Skip to main content
See every side of every news story
Published loading...Updated

FutureHouse Spins Out Edison Scientific for AI Research

Kosmos automates research by analyzing 1,500 papers and coding 42,000 lines per run, with 79.4% of its claims validated by PhD experts, Edison says.

  • Edison Scientific, the US-based firm, has been building Kosmos for about two years, a collection of AI agents Edison claims produced seven discoveries, four of which were novel.
  • By processing pre-prepared datasets, a typical Kosmos run lasts up to 12 hours, searches around 1,500 papers, writes 42,000 lines of code, but Edison says Kosmos `actually has completed probably 10 per cent of the task`.
  • On specific claims, PhD evaluators found that 79.4 supported overall, including 85.5 for data claims and 82.1 for literature, but critics dispute the SOD2 finding's novelty.
  • Responses have been mixed, with Sam Rodriques, Edison spokesperson and leader, saying Kosmos should collaborate with scientists and admitting some findings may be flawed, while the team has engaged with social media queries.
  • Experts warn Kosmos could speed research—20 cycles equal around six months of human work—but Ben Glocker, Imperial College London, and Noah Giansiracusa, Bentley University, urge human oversight.
Insights by Ground AI

22 Articles

Think freely.Subscribe and get full access to Ground NewsSubscriptions start at $9.99/yearSubscribe

Bias Distribution

  • 100% of the sources are Center
100% Center

Factuality Info Icon

To view factuality data please Upgrade to Premium

Ownership

Info Icon

To view ownership data please Upgrade to Vantage

TechCrunch broke the news in United States on Wednesday, March 5, 2025.
Too Big Arrow Icon
Sources are mostly out of (0)

Similar News Topics

News
Feed Dots Icon
For You
Search Icon
Search
Blindspot LogoBlindspotLocal