Deepmind dismantles its AlphaFold team as key authors leave for Anthropic
Quick Answer
Google Deepmind has disbanded its AlphaFold team, with key researchers leaving for Anthropic and other projects.
Quick Take
This shift marks a departure from their previous focus on long-term scientific challenges, as the team now pivots towards Gemini-powered systems and broader scientific automation.
Key Points
- Nearly 25% of AlphaFold authors have left Google Deepmind.
- Deepmind's new focus includes projects on Gemini, enzyme design, and genomics.
- Key AlphaFold researchers, including John Jumper, have joined Anthropic.
- The restructuring indicates a strategic shift away from dedicated research teams.
- Deepmind aims to compete with OpenAI and Anthropic in frontier AI development.
DeepSignal Analysis
What happened
Google Deepmind has disbanded its AlphaFold team, with many key researchers leaving for Anthropic and other projects. Nearly a quarter of the original authors have departed the company, marking a significant shift in focus towards Gemini-powered systems and broader scientific automation.
Key evidence
- Most of the original authors of the AlphaFold papers were reassigned internally over the past year, with nearly a quarter leaving Google Deepmind entirely.
- Deepmind confirmed that researchers are now working on projects related to Google's Gemini large language model, enzyme design, nuclear fusion, and genomics.
- Key researchers, including John Jumper and Jonas Adler, have moved to Anthropic and Google's internal 'Code Strike' team, which aims to enhance AI coding capabilities.
Why it matters
The disbanding of the AlphaFold team represents a departure from Deepmind's previous strategy of focusing on long-term scientific challenges. This shift raises questions about the future direction of the company and its commitment to groundbreaking scientific research, especially as it competes with other AI firms like OpenAI and Anthropic.
What to watch
Source Excerpt
The majority of the researchers behind AlphaFold are now working on other projects, and almost a quarter have left Google Deepmind altogether. The restructuring marks a sharp turn away from the strategy that put the lab on the map.
Want this in your inbox every morning?
Daily brief at your local 8am — bilingual EN/中文, free.
More from The Decoder
See more →
An AI model programmed nonstop for 19 days on a single MirrorCode task that cost $2,600 to run
Epoch AI's MirrorCode benchmark reveals Claude Opus 4.7 as the leader with a 56% solve rate, reconstructing a 16,000-line toolkit in 14 hours. Despite this, all models tested struggle with the most complex tasks, highlighting limitations in current AI capabilities. The single task consumed $2,600 over 19 days, raising questions about cost-effectiveness in AI development.
