
Hard Fork Where Is All the A.I.-Driven Scientific Progress?
138 snips
Dec 26, 2025 In a lively discussion, Sam Rodriques, founder and CEO of FutureHouse and Edison Scientific, unpacks his groundbreaking AI tool, Kosmos, which condenses months of research into just 12 hours. He challenges the hype surrounding AI's potential to cure diseases, citing real-world bottlenecks like patient recruitment and lengthy trial processes. Sam emphasizes that while AI can enhance data analysis and streamline scientific workflows, validation of its outputs remains crucial. This episode dives deep into the realistic applications of AI in scientific discovery and what the future may hold.
AI Snips
Chapters
Transcript
Episode notes
AI That Feels Like A Scientist
- Cosmos feels like an actual AI scientist that delivers deep, sometimes wrong, insights about 80% of the time.
- Edison measured Cosmos by giving it datasets and objectives and finding it reproduced months-long human results overnight.
Benchmarking Against Unpublished Work
- Edison validated the six-month claim by giving Cosmos unpublished datasets that human collaborators had previously analyzed.
- Cosmos found the same discoveries overnight that took researchers months to produce.
Use Structured World Models
- Build long-running structured world models so agents can maintain state and coordinate complex tasks over hours.
- Orchestrate many parallel and sequential agents toward a coherent goal rather than relying on single short prompts.

