A Nobel laureate has joined Anthropic!
Today, John Jumper, a core leader of AlphaFold, officially announced that he is leaving Google DeepMind, where he had worked for nearly nine years, to join Anthropic.
The Nobel laureate who rewrote the entire field of structural biology with an AI model has turned and left.
Hassabis quickly responded: "Thank you, John, for your extraordinary partnership over the past nine years! What we have achieved with AlphaFold has changed the world."
Nine years of collaboration and sharing a Nobel Prize – this is probably the most dignified farewell in the tech world.
Just two days earlier, Noam Shazeer, the legendary co-leader of the Transformer paper and co-head of Gemini, announced his departure from Google to join OpenAI.
In less than 72 hours, Google lost two trump cards.
One was bought back for $2.7 billion but couldn't be kept, and another was a 9-year relationship but couldn't be kept either.
Six months after graduating with a PhD, I directly led the AlphaFold team.
In the life sciences community, John Jumper is practically synonymous with "rewriting the entire discipline with AI."
Jumper was born in 1985 in Little Rock, Arkansas, an ordinary small town in the American South.
I earned double degrees in mathematics and physics from Vanderbilt University for my undergraduate studies, and then went on to pursue a PhD at the University of Chicago, specializing in theoretical chemistry. Specifically, I used computational methods to simulate the dynamic behavior of proteins.
Mathematics gave him the intuition to model, physics gave him the understanding of complex systems, and theoretical chemistry made him understand the problem of proteins better than any pure AI researcher.
These three directions combined represent the rarest combination of knowledge for solving the protein folding problem.
After receiving his PhD in 2017, Jumper joined DeepMind immediately.
It is worth noting that at that time he had almost no experience in deep learning, and the most prominent feature on his resume was not his mastery of neural networks, but his understanding of protein physics.
But that's exactly what Hassabis valued.
Immediately afterwards, he made a decision that no one expected—to let this young man, who had only graduated 6 months ago and was still learning deep learning on the job, directly lead the AlphaFold team.
There is no transition period, no "work as a researcher for a few years to accumulate experience".
Hassabis bet that understanding proteins is more important than understanding AI in solving the problem of protein folding. Jumper, on the other hand, is taking on the biggest gamble in the entire field of computational biology.
One person has increased biology 1000 times.
What happened in the following years can only be described as "outrageous"—
In 2018, AlphaFold made its debut at the CASP protein structure prediction competition, crushing traditional methods.
In 2020, AlphaFold 2 emerged, solving the protein folding problem that had plagued biologists for 50 years with a single AI model.
In 2021, Jumper led a team that calculated the 3D structures of almost all of the more than 50,000 human proteins. Ultimately, they achieved the generation of structures for approximately one million species and nearly 200 million known proteins.
Before AlphaFold, humans spent decades using experimental techniques such as X-ray crystallography and cryo-electron microscopy to solve approximately 200,000 protein structures.
Jumper's team grew 1000 times in one go.
It is no exaggeration to say that AlphaFold completed in a few months what biologists had not been able to do in the past hundred years.
In May 2024, AlphaFold 3 was released—it no longer just predicts proteins, but also the interactions between DNA, RNA, and small molecule drugs, making it an all-around solution. Its protein-ligand docking accuracy reached 76.4%, a 1.8-fold improvement over previous methods.
Five months later in Stockholm, John Jumper and Demis Hassabis stood together on the Nobel Prize in Chemistry podium.
That year, Jumper was 39 years old, making him the youngest Nobel laureate in chemistry in 70 years.
It only took him 7 years to go from a PhD graduate who had to learn deep learning on the job to standing in the spotlight in Stockholm.
Thus, the return on Hassabis's bet back then is probably among the highest in the history of human science.
So his departure today is a pain for Google DeepMind that goes beyond simply losing a director.
What's wrong with Google?
After the news broke, the comment section on X went wild.
User Chubby exclaimed, "This is a huge loss for Google, and absolutely insane for Anthropic!"
Some netizens remarked, "Anthropic has welcomed a Nobel laureate, and talent is continuously concentrating on OpenAI and Anthropic." Others exclaimed, "First it was Karpathy, and now it's the person behind AlphaFold. Anthropic is assembling an AI Avengers."
Logan Kilpatrick jokingly said he hoped Jumper would "win another Nobel Prize." While the tone was playful, upon closer inspection, it wasn't entirely an exaggeration.
After the initial shock, everyone was asking the same question – what exactly happened to Google?
Jumper didn't say, Anthropic didn't say, and Google didn't say.
Perhaps the closest answer comes from a comment by investor Lior Alexander.
"The cutting-edge AI lab is selling something that Google can't provide: the feeling that one person can change the trajectory of a company."
The person bought for $2.7 billion couldn't be kept.
Just two days before Jumper's official announcement, Noam Shazeer announced his departure from Google to join OpenAI as "Head of Architecture Research".
He was one of the core authors of "Attention Is All You Need," a foundational work on modern AI published in 2017. He designed multi-head attention, and he typed out the first usable implementation that outperformed state-of-the-art performance line by line.
Google spent $2.7 billion to bring him back from Character.AI.
Upon his return, Shazeer became co-head of Gemini and was instrumental in Google's counterattack against its large-scale model.
Less than two years later, he left too. Two days later, Jumper also left.
They are neither the first nor the last.
Over the past eight years, more than 20 top researchers who authored landmark papers have left DeepMind/Brain.
In 2025 alone, at least 11 executives will leave the company. DeepMind co-founder Mustafa Suleyman himself was poached by Microsoft in a $650 million acqui-hire round.
Life sciences: the next battleground for the AI trio
Returning to Anthropic, the planning began more than two months ago.
On April 3, Anthropic acquired biotechnology company Coefficient Bio for $400 million in stock. The team, consisting of fewer than 10 people, has already achieved industry-leading results in AI-driven antibody design.
At the same time, Anthropic is also building its own wet laboratory. Last October, it launched Claude for Life Sciences to help researchers accelerate drug discovery and biological experiment design, and this January it launched Claude for Healthcare for medical institutions.
They said the goal was to shorten the research and development cycle in life sciences by 10 times. And now, a Nobel Prize-winning protein scientist has taken on this task.
In fact, Anthropic is not the only one betting on life sciences.
In April of this year, OpenAI released GPT-Rosalind, an inference model specifically designed for biomedicine, focusing on drug discovery, genome analysis, and protein engineering. It has already partnered with leading pharmaceutical companies such as Amgen, Moderna, and Thermo Fisher.
The OpenAI Foundation has stated directly that it will invest no less than $1 billion in life sciences over the next year. With the recent recruitment of Shazeer to lead architecture research, OpenAI is making a strong push in this field.
Meanwhile, Google DeepMind's Hassabis subsidiary, Isomorphic Labs, raised $600 million last year and signed a collaboration agreement with Eli Lilly and Novartis with a total milestone value of up to $3 billion. AlphaFold's technological foundation remains an industry benchmark.
Three labs have simultaneously placed their bets in the same direction—rewriting life sciences with AI.
Jumper's choice is just the latest step in this grand scheme.
References:
https://x.com/JohnJumperSci/status/2068001285173834106
Edited by: Moses
This article is from the WeChat public account "New Zhiyuan", author: ASI Revelation.
