Our experiments to support founders to protect humanity
bluedot.org · by Dewi Erwan · clipped 2025-11-26
Preview clipped from the web into my Obsidian — read the full piece at the source.
Blog
Dewi Erwan • November 3, 2025
Summary
- BlueDot is recruiting founders to solve critical problems threatening civilization.
- Our approach: deep problem scoping, targeted recruitment, and capital coordination.
- First focus: pandemic PPE stockpiles.
Context
- BlueDot is a non-profit building the workforce that protects humanity.
- We’re a 5-person team. We train 1,000s of people each year in AI safety and biosecurity, and we place the highest-potential people into impactful jobs at AI companies, governments and non-profits.
- Since 2022, we’ve raised $34M.
- Today, ~2,000 people work full-time on AI safety. We believe this needs to increase to 100,000 by 2030 to prevent worst-case scenarios.
- To help scale the workforce, we’re testing a “startup studio” model alongside our courses. We’re attracting leaders, builders and entrepreneurs, and helping them start and scale the most important projects that protect humanity.
- We’re building this in collaboration with Jonah Weinbaum and Hugo Walrand.
Initial approach: AGI Strategy course
- Two months ago, we launched our AGI Strategy course alongside a $1M fund.
- We’ve attracted 15 founders to the course who’ve raised $1M+ each, and ~$50M in total.
- We’ve invited a handful of them to join our in-person “ incubator weeks ”, where we’ve helped them improve their ideas and launch new AI safety companies.
- We’ve deployed $130k in grants so far to 1) build a risk engine for AI insurance and 2) improve the cybersecurity of critical infrastructure.
- This approach works well for founders who already have a particular area of expertise, project ideas that leverage that expertise, and are motivated to work on AI safety.
- To test what it takes to drive founder-energy towards a shortlist of high-priority ideas, we’re running a second experiment in parallel.
Second experiment: Problem scoping and serial entrepreneurs
- Compared to other AI safety incubators, we’re taking a different approach:
-
- Doing deep in-house problem scoping before recruiting anyone, and
-
- Targeting serial entrepreneurs rather than junior talent.
-
- We’ve spent the past 3 weeks going deep on pandemic PPE stockpiles.
- We’ve done 50+ expert interviews, asked for intros from everyone, read many papers and blog posts, and shared a commentable doc with everything we’ve learnt and our understanding of the situation with those experts.
- We’re writing a “landscape analysis” to understand the current situation, and help us develop a strategy for how this problem will get solved.
- We’re headhunting serial entrepreneurs with relevant experience in the problem area we’re focused on, and making the most compelling pitch that we can for why they should work on solving this problem.
- Founding and scaling a successful company is exceptionally hard, and the best predictor of future success is past success.
- We make the pitch personal and concrete:
- “Here’s the threat and its scale. Your expertise in X is exactly what’s needed to make Y happen. This is why we believe it’s solvable. Here’s the funding situation (validated business model or $100M+ in philanthropy). Here are the experts we can introduce you to. Your work could save millions of lives.”