The Knowledge Project

Greg Brockman: Inside the 72 Hours That Almost Killed OpenAI

1h 12min

Greg Brockman recounts the founding of OpenAI, detailing his personal mission for Artificial General Intelligence (AGI) and the critical transition from a non-profit to a for-profit entity to secure the massive computational resources required. He shares insights into key technological breakthroughs like Dota and GPT-4, and provides a firsthand account of the dramatic 72 hours surrounding Sam Altman's firing, emphasizing the extraordinary employee loyalty that ultimately saved the company. Brockman also discusses OpenAI's strategic approach to iterative deployment, safety as a core product feature, and the profound future impact of AI on society, jobs, and personal empowerment.

Summarized by Podsumo

🎧 Listen 🎙️ Ask about this episode

✨ Key Takeaways

1

OpenAI's Origin Story: Greg Brockman's journey from Stripe, the initial gathering of top researchers in 2015, and the Napa offsite that laid out the foundational technical plan for reinforcement and unsupervised learning.
2

The Shift to For-Profit: The crucial realization in 2017 that achieving AGI required immense computational resources, leading to the decision by Elon Musk, Sam Altman, Ilya Sutskever, and Greg Brockman to create a for-profit structure, as non-profit fundraising had a clear cap.
3

The 72-Hour Crisis: Brockman's firsthand account of Sam Altman's firing, his immediate resignation, the overwhelming employee loyalty that crashed Google Docs with a petition, and how not a single person accepted a competing offer during the chaos.
4

AI's Exponential Growth & Impact: Discussion on AI now writing most code, solving open math and physics problems (even contradicting community expectations), and the vision for personal AI agents that proactively manage long-term goals.
5

Iterative Deployment & Safety: OpenAI's strategy of iterative deployment to learn from real-world usage (e.g., GPT-3's unexpected misuse for medical spam) and the commitment to safety as a core product feature and societal resilience.

💬 Notable Quotes

"“If you can actually make a difference in how AI will play out in the world like that would be a life well lived.”"
"“The actual writing of code, currently, the AI is much better than humans at writing code, given the right context, given the right structure.”"
"“I think that my view of AI is it is about empowerment. It is about human agency.”"