# @scspai on YouTube

- **Type:** Video
- **Original URL:** https://youtube.com/watch?v=CYjGi1Y1fDg
- **Gondola URL:** https://gondola.cc/posts/68254426-scspai-youtube
- **Thumbnail:** https://img.gondola.cc/tr:w-,h-,fo-auto/postThumbnails/5ae4cd7b66.jpg
- **Posted:** 2026-07-21T20:00:27.000+00:00
- **Account Owner:** Special Competitive Studies Project (@scspai) — https://gondola.cc/scspai

## Caption

What happens to the scientific method when AI can explore a hundred research paths at once?

At SCSP's AI+ Discovery Summit, Dr. Nathan Frey, who leads drug discovery and development research at Anthropic, sits down with MIT Technology Review's Amy Nordrum to talk about how frontier AI models are changing the way science gets done.

In this conversation, Frey explains why he believes AI's biggest near-term impact is cutting "research toil," letting scientists pursue more ideas in parallel and kill dead ends faster. He pushes back on the claim that biology is too messy for AI, arguing that models are no longer limited to existing data because designing experiments and building new datasets on the fly is now part of the workflow. He also breaks down Claude for Science, Anthropic's new product for researchers, including its reviewer agents that check work in the background and flag errors automatically.

Topics covered:

Why problem selection is the hardest skill in science, and how AI helps
Anthropic's outcomes-driven approach vs. benchmark chasing
Interpretability and why scientists still need to understand the "why"
The gap between AI for discovery and AI for commercialization
Where Claude's biotech capabilities are headed next

Recorded live at the AI+ Discovery Summit, hosted by the Special Competitive Studies Project (SCSP).

Topics covered:

Why problem selection is the hardest skill in science, and how AI helps
Anthropic's outcomes-driven approach vs. benchmark chasing
What John Jumper's move from Google DeepMind to Anthropic signals
Interpretability and why scientists still need to understand the "why"
The gap between AI for discovery and AI for commercialization
Where Claude's biotech capabilities are headed next

Recorded live at the AI+ Discovery Summit, hosted by the Special Competitive Studies Project (SCSP).

Keep up with SCSP on our website and social media:
https://www.scsp.ai/ 
https://www.youtube.com/@SCSPAI 
https://www.instagram.com/scsp.ai/?hl=en 
https://www.linkedin.com/company/specialcompetitivestudiesproject/posts/
https://substack.com/@scsp/notes
https://www.tiktok.com/@scsp.ai
https://x.com/scsp_ai


#AI #DrugDiscovery #Anthropic #Claude #Biotech #ArtificialIntelligence #ScientificDiscovery #SCSP

## Stats

- **Views:** 160
- **Likes:** 5
- **Shares:** 0
- **Comments:** 0

## Tags

scsp, claude, anthropic, scientificdiscovery, drugdiscovery, biotech, ai, artificialintelligence

---
Copyright (c) Gondola