Anthropic says its AI, Claude, discovered a new enzyme system that it compares to the machinery behind Crispr gene editing. It is the first result from the company's new wet lab. Nearly 950 Claude agents worked for 21 hours, using 210 million tokens, searching a huge DNA database. One spotted an unusual repeating pattern and flagged it for humans to check. Lab tests confirmed a previously unknown enzyme system found in bacteriophages, viruses that infect bacteria. Scientists say it is still unclear what it does.
Anthropic says human scientists only set the first prompt and did the lab work. The company chose to share the finding early to show what Claude can do. It is still unclear whether the enzyme system has real uses, let alone whether it will be as big as Crispr. Anthropic admits the announcement is premature. The news comes as Anthropic prepares to go public and tries to attract scientists, while AI firms race to prove their models can aid research.
If AI can genuinely speed up science, it could change medicine and biotech. But early hype risks overstating results, so experts will watch whether the discovery leads to real breakthroughs.

Anthropic says its AI, Claude, discovered a new enzyme system that it compares to the machinery behind Crispr gene editing. It is the first result from the company's new wet lab. Nearly 950 Claude agents worked for 21 hours, using 210 million tokens, searching a huge DNA database. One spotted an unusual repeating pattern and flagged it for humans to check. Lab tests confirmed a previously unknown enzyme system found in bacteriophages, viruses that infect bacteria. Scientists say it is still unclear what it does.

Anthropic says human scientists only set the first prompt and did the lab work. The company chose to share the finding early to show what Claude can do. It is still unclear whether the enzyme system has real uses, let alone whether it will be as big as Crispr. Anthropic admits the announcement is premature. The news comes as Anthropic prepares to go public and tries to attract scientists, while AI firms race to prove their models can aid research.

If AI can genuinely speed up science, it could change medicine and biotech. But early hype risks overstating results, so experts will watch whether the discovery leads to real breakthroughs.

πŸ“° Source: News Source
theverge.com β†—
Was this article useful?