billion and reportedly eyeing an IPO. Altman, Lightcap, Friar, and co-founder Greg Brockman are effectively holding the fort, with newer hire Denise Dresser (ex-Slack CEO) seen as a candidate for expanded responsibility. That's a small circle to run a company at this scale, especially one that's simultaneously racing Anthropic on coding tools, trying to launch a consumer super app, and reportedly preparing for public markets.

Simo was widely viewed as the person who'd take on broader operational responsibility if and when OpenAI actually goes public. Her stepping back doesn't kill that plan, but it does remove the person who was arguably best positioned to run the business side of an IPO-track company.

🔮 Prediction: Expect OpenAI to elevate someone into a Simo-shaped hole within the next two quarters, likely Dresser or a new external hire, rather than leave the applications side unowned heading into a potential public offering. Watch whether Brockman's temporary product oversight becomes permanent, because that would be the clearest signal yet of how thin the OpenAI leadership layer actually is.

What do you make of the reshuffle? Is it a distraction or a genuine risk to OpenAI's IPO timeline? Reply and let me know.

30-Second AI Play

🖼️ Stop Meta's AI From Using Your Instagram Photos

Meta just launched Muse Image, a new AI image generator baked into Instagram, WhatsApp, and the Meta AI app. The catch: if your Instagram account is public, anyone can tag your profile and use your photos as raw material for their AI-generated images, and you won't be notified when they do it. Here's how to shut that down in under a minute.

  1. Open Instagram and go to Settings and activity.

  2. Tap into Privacy, then look for Sharing and reuse.

  3. Find the toggle labeled "Allow people to use your content on Instagram with AI features on Meta."

  4. Switch it off for both posts and reels.

  5. If you want maximum protection, consider switching your account to private altogether, since that blocks the reuse feature entirely.

💡 Pro tip: Only private accounts and users under 18 are excluded by default, meaning every public adult account starts opted in. Do this now rather than later, since Meta doesn't notify you retroactively about images already generated using your photos.

Advertise to 180k engineers and CTOs choosing what tools their companies build with.

Other Relevant AI News!

📈 Meta shares jumped 6% on Friday, capping a 15% weekly gain, its best week since early 2024, after an internal memo suggested the company has engineered major cost savings on its AI data center buildout.

🎓 A new survey of over 400 educators finds cheating isn't actually teachers' biggest AI worry. Roughly half say the real problem is no longer being able to tell what students actually understand.

🌐 OpenAI is shutting down Atlas, its standalone AI browser, less than a year after launch, folding its agentic features into ChatGPT's desktop app and a new Chrome extension instead.

💸 Palo Alto Networks CEO Nikesh Arora says AI token costs need to fall roughly 90% over the next two years before enterprise AI deployment truly makes economic sense at scale.

Golden Nuggets

  • 🧪 A Nobel laureate leaving Berkeley for Tsinghua signals China's serious push to recruit AI-adjacent scientific talent, not just AI researchers.

  • 🤖 OpenAI's leadership bench just got thinner right as it eyes an IPO and battles Anthropic on every front.

  • 📸 Meta's Muse Image opts every public Instagram account in by default, so take thirty seconds and opt back out.

Would love to hear your thoughts! Send me your thoughts by replying to this email (yes, I read them all :)

Until our next AI rendezvous,

Anthony | Founder of Uncover AI