
What happened
Johns Hopkins astrophysicist Brice Ménard used Anthropic's Claude Science to build the first complete ultraviolet map of the sky, with AI agents downloading and merging data from multiple space missions and filling gaps by inpainting. In tests, predictions averaged about ten percent deviation from actual measurements.
Why it matters
The result points to AI taking on the tedious data preparation that has kept this kind of project on the shelf, and the map is meant to serve as teaching material. Ménard suspects many scientists have been putting off similar projects that AI could now make possible.
WHO IT HITSResearch groups and university astronomy teams that shelve large data-crunching projects will see a template in which AI agents handle the downloading, calibration, and gap-filling themselves. The roughly ten percent average deviation from actual measurements in tests gives those groups a first benchmark for trusting such output.
Summaries like this, in your inbox every morning.
What made this project possible is partly a physical limit: ultraviolet light is blocked by the ozone layer, so it can only be observed from space, and NASA's GALEX mission left about two-thirds of the sky covered while skipping bright star-forming regions, where dust lit by starlight — clouds around young stars, rings from stellar explosions — shows up most clearly.
The gap was filled not by a new telescope but by Claude Science coordinating AI agents that pulled data from multiple space missions, calibrated it, merged it, and reconstructed the missing areas through inpainting, a technique in which a model learns from existing data to fill holes. The accuracy figures suggest the reconstructed regions track the real sky closely, though the map is positioned first as teaching material rather than as a finished scientific instrument.
The broader signal Ménard draws is about backlog rather than breakthrough: he suspects many scientists have been putting off similar projects that AI could now make possible. That framing — AI as the reason a dormant project finally gets done — is the claim Anthropic is putting forward with the release.
Pick your industry and the AI tools you use, and get news related to your work every day.
Free · 30 seconds with Google · unsubscribe anytimeWhat is AIToday? →
Ask AI anything about this article. The AI reads this article, earlier AIToday articles, and Wikipedia, and cites its sources. Q&As are published on this page for other readers too.
In a post on X on Friday, OpenAI said Jasmine Wang, Tomek Korbak and Mikita Balesni were dismissed for violati…

Sophos CTO John Peterson says agents built through OpenAI's Daybreak cut average MDR case response time from a…

Innodisk and Apacer both reported year-over-year and month-over-month revenue growth for September 2026, suppo…

Applied Materials and Intel are widening their EPIC Center collaboration to cover next-generation transistors…

Anthropic announced Anthropic Cyber Mission on October 8, starting with the Critical Infrastructure Defense Pr…

Nvidia-backed AI data centre operator Firmus withdrew its planned $5 billion IPO on Friday, citing market vola…
