Zuckerberg offered him 1.5 billion to change jobs – the AI expert said: “Nope”

Imagine Mark Zuckerberg offers you $1.5 billion to quit your job and join his team. Would you accept the offer? Probably, right? Andrew Tulloch didn't. That's a rejection.

But how did this happen?

According to the Wall Street Journal, Meta CEO Zuckerberg wanted to acquire the AI startup Thinking Machines Lab (TML), founded by former OpenAI CTO Mira Murati. The startup is currently considered one of the most interesting projects in the field of generative AI and was valued at around $12 billion just a few months after its founding.

However, Murati is said to have rejected Zuckerberg's takeover offer.

Zuckerberg then attempted to gain access to TML's AI expertise through other means. According to the Wall Street Journal report, he launched a targeted poaching campaign to recruit key members of the startup team for Meta.

Zuckerberg was particularly interested in Andrew Tulloch, co-founder of TML and one of the leading ML engineers of his generation. Tulloch was reportedly offered a compensation package worth up to $1.5 billion – spread over six years, with the prospect of significantly higher earnings if Meta's shares performed well.

But Tulloch declined, as did all the other Thinking Machines Lab team members approached by Meta.

Meta denied the reports. In a statement to the Wall Street Journal, it said Meta did not want to acquire TML and that the amount of the job offer was "inaccurate and ridiculous."

Who is this man who says no to more than a billion? Andrew Tulloch describes himself on his LinkedIn profile simply as "Member of Technical Staff" at Thinking Machines Lab . A rather modest description considering that, according to media reports, he is the co-founder of a multi-billion dollar AI startup. This reticence characterizes his entire resume: Even at his previous employer, OpenAI, Tulloch uses the same title, even though he and Murati were instrumental in the development of the GPT models 40 and 4.5.

A quick look back: The Australian-born man completed his bachelor's degree in mathematics at the University of Sydney with outstanding results and even received the prestigious "University Medal" for his work.

This was followed by a Master's degree in Mathematical Statistics and Machine Learning at the University of Cambridge, which he also completed with distinction.

Tulloch began his career while still studying for his bachelor's degree at Goldman Sachs, where he developed quantitative models for the trading team. Shortly after graduating in 2012, he joined Facebook (now Meta), where he spent the next eleven years. As a Distinguished Engineer, he played a key role in building the ML infrastructure and, among other things, was involved in the development of the PyTorch framework.

In 2023, Tulloch moved from Meta to OpenAI, where he played a key role in training the GPT-4o and GPT-4.5 models. Just one year later, he founded the startup Thinking Machines Lab with Mira Murati. The company develops flexible and multimodal AI systems, deliberately pursuing a counter-approach to the centralized models of major tech companies.

Thinking Machines Lab now employs around 50 people. Even before its official market launch, the company reached a valuation of over $12 billion in its most recent seed funding round.

Why did Tulloch reject the offer? He hasn't officially commented yet. A request from Gründerszene has so far gone unanswered. Sources close to him say Tulloch supports Murati's vision, is critical of Meta's heavy reliance on advertising, and doesn't want to let money take away from the creative freedom he currently enjoys.

However, his decision may also have been simply financially prudent. As co-founder of Thinking Machines Lab, he likely holds shares that could become significantly more valuable in the future than Meta's multi-billion dollar offering.

Should TML launch a successful product in the coming years or move toward a strategic exit, Tulloch is likely to benefit significantly more financially than an employment contract—even at Meta—would allow. Outside observers, therefore, suggest in the Financial Express that his decision may have been less idealistic and more a calculated bet on higher long-term equity values.

It's no secret that Zuckerberg has been specifically seeking AI talent for some time, willingly investing large sums in the process. OpenAI employees are particularly high on his wish list—and this strategy has recently proven quite successful.

According to the Wall Street Journal, Zuckerberg even personally takes care of the recruitment.

For example, Lucas Beyer, Alexander Kolesnikov, and Xiaohua Zhai moved from OpenAI to Meta. All three had previously played key roles in establishing OpenAI's Zurich office and are now part of Meta's newly formed Superintelligence team. According to media reports, they even received a signing bonus of around $100 million – a sum that Beyer has denied as "fake news."

In addition, in July 2025, Meta hired two more OpenAI researchers specializing in multimodal AI: Allan Jabri and Lu Liu, who will also support the Superintelligence team, The Information reported, citing a source. Shengjia Zhao, one of the key minds behind ChatGPT and GPT-4, also moved directly from OpenAI to Meta, taking on the role of Chief Scientist in the Superintelligence Lab.

In addition to established experts from the OpenAI community, Zuckerberg also has his eye on young, up-and-coming talent. One example is Matt Deitke: The 24-year-old dropped out of his PhD program to found an AI startup.

Zuckerberg was desperate to win over Deitke for Meta's new AI initiative. According to a New York Times report, the initial offer was approximately $125 million in stock and cash over four years. Deitke initially declined, wanting to continue running his startup Vercept.

Zuckerberg then reportedly met with Deitke personally and doubled his offer to approximately $250 million—with up to $100 million in the first year alone. This sum was apparently so exceptionally high that Deitke reportedly sought advice from other researchers before finally accepting.

The New York Times compares the competition for top AI talent to the bidding war for NBA stars: Young AI researchers are now being traded similarly to Steph Curry or LeBron James—equipped with nine-figure compensation packages over several years. But unlike in the NBA, where salaries are capped by the salary cap, companies like Meta, Google, and OpenAI have no upper limits—making the competition for these top talents even more intense.