Anthropic Launched Claude Science to Let Literally Anyone Play Mad Scientist
Forget fighting with messy Python scripts and clunky academic databases. The smart folks at Anthropic just dropped a dedicated workbench that acts as a super-smart research assistant. And yes, you don't even need a PhD to play with it.
The newly released application acts as a unified hub designed to replace the chaotic workflow of modern researchers who usually waste half their day jumping between Jupyter notebooks, terminal clusters, and obscure biological databases. Instead of writing custom parsing code for every minor experiment, users can simply type a request in plain English.
Under the hood, the system runs a coordinating agent backed by over 60 curated scientific skills and direct connections to major databases like UniProt and PDB. To prevent the AI from hallucinating a cure for aging, a separate reviewer agent double-checks all citations and math, leaving a clear breadcrumb trail so other humans can actually replicate the results.
During the presentation, Dario Amodei, the chief of Anthropic, admitted that biology is far messier than clean computer code, making it the ultimate test for general-purpose AI. The tool also plugs directly into Nvidia's BioNeMo models, meaning researchers can hook up things like Evo 2 or OpenFold3 without crying over dependency hell.
The app runs locally on macOS and Linux, scaling calculations from a single laptop GPU up to massive high-performance clusters via Modal. Although built for serious labs, Anthropic opened the beta to all paying users on Pro, Team, and Enterprise tiers, while simultaneously launching a program to fund 50 scientific projects with thirty-thousand-dollar grants.
Giving keys to the kingdom to anyone with a premium subscription is a wild gamble that might either democratize breakthroughs or flood preprint servers with AI-generated scientific noise. The line between a brilliant breakthrough and a hallucinated disaster has never been this thin.
Source: Anthropic
Comments
This is where the magic happens: AI reads your discussion and rewrites the article based on the most interesting comments. Each strong comment adds points to the meter below. Once the meter is full, the article updates live — no page reload needed.